gnupg/tools/gpgconf-comp.c

3462 lines
101 KiB
C
Raw Normal View History

/* gpgconf-comp.c - Configuration utility for GnuPG.
* Copyright (C) 2004, 2007-2011 Free Software Foundation, Inc.
* Copyright (C) 2016 Werner Koch
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
* Copyright (C) 2020 g10 Code GmbH
2006-06-20 19:21:37 +02:00
*
* This file is part of GnuPG.
*
* GnuPG is free software; you can redistribute it and/or modify it
* under the terms of the GNU General Public License as published by
2007-07-04 21:49:40 +02:00
* the Free Software Foundation; either version 3 of the License, or
2006-06-20 19:21:37 +02:00
* (at your option) any later version.
*
* GnuPG is distributed in the hope that it will be useful, but
* WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with GnuPG; if not, see <https://www.gnu.org/licenses/>.
2006-06-20 19:21:37 +02:00
*/
#if HAVE_CONFIG_H
#include <config.h>
#endif
#include <stdlib.h>
#include <stdio.h>
#include <string.h>
#include <fcntl.h>
#include <unistd.h>
#include <sys/types.h>
#include <errno.h>
#include <time.h>
#include <stdarg.h>
2010-03-29 14:57:11 +02:00
#ifdef HAVE_SIGNAL_H
# include <signal.h>
#endif
#include <ctype.h>
2007-06-26 15:48:44 +02:00
#ifdef HAVE_W32_SYSTEM
# define WIN32_LEAN_AND_MEAN 1
# include <windows.h>
#else
# include <pwd.h>
# include <grp.h>
#endif
#include "../common/util.h"
#include "../common/i18n.h"
#include "../common/exechelp.h"
#include "../common/sysutils.h"
#include "../common/status.h"
#include "../common/gc-opt-flags.h"
#include "gpgconf.h"
2007-06-26 15:48:44 +02:00
#if (__GNUC__ > 2 || (__GNUC__ == 2 && __GNUC_MINOR__ >= 5 ))
void gc_error (int status, int errnum, const char *fmt, ...) \
__attribute__ ((format (printf, 3, 4)));
#endif
/* Output a diagnostic message. If ERRNUM is not 0, then the output
is followed by a colon, a white space, and the error string for the
error number ERRNUM. In any case the output is finished by a
newline. The message is prepended by the program name, a colon,
and a whitespace. The output may be further formatted or
redirected by the jnlib logging facility. */
void
gc_error (int status, int errnum, const char *fmt, ...)
{
va_list arg_ptr;
va_start (arg_ptr, fmt);
log_logv (GPGRT_LOGLVL_ERROR, fmt, arg_ptr);
va_end (arg_ptr);
if (errnum)
log_printf (": %s\n", strerror (errnum));
else
log_printf ("\n");
if (status)
{
log_printf (NULL);
log_printf ("fatal error (exit status %i)\n", status);
gpgconf_failure (gpg_error_from_errno (errnum));
}
}
/* Forward declaration. */
static void gpg_agent_runtime_change (int killflag);
static void scdaemon_runtime_change (int killflag);
tpm2d: Add tpm2daemon code * tpm2d: New directory. * Makefile.am (SUBDIRS): Add directory. * configure.ac: Detect libtss and decide whether to build tpm2d. * am/cmacros.am: Add a define. * util.h (GNUPG_MODULE_NAME_TPM2DAEMON): New. * common/homedir.c (gnupg_module_name): Add tpm2d. * common/mapstrings.c (macros): Add "TPM2DAEMON". * tools/gpgconf.h (GC_COMPONENT_TPM2DAEMON): New. * tools/gpgconf-comp.c (known_options_tpm2daemon): New. (gc_component): Add TPM2. (tpm2daemon_runtime_change): New. * tpm2d/Makefile.am: New. * tpm2d/command.c: New. * tpm2d/ibm-tss.h: New. * tpm2d/tpm2.c: New. * tpm2d/tpm2.h: New. * tpm2d/tpm2daemon.c: New. * tpm2d/tpm2daemon.h: New. --- This commit adds and plumbs in a tpm2daemon to the build to mirror the operation of scdaemon. The architecture of the code is that tpm2daemon.c itself is pretty much a clone of scd/scdaemon.c just with updated function prefixes (this argues there could be some further consolidation of the daemon handling code). Note that although this commit causes the daemon to be built and installed, nothing actually starts it or uses it yet. Command handling ---------------- command.c is copied from the command handler in scd.c except that the command implementation is now done in terms of tpm2 commands and the wire protocol is far simpler. The tpm2daemon only responds to 4 commands IMPORT: import a standard s-expression private key and export it to TPM2 format. This conversion cannot be undone and the private key now can *only* be used by the TPM2. To anyone who gets hold of the private key now, it's just an encrypted binary blob. PKSIGN: create a signature from the tpm2 key. The TPM2 form private key is retrieved by KEYDATA and the hash to be signed by EXTRA. Note there is no hash specifier because the tpm2 tss deduces the hash type from the length of the EXTRA data. This is actually a limitation of the tpm2 command API and it will be interesting to see how this fares if the tpm2 ever supports say sha3-256 hashes. PKDECRYPT: decrypt (RSA case) or derive (ECC case) a symmetric key. The tpm2 for private key is retrieved by KEYDATA and the information used to create the symmetric key by EXTRA. KILLTPM2D: stop the daemon All the tpm2 primitives used by command.c are in tpm2.h and all the tpm2 specific gunk is confined to tpm2.c, which is the only piece of this that actually does calls into the tss library. Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com> Changes from James' patch: - gpgconf: The displayed name is "TPM" and not "TPM2". That string is used by GUIs and should be something the user understands. For example we also use "network" instead of "Dirmngr". - Removed some commented includes. - Use 16 as emulation of GPG_ERR_SOURCE_TPM2. - Silenced a C90 compiler warning and flags unused parameters. - Removed "if HAVE_LIBS" from tpm2/Makefile.am and add missing files so that make distcheck works. Signed-off-by: Werner Koch <wk@gnupg.org>
2021-03-09 22:50:28 +01:00
static void tpm2daemon_runtime_change (int killflag);
static void dirmngr_runtime_change (int killflag);
static void keyboxd_runtime_change (int killflag);
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
/* STRING_ARRAY is a malloced array with malloced strings. It is used
* a space to store strings so that other objects may point to these
* strings. It shall never be shrinked or any items changes.
* STRING_ARRAY itself may be reallocated to increase the size of the
* table. STRING_ARRAY_USED is the number of items currently used,
* STRING_ARRAY_SIZE is the number of calloced slots. */
static char **string_array;
static size_t string_array_used;
static size_t string_array_size;
/* Option configuration. */
/* An option might take an argument, or not. Argument types can be
basic or complex. Basic types are generic and easy to validate.
Complex types provide more specific information about the intended
use, but can be difficult to validate. If you add to this enum,
don't forget to update GC_ARG_TYPE below. YOU MUST NOT CHANGE THE
NUMBERS OF THE EXISTING ENTRIES, AS THEY ARE PART OF THE EXTERNAL
INTERFACE. */
typedef enum
{
/* Basic argument types. */
/* No argument. */
GC_ARG_TYPE_NONE = 0,
/* A String argument. */
GC_ARG_TYPE_STRING = 1,
/* A signed integer argument. */
GC_ARG_TYPE_INT32 = 2,
/* An unsigned integer argument. */
GC_ARG_TYPE_UINT32 = 3,
/* ADD NEW BASIC TYPE ENTRIES HERE. */
/* Complex argument types. */
/* A complete filename. */
GC_ARG_TYPE_FILENAME = 32,
/* An LDAP server in the format
HOSTNAME:PORT:USERNAME:PASSWORD:BASE_DN. */
GC_ARG_TYPE_LDAP_SERVER = 33,
/* A 40 character fingerprint. */
GC_ARG_TYPE_KEY_FPR = 34,
/* A user ID or key ID or fingerprint for a certificate. */
GC_ARG_TYPE_PUB_KEY = 35,
/* A user ID or key ID or fingerprint for a certificate with a key. */
GC_ARG_TYPE_SEC_KEY = 36,
/* A alias list made up of a key, an equal sign and a space
separated list of values. */
GC_ARG_TYPE_ALIAS_LIST = 37,
/* ADD NEW COMPLEX TYPE ENTRIES HERE. */
/* The number of the above entries. */
GC_ARG_TYPE_NR
} gc_arg_type_t;
/* For every argument, we record some information about it in the
following struct. */
static const struct
{
/* For every argument type exists a basic argument type that can be
used as a fallback for input and validation purposes. */
gc_arg_type_t fallback;
/* Human-readable name of the type. */
const char *name;
} gc_arg_type[GC_ARG_TYPE_NR] =
{
/* The basic argument types have their own types as fallback. */
{ GC_ARG_TYPE_NONE, "none" },
{ GC_ARG_TYPE_STRING, "string" },
{ GC_ARG_TYPE_INT32, "int32" },
{ GC_ARG_TYPE_UINT32, "uint32" },
/* Reserved basic type entries for future extension. */
{ GC_ARG_TYPE_NR, NULL }, { GC_ARG_TYPE_NR, NULL },
{ GC_ARG_TYPE_NR, NULL }, { GC_ARG_TYPE_NR, NULL },
{ GC_ARG_TYPE_NR, NULL }, { GC_ARG_TYPE_NR, NULL },
{ GC_ARG_TYPE_NR, NULL }, { GC_ARG_TYPE_NR, NULL },
{ GC_ARG_TYPE_NR, NULL }, { GC_ARG_TYPE_NR, NULL },
{ GC_ARG_TYPE_NR, NULL }, { GC_ARG_TYPE_NR, NULL },
{ GC_ARG_TYPE_NR, NULL }, { GC_ARG_TYPE_NR, NULL },
{ GC_ARG_TYPE_NR, NULL }, { GC_ARG_TYPE_NR, NULL },
{ GC_ARG_TYPE_NR, NULL }, { GC_ARG_TYPE_NR, NULL },
{ GC_ARG_TYPE_NR, NULL }, { GC_ARG_TYPE_NR, NULL },
{ GC_ARG_TYPE_NR, NULL }, { GC_ARG_TYPE_NR, NULL },
{ GC_ARG_TYPE_NR, NULL }, { GC_ARG_TYPE_NR, NULL },
{ GC_ARG_TYPE_NR, NULL }, { GC_ARG_TYPE_NR, NULL },
{ GC_ARG_TYPE_NR, NULL }, { GC_ARG_TYPE_NR, NULL },
/* The complex argument types have a basic type as fallback. */
{ GC_ARG_TYPE_STRING, "filename" },
{ GC_ARG_TYPE_STRING, "ldap server" },
{ GC_ARG_TYPE_STRING, "key fpr" },
{ GC_ARG_TYPE_STRING, "pub key" },
{ GC_ARG_TYPE_STRING, "sec key" },
{ GC_ARG_TYPE_STRING, "alias list" },
};
/* Every option has an associated expert level, than can be used to
hide advanced and expert options from beginners. If you add to
this list, don't forget to update GC_LEVEL below. YOU MUST NOT
CHANGE THE NUMBERS OF THE EXISTING ENTRIES, AS THEY ARE PART OF THE
EXTERNAL INTERFACE. */
typedef enum
{
/* The basic options should always be displayed. */
GC_LEVEL_BASIC,
/* The advanced options may be hidden from beginners. */
GC_LEVEL_ADVANCED,
/* The expert options should only be displayed to experts. */
GC_LEVEL_EXPERT,
/* The invisible options should normally never be displayed. */
GC_LEVEL_INVISIBLE,
/* The internal options are never exported, they mark options that
are recorded for internal use only. */
GC_LEVEL_INTERNAL,
/* ADD NEW ENTRIES HERE. */
/* The number of the above entries. */
GC_LEVEL_NR
} gc_expert_level_t;
/* A description for each expert level. */
static const struct
{
const char *name;
} gc_level[] =
{
{ "basic" },
{ "advanced" },
{ "expert" },
{ "invisible" },
{ "internal" }
};
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
/* Option flags. The flags which are used by the components are defined
by gc-opt-flags.h, included above.
YOU MUST NOT CHANGE THE NUMBERS OF THE EXISTING FLAGS, AS THEY ARE
PART OF THE EXTERNAL INTERFACE. */
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
/* Some entries in the emitted option list are not options, but mark
the beginning of a new group of options. These entries have the
GROUP flag set. Note that this is internally also known as a
header line. */
#define GC_OPT_FLAG_GROUP (1UL << 0)
/* The ARG_OPT flag for an option indicates that the argument is
optional. This is never set for GC_ARG_TYPE_NONE options. */
#define GC_OPT_FLAG_ARG_OPT (1UL << 1)
/* The LIST flag for an option indicates that the option can occur
several times. A comma separated list of arguments is used as the
argument value. */
#define GC_OPT_FLAG_LIST (1UL << 2)
gpgconf: Further simplify the gpgconf option processing. * common/gc-opt-flags.h (GC_OPT_FLAG_RUNTIME): Move to ... * tools/gpgconf-comp.c: here. (known_options_scdaemon): Remove "options". (known_options_dirmngr): Remove "options". (known_options_gpgsm): Remove "options". (known_options_gpg): Remove "options" and "keyserver". (struct gc_option_s): Rename active t gpgconf_list. (gc_component_list_options): Do not act upon active. (option_check_validity): Ditto. (is_known_option): Make it work correctly for unknown options. (retrieve_options_from_program): Use renamed flag gpgconf_list only to detect duplicated items from --gpgconf-list. Do not set runtime. Only e set the options if set by --gpgconf-list; never clear them. * agent/gpg-agent.c: Simplify the --gpgconf-list output. * dirmngr/dirmngr.c: Ditto. * g10/gpg.c: Ditto. * kbx/keyboxd.c: Ditto. * scd/scdaemon.c: Ditto. * sm/gpgsm.c: Ditto. * tests/openpgp/gpgconf.scm: Use "compliance" instead of "keyserver" for the string arg test. -- There is no need to read the list of options from the components unless they convey a default value. It is better to consult only the list we have in gpgconf-comp.c to decide on whether an option should be displayed. Right, this might mess up thing if a newer gpgconf version is used with an older component, but we already print warnings in this case and in general we do not want to support this anymore - the times of gpg 1.4. and 2.0 are long over now. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-14 18:53:43 +01:00
/* The RUNTIME flag for an option indicates that the option can be
changed at runtime. */
#define GC_OPT_FLAG_RUNTIME (1UL << 3)
/* A human-readable description for each flag. */
static const struct
{
const char *name;
} gc_flag[] =
{
{ "group" },
{ "optional arg" },
{ "list" },
{ "runtime" },
{ "default" },
{ "default desc" },
{ "no arg desc" },
{ "no change" }
};
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
/* Each option we want to support in gpgconf has the needed
* information in a static list per componenet. This struct describes
* the info for a single option. */
struct known_option_s
{
/* If this is NULL, then this is a terminator in an array of unknown
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
* length. Otherwise it is the name of the option described by this
* entry. The name must not contain a colon. */
const char *name;
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
/* The option flags. */
unsigned long flags;
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
/* The expert level. */
gc_expert_level_t level;
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
/* The complex type of the option argument; the default of 0 is used
* for a standard type as returned by --dump-option-table. */
gc_arg_type_t arg_type;
};
typedef struct known_option_s known_option_t;
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
/* The known options of the GC_COMPONENT_GPG_AGENT component. */
static known_option_t known_options_gpg_agent[] =
{
{ "verbose", GC_OPT_FLAG_LIST|GC_OPT_FLAG_RUNTIME, GC_LEVEL_BASIC },
{ "quiet", GC_OPT_FLAG_NONE|GC_OPT_FLAG_RUNTIME, GC_LEVEL_BASIC },
{ "disable-scdaemon", GC_OPT_FLAG_NONE, GC_LEVEL_ADVANCED },
{ "enable-ssh-support", GC_OPT_FLAG_NONE, GC_LEVEL_BASIC },
{ "ssh-fingerprint-digest", GC_OPT_FLAG_RUNTIME, GC_LEVEL_EXPERT },
{ "enable-putty-support", GC_OPT_FLAG_NONE, GC_LEVEL_BASIC },
{ "enable-extended-key-format", GC_OPT_FLAG_RUNTIME, GC_LEVEL_INVISIBLE },
{ "debug-level", GC_OPT_FLAG_ARG_OPT|GC_OPT_FLAG_RUNTIME, GC_LEVEL_ADVANCED},
{ "log-file", GC_OPT_FLAG_RUNTIME, GC_LEVEL_ADVANCED,
/**/ GC_ARG_TYPE_FILENAME },
{ "faked-system-time", GC_OPT_FLAG_NONE, GC_LEVEL_INVISIBLE },
{ "default-cache-ttl", GC_OPT_FLAG_RUNTIME, GC_LEVEL_BASIC },
{ "default-cache-ttl-ssh", GC_OPT_FLAG_RUNTIME, GC_LEVEL_ADVANCED },
{ "max-cache-ttl", GC_OPT_FLAG_RUNTIME, GC_LEVEL_EXPERT },
{ "max-cache-ttl-ssh", GC_OPT_FLAG_RUNTIME, GC_LEVEL_EXPERT },
{ "ignore-cache-for-signing", GC_OPT_FLAG_RUNTIME, GC_LEVEL_BASIC },
{ "allow-emacs-pinentry", GC_OPT_FLAG_RUNTIME, GC_LEVEL_ADVANCED },
{ "grab", GC_OPT_FLAG_RUNTIME, GC_LEVEL_EXPERT },
{ "no-allow-external-cache", GC_OPT_FLAG_RUNTIME, GC_LEVEL_BASIC },
{ "no-allow-mark-trusted", GC_OPT_FLAG_RUNTIME, GC_LEVEL_ADVANCED },
{ "no-allow-loopback-pinentry", GC_OPT_FLAG_RUNTIME, GC_LEVEL_EXPERT },
{ "enforce-passphrase-constraints", GC_OPT_FLAG_RUNTIME, GC_LEVEL_EXPERT },
{ "min-passphrase-len", GC_OPT_FLAG_RUNTIME, GC_LEVEL_ADVANCED },
{ "min-passphrase-nonalpha", GC_OPT_FLAG_RUNTIME, GC_LEVEL_EXPERT },
{ "check-passphrase-pattern", GC_OPT_FLAG_RUNTIME, GC_LEVEL_EXPERT,
/**/ GC_ARG_TYPE_FILENAME },
{ "max-passphrase-days", GC_OPT_FLAG_RUNTIME, GC_LEVEL_EXPERT },
{ "enable-passphrase-history", GC_OPT_FLAG_RUNTIME, GC_LEVEL_EXPERT },
{ "pinentry-timeout", GC_OPT_FLAG_RUNTIME, GC_LEVEL_ADVANCED },
{ NULL }
};
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
/* The known options of the GC_COMPONENT_SCDAEMON component. */
static known_option_t known_options_scdaemon[] =
{
{ "verbose", GC_OPT_FLAG_LIST|GC_OPT_FLAG_RUNTIME, GC_LEVEL_BASIC },
{ "quiet", GC_OPT_FLAG_NONE, GC_LEVEL_BASIC },
{ "no-greeting", GC_OPT_FLAG_NONE, GC_LEVEL_INVISIBLE },
{ "reader-port", GC_OPT_FLAG_RUNTIME, GC_LEVEL_BASIC },
{ "ctapi-driver", GC_OPT_FLAG_RUNTIME, GC_LEVEL_ADVANCED },
{ "pcsc-driver", GC_OPT_FLAG_RUNTIME, GC_LEVEL_ADVANCED },
{ "disable-ccid", GC_OPT_FLAG_RUNTIME, GC_LEVEL_EXPERT },
{ "disable-pinpad", GC_OPT_FLAG_NONE|GC_OPT_FLAG_RUNTIME, GC_LEVEL_BASIC },
{ "enable-pinpad-varlen", GC_OPT_FLAG_RUNTIME, GC_LEVEL_BASIC },
{ "card-timeout", GC_OPT_FLAG_RUNTIME, GC_LEVEL_BASIC },
{ "application-priority", GC_OPT_FLAG_RUNTIME, GC_LEVEL_ADVANCED },
{ "debug-level", GC_OPT_FLAG_ARG_OPT|GC_OPT_FLAG_RUNTIME, GC_LEVEL_ADVANCED},
{ "log-file", GC_OPT_FLAG_RUNTIME, GC_LEVEL_ADVANCED,
GC_ARG_TYPE_FILENAME },
{ "deny-admin", GC_OPT_FLAG_RUNTIME, GC_LEVEL_BASIC },
{ NULL }
};
tpm2d: Add tpm2daemon code * tpm2d: New directory. * Makefile.am (SUBDIRS): Add directory. * configure.ac: Detect libtss and decide whether to build tpm2d. * am/cmacros.am: Add a define. * util.h (GNUPG_MODULE_NAME_TPM2DAEMON): New. * common/homedir.c (gnupg_module_name): Add tpm2d. * common/mapstrings.c (macros): Add "TPM2DAEMON". * tools/gpgconf.h (GC_COMPONENT_TPM2DAEMON): New. * tools/gpgconf-comp.c (known_options_tpm2daemon): New. (gc_component): Add TPM2. (tpm2daemon_runtime_change): New. * tpm2d/Makefile.am: New. * tpm2d/command.c: New. * tpm2d/ibm-tss.h: New. * tpm2d/tpm2.c: New. * tpm2d/tpm2.h: New. * tpm2d/tpm2daemon.c: New. * tpm2d/tpm2daemon.h: New. --- This commit adds and plumbs in a tpm2daemon to the build to mirror the operation of scdaemon. The architecture of the code is that tpm2daemon.c itself is pretty much a clone of scd/scdaemon.c just with updated function prefixes (this argues there could be some further consolidation of the daemon handling code). Note that although this commit causes the daemon to be built and installed, nothing actually starts it or uses it yet. Command handling ---------------- command.c is copied from the command handler in scd.c except that the command implementation is now done in terms of tpm2 commands and the wire protocol is far simpler. The tpm2daemon only responds to 4 commands IMPORT: import a standard s-expression private key and export it to TPM2 format. This conversion cannot be undone and the private key now can *only* be used by the TPM2. To anyone who gets hold of the private key now, it's just an encrypted binary blob. PKSIGN: create a signature from the tpm2 key. The TPM2 form private key is retrieved by KEYDATA and the hash to be signed by EXTRA. Note there is no hash specifier because the tpm2 tss deduces the hash type from the length of the EXTRA data. This is actually a limitation of the tpm2 command API and it will be interesting to see how this fares if the tpm2 ever supports say sha3-256 hashes. PKDECRYPT: decrypt (RSA case) or derive (ECC case) a symmetric key. The tpm2 for private key is retrieved by KEYDATA and the information used to create the symmetric key by EXTRA. KILLTPM2D: stop the daemon All the tpm2 primitives used by command.c are in tpm2.h and all the tpm2 specific gunk is confined to tpm2.c, which is the only piece of this that actually does calls into the tss library. Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com> Changes from James' patch: - gpgconf: The displayed name is "TPM" and not "TPM2". That string is used by GUIs and should be something the user understands. For example we also use "network" instead of "Dirmngr". - Removed some commented includes. - Use 16 as emulation of GPG_ERR_SOURCE_TPM2. - Silenced a C90 compiler warning and flags unused parameters. - Removed "if HAVE_LIBS" from tpm2/Makefile.am and add missing files so that make distcheck works. Signed-off-by: Werner Koch <wk@gnupg.org>
2021-03-09 22:50:28 +01:00
/* The known options of the GC_COMPONENT_TPM2DAEMON component. */
static known_option_t known_options_tpm2daemon[] =
{
{ "verbose", GC_OPT_FLAG_LIST|GC_OPT_FLAG_RUNTIME, GC_LEVEL_BASIC },
{ "quiet", GC_OPT_FLAG_NONE, GC_LEVEL_BASIC },
{ "no-greeting", GC_OPT_FLAG_NONE, GC_LEVEL_INVISIBLE },
{ "debug-level", GC_OPT_FLAG_ARG_OPT|GC_OPT_FLAG_RUNTIME, GC_LEVEL_ADVANCED},
{ "log-file", GC_OPT_FLAG_RUNTIME, GC_LEVEL_ADVANCED,
GC_ARG_TYPE_FILENAME },
{ "deny-admin", GC_OPT_FLAG_RUNTIME, GC_LEVEL_BASIC },
{ "parent", GC_OPT_FLAG_RUNTIME, GC_LEVEL_ADVANCED },
{ NULL }
};
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
/* The known options of the GC_COMPONENT_GPG component. */
static known_option_t known_options_gpg[] =
{
{ "verbose", GC_OPT_FLAG_LIST, GC_LEVEL_BASIC },
{ "quiet", GC_OPT_FLAG_NONE, GC_LEVEL_BASIC },
{ "no-greeting", GC_OPT_FLAG_NONE, GC_LEVEL_INVISIBLE },
{ "default-key", GC_OPT_FLAG_NONE, GC_LEVEL_BASIC },
{ "encrypt-to", GC_OPT_FLAG_NONE, GC_LEVEL_BASIC },
{ "group", GC_OPT_FLAG_LIST, GC_LEVEL_ADVANCED,
GC_ARG_TYPE_ALIAS_LIST},
{ "compliance", GC_OPT_FLAG_NONE, GC_LEVEL_EXPERT },
{ "default-new-key-algo", GC_OPT_FLAG_NONE, GC_LEVEL_INVISIBLE },
{ "trust-model", GC_OPT_FLAG_NONE, GC_LEVEL_INVISIBLE },
{ "debug-level", GC_OPT_FLAG_ARG_OPT, GC_LEVEL_ADVANCED },
{ "log-file", GC_OPT_FLAG_NONE, GC_LEVEL_ADVANCED,
GC_ARG_TYPE_FILENAME },
{ "auto-key-locate", GC_OPT_FLAG_NONE, GC_LEVEL_ADVANCED },
{ "auto-key-import", GC_OPT_FLAG_NONE, GC_LEVEL_BASIC },
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
{ "auto-key-retrieve", GC_OPT_FLAG_NONE, GC_LEVEL_EXPERT },
{ "include-key-block", GC_OPT_FLAG_NONE, GC_LEVEL_BASIC },
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
{ "disable-dirmngr", GC_OPT_FLAG_NONE, GC_LEVEL_EXPERT },
{ "max-cert-depth", GC_OPT_FLAG_NONE, GC_LEVEL_INVISIBLE },
{ "completes-needed", GC_OPT_FLAG_NONE, GC_LEVEL_INVISIBLE },
{ "marginals-needed", GC_OPT_FLAG_NONE, GC_LEVEL_INVISIBLE },
/* The next is a pseudo option which we read via --gpgconf-list */
{ "default_pubkey_algo",
(GC_OPT_FLAG_ARG_OPT|GC_OPT_FLAG_NO_CHANGE), GC_LEVEL_INVISIBLE },
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
{ NULL }
};
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
/* The known options of the GC_COMPONENT_GPGSM component. */
static known_option_t known_options_gpgsm[] =
{
{ "verbose", GC_OPT_FLAG_LIST, GC_LEVEL_BASIC },
{ "quiet", GC_OPT_FLAG_NONE, GC_LEVEL_BASIC },
{ "no-greeting", GC_OPT_FLAG_NONE, GC_LEVEL_INVISIBLE },
{ "default-key", GC_OPT_FLAG_NONE, GC_LEVEL_BASIC },
{ "encrypt-to", GC_OPT_FLAG_NONE, GC_LEVEL_BASIC },
{ "disable-dirmngr", GC_OPT_FLAG_NONE, GC_LEVEL_EXPERT },
{ "p12-charset", GC_OPT_FLAG_NONE, GC_LEVEL_ADVANCED },
{ "keyserver", GC_OPT_FLAG_LIST, GC_LEVEL_BASIC,
GC_ARG_TYPE_LDAP_SERVER },
{ "compliance", GC_OPT_FLAG_NONE, GC_LEVEL_EXPERT },
{ "debug-level", GC_OPT_FLAG_ARG_OPT, GC_LEVEL_ADVANCED },
{ "log-file", GC_OPT_FLAG_NONE, GC_LEVEL_ADVANCED,
GC_ARG_TYPE_FILENAME },
{ "faked-system-time", GC_OPT_FLAG_NONE, GC_LEVEL_INVISIBLE },
{ "disable-crl-checks", GC_OPT_FLAG_NONE, GC_LEVEL_BASIC },
{ "enable-crl-checks", GC_OPT_FLAG_NONE, GC_LEVEL_INVISIBLE },
{ "enable-ocsp", GC_OPT_FLAG_NONE, GC_LEVEL_ADVANCED },
{ "include-certs", GC_OPT_FLAG_NONE, GC_LEVEL_EXPERT },
{ "disable-policy-checks", GC_OPT_FLAG_NONE, GC_LEVEL_ADVANCED },
{ "auto-issuer-key-retrieve", GC_OPT_FLAG_NONE, GC_LEVEL_BASIC },
{ "cipher-algo", GC_OPT_FLAG_NONE, GC_LEVEL_ADVANCED },
{ "disable-trusted-cert-crl-check", GC_OPT_FLAG_NONE, GC_LEVEL_EXPERT },
/* Pseudo option follows. */
{ "default_pubkey_algo",
(GC_OPT_FLAG_ARG_OPT|GC_OPT_FLAG_NO_CHANGE), GC_LEVEL_INVISIBLE },
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
{ NULL }
};
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
/* The known options of the GC_COMPONENT_DIRMNGR component. */
static known_option_t known_options_dirmngr[] =
{
{ "verbose", GC_OPT_FLAG_LIST, GC_LEVEL_BASIC },
{ "quiet", GC_OPT_FLAG_NONE, GC_LEVEL_BASIC },
{ "no-greeting", GC_OPT_FLAG_NONE, GC_LEVEL_INVISIBLE },
{ "resolver-timeout", GC_OPT_FLAG_NONE, GC_LEVEL_INVISIBLE },
{ "nameserver", GC_OPT_FLAG_NONE, GC_LEVEL_INVISIBLE },
{ "debug-level", GC_OPT_FLAG_ARG_OPT, GC_LEVEL_ADVANCED },
{ "log-file", GC_OPT_FLAG_NONE, GC_LEVEL_ADVANCED,
GC_ARG_TYPE_FILENAME },
{ "faked-system-time", GC_OPT_FLAG_NONE, GC_LEVEL_INVISIBLE },
{ "batch", GC_OPT_FLAG_NONE, GC_LEVEL_BASIC },
{ "force", GC_OPT_FLAG_NONE, GC_LEVEL_BASIC },
{ "use-tor", GC_OPT_FLAG_NONE, GC_LEVEL_BASIC },
{ "keyserver", GC_OPT_FLAG_NONE, GC_LEVEL_BASIC },
{ "disable-http", GC_OPT_FLAG_NONE, GC_LEVEL_ADVANCED },
{ "ignore-http-dp", GC_OPT_FLAG_NONE, GC_LEVEL_ADVANCED },
{ "http-proxy", GC_OPT_FLAG_NONE, GC_LEVEL_ADVANCED },
{ "honor-http-proxy", GC_OPT_FLAG_NONE, GC_LEVEL_ADVANCED },
{ "disable-ldap", GC_OPT_FLAG_NONE, GC_LEVEL_ADVANCED },
{ "ignore-ldap-dp", GC_OPT_FLAG_NONE, GC_LEVEL_ADVANCED },
{ "ldap-proxy", GC_OPT_FLAG_NONE, GC_LEVEL_BASIC },
{ "only-ldap-proxy", GC_OPT_FLAG_NONE, GC_LEVEL_ADVANCED },
{ "add-servers", GC_OPT_FLAG_NONE, GC_LEVEL_EXPERT },
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
{ "ldaptimeout", GC_OPT_FLAG_NONE, GC_LEVEL_BASIC },
{ "max-replies", GC_OPT_FLAG_NONE, GC_LEVEL_BASIC },
{ "allow-ocsp", GC_OPT_FLAG_NONE, GC_LEVEL_BASIC },
{ "ocsp-responder", GC_OPT_FLAG_NONE, GC_LEVEL_ADVANCED },
{ "ocsp-signer", GC_OPT_FLAG_NONE, GC_LEVEL_ADVANCED },
{ "allow-version-check", GC_OPT_FLAG_NONE, GC_LEVEL_BASIC },
{ "ignore-ocsp-service-url", GC_OPT_FLAG_NONE, GC_LEVEL_ADVANCED },
{ NULL }
};
/* The known options of the GC_COMPONENT_KEYBOXD component. */
static known_option_t known_options_keyboxd[] =
{
{ "verbose", GC_OPT_FLAG_LIST, GC_LEVEL_BASIC },
{ "quiet", GC_OPT_FLAG_NONE, GC_LEVEL_BASIC },
{ "log-file", GC_OPT_FLAG_NONE, GC_LEVEL_ADVANCED,
GC_ARG_TYPE_FILENAME },
{ "faked-system-time", GC_OPT_FLAG_NONE, GC_LEVEL_INVISIBLE },
{ NULL }
};
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
/* The known options of the GC_COMPONENT_PINENTRY component. */
static known_option_t known_options_pinentry[] =
{
{ NULL }
};
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
/* Our main option info object. We copy all required information from the
* gpgrt_opt_t items but convert the flags value to bit flags. */
struct gc_option_s
{
const char *name; /* The same as gpgrt_opt_t.long_opt. */
const char *desc; /* The same as gpgrt_opt_t.description. */
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
unsigned int is_header:1; /* This is a header item. */
unsigned int is_list:1; /* This is a list style option. */
unsigned int opt_arg:1; /* The option's argument is optional. */
unsigned int runtime:1; /* The option is runtime changeable. */
gpgconf: Further simplify the gpgconf option processing. * common/gc-opt-flags.h (GC_OPT_FLAG_RUNTIME): Move to ... * tools/gpgconf-comp.c: here. (known_options_scdaemon): Remove "options". (known_options_dirmngr): Remove "options". (known_options_gpgsm): Remove "options". (known_options_gpg): Remove "options" and "keyserver". (struct gc_option_s): Rename active t gpgconf_list. (gc_component_list_options): Do not act upon active. (option_check_validity): Ditto. (is_known_option): Make it work correctly for unknown options. (retrieve_options_from_program): Use renamed flag gpgconf_list only to detect duplicated items from --gpgconf-list. Do not set runtime. Only e set the options if set by --gpgconf-list; never clear them. * agent/gpg-agent.c: Simplify the --gpgconf-list output. * dirmngr/dirmngr.c: Ditto. * g10/gpg.c: Ditto. * kbx/keyboxd.c: Ditto. * scd/scdaemon.c: Ditto. * sm/gpgsm.c: Ditto. * tests/openpgp/gpgconf.scm: Use "compliance" instead of "keyserver" for the string arg test. -- There is no need to read the list of options from the components unless they convey a default value. It is better to consult only the list we have in gpgconf-comp.c to decide on whether an option should be displayed. Right, this might mess up thing if a newer gpgconf version is used with an older component, but we already print warnings in this case and in general we do not want to support this anymore - the times of gpg 1.4. and 2.0 are long over now. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-14 18:53:43 +01:00
unsigned int gpgconf_list:1; /* Mentioned by --gpgconf-list. */
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
unsigned int has_default:1; /* The option has a default value. */
unsigned int def_in_desc:1; /* The default is in the descrition. */
unsigned int no_arg_desc:1; /* The argument has a default ???. */
unsigned int no_change:1; /* User shall not change the option. */
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
unsigned int attr_ignore:1; /* The ARGPARSE_ATTR_IGNORE. */
unsigned int attr_force:1; /* The ARGPARSE_ATTR_FORCE. */
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
/* The expert level - copied from known_options. */
gc_expert_level_t level;
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
/* The complex type - copied from known_options. */
gc_arg_type_t arg_type;
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
/* The default value for this option. This is NULL if the option is
not present in the component, the empty string if no default is
available, and otherwise a quoted string. This is currently
malloced.*/
char *default_value;
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
/* The current value of this option. */
char *value;
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
/* The new flags for this option. The only defined flag is actually
GC_OPT_FLAG_DEFAULT, and it means that the option should be
deleted. In this case, NEW_VALUE is NULL. */
unsigned long new_flags;
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
/* The new value of this option. */
char *new_value;
};
typedef struct gc_option_s gc_option_t;
/* The information associated with each component. */
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
static struct
{
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
/* The name of the component. Some components don't have an
* associated program, but are implemented directly by GPGConf. In
* this case, PROGRAM is NULL. */
char *program;
/* The displayed name of this component. Must not contain a colon
* (':') character. */
const char *name;
/* The gettext domain for the description DESC. If this is NULL,
then the description is not translated. */
const char *desc_domain;
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
/* The description of this component. */
const char *desc;
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
/* The module name (GNUPG_MODULE_NAME_foo) as defined by
* ../common/util.h. This value is used to get the actual installed
* path of the program. 0 is used if no program for the component
* is available. */
char module_name;
/* The name for the configuration filename of this component. */
const char *option_config_filename;
/* The static table of known options for this component. */
known_option_t *known_options;
/* The runtime change callback. If KILLFLAG is true the component
is killed and not just reloaded. */
void (*runtime_change) (int killflag);
/* The table of known options as read from the component including
* header lines and such. This is suitable to be passed to
* gpgrt_argparser. Will be filled in by
* retrieve_options_from_program. */
gpgrt_opt_t *opt_table;
/* The full table including data from OPT_TABLE. The end of the
* table is marked by NULL entry for NAME. Will be filled in by
* retrieve_options_from_program. */
gc_option_t *options;
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
} gc_component[GC_COMPONENT_NR] =
{
/* Note: The order of the items must match the order given in the
* gc_component_id_t enumeration. The order is often used by
* frontends to display the backend options thus do not change the
* order without considering the user experience. */
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
{ NULL }, /* DUMMY for GC_COMPONENT_ANY */
{ GPG_NAME, GPG_DISP_NAME, "gnupg", N_("OpenPGP"),
GNUPG_MODULE_NAME_GPG, GPG_NAME ".conf",
known_options_gpg },
{ GPGSM_NAME, GPGSM_DISP_NAME, "gnupg", N_("S/MIME"),
GNUPG_MODULE_NAME_GPGSM, GPGSM_NAME ".conf",
known_options_gpgsm },
{ KEYBOXD_NAME, KEYBOXD_DISP_NAME, "gnupg", N_("Public Keys"),
GNUPG_MODULE_NAME_KEYBOXD, KEYBOXD_NAME ".conf",
known_options_keyboxd, keyboxd_runtime_change },
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
{ GPG_AGENT_NAME, GPG_AGENT_DISP_NAME, "gnupg", N_("Private Keys"),
GNUPG_MODULE_NAME_AGENT, GPG_AGENT_NAME ".conf",
known_options_gpg_agent, gpg_agent_runtime_change },
{ SCDAEMON_NAME, SCDAEMON_DISP_NAME, "gnupg", N_("Smartcards"),
GNUPG_MODULE_NAME_SCDAEMON, SCDAEMON_NAME ".conf",
known_options_scdaemon, scdaemon_runtime_change},
tpm2d: Add tpm2daemon code * tpm2d: New directory. * Makefile.am (SUBDIRS): Add directory. * configure.ac: Detect libtss and decide whether to build tpm2d. * am/cmacros.am: Add a define. * util.h (GNUPG_MODULE_NAME_TPM2DAEMON): New. * common/homedir.c (gnupg_module_name): Add tpm2d. * common/mapstrings.c (macros): Add "TPM2DAEMON". * tools/gpgconf.h (GC_COMPONENT_TPM2DAEMON): New. * tools/gpgconf-comp.c (known_options_tpm2daemon): New. (gc_component): Add TPM2. (tpm2daemon_runtime_change): New. * tpm2d/Makefile.am: New. * tpm2d/command.c: New. * tpm2d/ibm-tss.h: New. * tpm2d/tpm2.c: New. * tpm2d/tpm2.h: New. * tpm2d/tpm2daemon.c: New. * tpm2d/tpm2daemon.h: New. --- This commit adds and plumbs in a tpm2daemon to the build to mirror the operation of scdaemon. The architecture of the code is that tpm2daemon.c itself is pretty much a clone of scd/scdaemon.c just with updated function prefixes (this argues there could be some further consolidation of the daemon handling code). Note that although this commit causes the daemon to be built and installed, nothing actually starts it or uses it yet. Command handling ---------------- command.c is copied from the command handler in scd.c except that the command implementation is now done in terms of tpm2 commands and the wire protocol is far simpler. The tpm2daemon only responds to 4 commands IMPORT: import a standard s-expression private key and export it to TPM2 format. This conversion cannot be undone and the private key now can *only* be used by the TPM2. To anyone who gets hold of the private key now, it's just an encrypted binary blob. PKSIGN: create a signature from the tpm2 key. The TPM2 form private key is retrieved by KEYDATA and the hash to be signed by EXTRA. Note there is no hash specifier because the tpm2 tss deduces the hash type from the length of the EXTRA data. This is actually a limitation of the tpm2 command API and it will be interesting to see how this fares if the tpm2 ever supports say sha3-256 hashes. PKDECRYPT: decrypt (RSA case) or derive (ECC case) a symmetric key. The tpm2 for private key is retrieved by KEYDATA and the information used to create the symmetric key by EXTRA. KILLTPM2D: stop the daemon All the tpm2 primitives used by command.c are in tpm2.h and all the tpm2 specific gunk is confined to tpm2.c, which is the only piece of this that actually does calls into the tss library. Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com> Changes from James' patch: - gpgconf: The displayed name is "TPM" and not "TPM2". That string is used by GUIs and should be something the user understands. For example we also use "network" instead of "Dirmngr". - Removed some commented includes. - Use 16 as emulation of GPG_ERR_SOURCE_TPM2. - Silenced a C90 compiler warning and flags unused parameters. - Removed "if HAVE_LIBS" from tpm2/Makefile.am and add missing files so that make distcheck works. Signed-off-by: Werner Koch <wk@gnupg.org>
2021-03-09 22:50:28 +01:00
{ TPM2DAEMON_NAME, TPM2DAEMON_DISP_NAME, "gnupg", N_("TPM"),
GNUPG_MODULE_NAME_TPM2DAEMON, TPM2DAEMON_NAME ".conf",
known_options_tpm2daemon, tpm2daemon_runtime_change},
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
{ DIRMNGR_NAME, DIRMNGR_DISP_NAME, "gnupg", N_("Network"),
GNUPG_MODULE_NAME_DIRMNGR, DIRMNGR_NAME ".conf",
known_options_dirmngr, dirmngr_runtime_change },
{ "pinentry", "Pinentry", "gnupg", N_("Passphrase Entry"),
GNUPG_MODULE_NAME_PINENTRY, NULL,
known_options_pinentry }
};
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
/* Structure used to collect error output of the component programs. */
struct error_line_s;
typedef struct error_line_s *error_line_t;
struct error_line_s
{
error_line_t next; /* Link to next item. */
const char *fname; /* Name of the config file (points into BUFFER). */
unsigned int lineno; /* Line number of the config file. */
const char *errtext; /* Text of the error message (points into BUFFER). */
char buffer[1]; /* Helper buffer. */
};
/* Initialization and finalization. */
static void
gc_option_free (gc_option_t *o)
{
if (o == NULL || o->name == NULL)
return;
xfree (o->value);
gc_option_free (o + 1);
}
static void
gc_components_free (void)
{
int i;
for (i = 0; i < DIM (gc_component); i++)
gc_option_free (gc_component[i].options);
}
void
gc_components_init (void)
{
atexit (gc_components_free);
}
/* Engine specific support. */
static void
gpg_agent_runtime_change (int killflag)
{
gpg_error_t err = 0;
const char *pgmname;
const char *argv[5];
pid_t pid = (pid_t)(-1);
int i = 0;
pgmname = gnupg_module_name (GNUPG_MODULE_NAME_CONNECT_AGENT);
if (!gnupg_default_homedir_p ())
{
argv[i++] = "--homedir";
argv[i++] = gnupg_homedir ();
}
argv[i++] = "--no-autostart";
argv[i++] = killflag? "KILLAGENT" : "RELOADAGENT";
argv[i++] = NULL;
if (!err)
err = gnupg_spawn_process_fd (pgmname, argv, -1, -1, -1, &pid);
if (!err)
2010-06-24 12:51:30 +02:00
err = gnupg_wait_process (pgmname, pid, 1, NULL);
if (err)
gc_error (0, 0, "error running '%s %s': %s",
pgmname, argv[1], gpg_strerror (err));
gnupg_release_process (pid);
}
static void
scdaemon_runtime_change (int killflag)
{
gpg_error_t err = 0;
const char *pgmname;
const char *argv[9];
pid_t pid = (pid_t)(-1);
int i = 0;
(void)killflag; /* For scdaemon kill and reload are synonyms. */
/* We use "GETINFO app_running" to see whether the agent is already
running and kill it only in this case. This avoids an explicit
starting of the agent in case it is not yet running. There is
obviously a race condition but that should not harm too much. */
pgmname = gnupg_module_name (GNUPG_MODULE_NAME_CONNECT_AGENT);
if (!gnupg_default_homedir_p ())
{
argv[i++] = "--homedir";
argv[i++] = gnupg_homedir ();
}
argv[i++] = "-s";
argv[i++] = "--no-autostart";
argv[i++] = "GETINFO scd_running";
argv[i++] = "/if ${! $?}";
argv[i++] = "scd killscd";
argv[i++] = "/end";
argv[i++] = NULL;
if (!err)
err = gnupg_spawn_process_fd (pgmname, argv, -1, -1, -1, &pid);
if (!err)
err = gnupg_wait_process (pgmname, pid, 1, NULL);
if (err)
gc_error (0, 0, "error running '%s %s': %s",
pgmname, argv[4], gpg_strerror (err));
gnupg_release_process (pid);
}
tpm2d: Add tpm2daemon code * tpm2d: New directory. * Makefile.am (SUBDIRS): Add directory. * configure.ac: Detect libtss and decide whether to build tpm2d. * am/cmacros.am: Add a define. * util.h (GNUPG_MODULE_NAME_TPM2DAEMON): New. * common/homedir.c (gnupg_module_name): Add tpm2d. * common/mapstrings.c (macros): Add "TPM2DAEMON". * tools/gpgconf.h (GC_COMPONENT_TPM2DAEMON): New. * tools/gpgconf-comp.c (known_options_tpm2daemon): New. (gc_component): Add TPM2. (tpm2daemon_runtime_change): New. * tpm2d/Makefile.am: New. * tpm2d/command.c: New. * tpm2d/ibm-tss.h: New. * tpm2d/tpm2.c: New. * tpm2d/tpm2.h: New. * tpm2d/tpm2daemon.c: New. * tpm2d/tpm2daemon.h: New. --- This commit adds and plumbs in a tpm2daemon to the build to mirror the operation of scdaemon. The architecture of the code is that tpm2daemon.c itself is pretty much a clone of scd/scdaemon.c just with updated function prefixes (this argues there could be some further consolidation of the daemon handling code). Note that although this commit causes the daemon to be built and installed, nothing actually starts it or uses it yet. Command handling ---------------- command.c is copied from the command handler in scd.c except that the command implementation is now done in terms of tpm2 commands and the wire protocol is far simpler. The tpm2daemon only responds to 4 commands IMPORT: import a standard s-expression private key and export it to TPM2 format. This conversion cannot be undone and the private key now can *only* be used by the TPM2. To anyone who gets hold of the private key now, it's just an encrypted binary blob. PKSIGN: create a signature from the tpm2 key. The TPM2 form private key is retrieved by KEYDATA and the hash to be signed by EXTRA. Note there is no hash specifier because the tpm2 tss deduces the hash type from the length of the EXTRA data. This is actually a limitation of the tpm2 command API and it will be interesting to see how this fares if the tpm2 ever supports say sha3-256 hashes. PKDECRYPT: decrypt (RSA case) or derive (ECC case) a symmetric key. The tpm2 for private key is retrieved by KEYDATA and the information used to create the symmetric key by EXTRA. KILLTPM2D: stop the daemon All the tpm2 primitives used by command.c are in tpm2.h and all the tpm2 specific gunk is confined to tpm2.c, which is the only piece of this that actually does calls into the tss library. Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com> Changes from James' patch: - gpgconf: The displayed name is "TPM" and not "TPM2". That string is used by GUIs and should be something the user understands. For example we also use "network" instead of "Dirmngr". - Removed some commented includes. - Use 16 as emulation of GPG_ERR_SOURCE_TPM2. - Silenced a C90 compiler warning and flags unused parameters. - Removed "if HAVE_LIBS" from tpm2/Makefile.am and add missing files so that make distcheck works. Signed-off-by: Werner Koch <wk@gnupg.org>
2021-03-09 22:50:28 +01:00
static void
tpm2daemon_runtime_change (int killflag)
{
gpg_error_t err = 0;
const char *pgmname;
const char *argv[9];
pid_t pid = (pid_t)(-1);
int i = 0;
(void)killflag; /* For scdaemon kill and reload are synonyms. */
/* We use "GETINFO app_running" to see whether the agent is already
running and kill it only in this case. This avoids an explicit
starting of the agent in case it is not yet running. There is
obviously a race condition but that should not harm too much. */
pgmname = gnupg_module_name (GNUPG_MODULE_NAME_CONNECT_AGENT);
if (!gnupg_default_homedir_p ())
{
argv[i++] = "--homedir";
argv[i++] = gnupg_homedir ();
}
argv[i++] = "-s";
argv[i++] = "--no-autostart";
argv[i++] = "GETINFO tpm2d_running";
argv[i++] = "/if ${! $?}";
argv[i++] = "scd killtpm2cd";
argv[i++] = "/end";
argv[i++] = NULL;
if (!err)
err = gnupg_spawn_process_fd (pgmname, argv, -1, -1, -1, &pid);
if (!err)
err = gnupg_wait_process (pgmname, pid, 1, NULL);
if (err)
gc_error (0, 0, "error running '%s %s': %s",
pgmname, argv[4], gpg_strerror (err));
gnupg_release_process (pid);
}
static void
dirmngr_runtime_change (int killflag)
{
gpg_error_t err = 0;
const char *pgmname;
const char *argv[6];
pid_t pid = (pid_t)(-1);
pgmname = gnupg_module_name (GNUPG_MODULE_NAME_CONNECT_AGENT);
argv[0] = "--no-autostart";
argv[1] = "--dirmngr";
argv[2] = killflag? "KILLDIRMNGR" : "RELOADDIRMNGR";
if (gnupg_default_homedir_p ())
argv[3] = NULL;
else
{
argv[3] = "--homedir";
argv[4] = gnupg_homedir ();
argv[5] = NULL;
}
if (!err)
err = gnupg_spawn_process_fd (pgmname, argv, -1, -1, -1, &pid);
if (!err)
2010-06-24 12:51:30 +02:00
err = gnupg_wait_process (pgmname, pid, 1, NULL);
if (err)
gc_error (0, 0, "error running '%s %s': %s",
pgmname, argv[2], gpg_strerror (err));
gnupg_release_process (pid);
}
static void
keyboxd_runtime_change (int killflag)
{
gpg_error_t err = 0;
const char *pgmname;
const char *argv[6];
pid_t pid = (pid_t)(-1);
pgmname = gnupg_module_name (GNUPG_MODULE_NAME_CONNECT_AGENT);
argv[0] = "--no-autostart";
argv[1] = "--keyboxd";
argv[2] = killflag? "KILLKEYBOXD" : "RELOADKEYBOXD";
if (gnupg_default_homedir_p ())
argv[3] = NULL;
else
{
argv[3] = "--homedir";
argv[4] = gnupg_homedir ();
argv[5] = NULL;
}
if (!err)
err = gnupg_spawn_process_fd (pgmname, argv, -1, -1, -1, &pid);
if (!err)
err = gnupg_wait_process (pgmname, pid, 1, NULL);
if (err)
gc_error (0, 0, "error running '%s %s': %s",
pgmname, argv[2], gpg_strerror (err));
gnupg_release_process (pid);
}
/* Launch the gpg-agent or the dirmngr if not already running. */
gpg_error_t
gc_component_launch (int component)
{
gpg_error_t err;
const char *pgmname;
const char *argv[5];
int i;
pid_t pid;
if (component < 0)
{
err = gc_component_launch (GC_COMPONENT_GPG_AGENT);
if (!err)
err = gc_component_launch (GC_COMPONENT_KEYBOXD);
if (!err)
err = gc_component_launch (GC_COMPONENT_DIRMNGR);
return err;
}
if (!(component == GC_COMPONENT_GPG_AGENT
|| component == GC_COMPONENT_KEYBOXD
|| component == GC_COMPONENT_DIRMNGR))
{
log_error ("%s\n", _("Component not suitable for launching"));
gpgconf_failure (0);
}
if (gc_component_check_options (component, NULL, NULL))
{
log_error (_("Configuration file of component %s is broken\n"),
gc_component[component].name);
if (!opt.quiet)
log_info (_("Note: Use the command \"%s%s\" to get details.\n"),
gc_component[component].name, " --gpgconf-test");
gpgconf_failure (0);
}
pgmname = gnupg_module_name (GNUPG_MODULE_NAME_CONNECT_AGENT);
i = 0;
if (!gnupg_default_homedir_p ())
{
argv[i++] = "--homedir";
argv[i++] = gnupg_homedir ();
}
if (component == GC_COMPONENT_DIRMNGR)
argv[i++] = "--dirmngr";
else if (component == GC_COMPONENT_KEYBOXD)
argv[i++] = "--keyboxd";
argv[i++] = "NOP";
argv[i] = NULL;
err = gnupg_spawn_process_fd (pgmname, argv, -1, -1, -1, &pid);
if (!err)
err = gnupg_wait_process (pgmname, pid, 1, NULL);
if (err)
gc_error (0, 0, "error running '%s%s%s': %s",
pgmname,
component == GC_COMPONENT_DIRMNGR? " --dirmngr"
: component == GC_COMPONENT_KEYBOXD? " --keyboxd":"",
" NOP",
gpg_strerror (err));
gnupg_release_process (pid);
return err;
}
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
static void
do_runtime_change (int component, int killflag)
{
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
int runtime[GC_COMPONENT_NR] = { 0 };
if (component < 0)
{
for (component = 0; component < GC_COMPONENT_NR; component++)
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
runtime [component] = 1;
}
else
{
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
log_assert (component >= 0 && component < GC_COMPONENT_NR);
runtime [component] = 1;
}
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
/* Do the restart for the selected components. */
for (component = GC_COMPONENT_NR-1; component >= 0; component--)
{
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
if (runtime[component] && gc_component[component].runtime_change)
(*gc_component[component].runtime_change) (killflag);
}
}
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
/* Unconditionally restart COMPONENT. */
void
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
gc_component_kill (int component)
{
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
do_runtime_change (component, 1);
}
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
/* Unconditionally reload COMPONENT or all components if COMPONENT is -1. */
void
gc_component_reload (int component)
{
do_runtime_change (component, 0);
}
/* More or less Robust version of dgettext. It has the side effect of
switching the codeset to utf-8 because this is what we want to
output. In theory it is possible to keep the original code set and
switch back for regular diagnostic output (redefine "_(" for that)
but given the nature of this tool, being something invoked from
other programs, it does not make much sense. */
static const char *
my_dgettext (const char *domain, const char *msgid)
{
#ifdef USE_SIMPLE_GETTEXT
if (domain)
{
static int switched_codeset;
char *text;
if (!switched_codeset)
{
switched_codeset = 1;
2010-03-15 14:08:51 +01:00
gettext_use_utf8 (1);
}
if (!strcmp (domain, "gnupg"))
domain = PACKAGE_GT;
/* FIXME: we have no dgettext, thus we can't switch. */
text = (char*)gettext (msgid);
return text ? text : msgid;
}
else
return msgid;
#elif defined(ENABLE_NLS)
if (domain)
{
static int switched_codeset;
char *text;
if (!switched_codeset)
{
switched_codeset = 1;
bind_textdomain_codeset (PACKAGE_GT, "utf-8");
bindtextdomain (DIRMNGR_NAME, LOCALEDIR);
bind_textdomain_codeset (DIRMNGR_NAME, "utf-8");
}
/* Note: This is a hack to actually use the gnupg2 domain as
long we are in a transition phase where gnupg 1.x and 1.9 may
coexist. */
if (!strcmp (domain, "gnupg"))
domain = PACKAGE_GT;
text = dgettext (domain, msgid);
return text ? text : msgid;
}
else
return msgid;
#else
(void)domain;
return msgid;
#endif
}
/* Percent-Escape special characters. The string is valid until the
next invocation of the function. */
char *
gc_percent_escape (const char *src)
{
static char *esc_str;
static int esc_str_len;
int new_len = 3 * strlen (src) + 1;
char *dst;
if (esc_str_len < new_len)
{
char *new_esc_str = xrealloc (esc_str, new_len);
esc_str = new_esc_str;
esc_str_len = new_len;
}
dst = esc_str;
while (*src)
{
if (*src == '%')
{
*(dst++) = '%';
*(dst++) = '2';
*(dst++) = '5';
}
else if (*src == ':')
{
/* The colon is used as field separator. */
*(dst++) = '%';
*(dst++) = '3';
*(dst++) = 'a';
}
else if (*src == ',')
{
/* The comma is used as list separator. */
*(dst++) = '%';
*(dst++) = '2';
*(dst++) = 'c';
}
else if (*src == '\n')
{
/* The newline is problematic in a line-based format. */
*(dst++) = '%';
*(dst++) = '0';
*(dst++) = 'a';
}
else
*(dst++) = *(src);
src++;
}
*dst = '\0';
return esc_str;
}
/* Percent-Deescape special characters. The string is valid until the
next invocation of the function. */
static char *
percent_deescape (const char *src)
{
static char *str;
static int str_len;
int new_len = 3 * strlen (src) + 1;
char *dst;
if (str_len < new_len)
{
char *new_str = xrealloc (str, new_len);
str = new_str;
str_len = new_len;
}
dst = str;
while (*src)
{
if (*src == '%')
{
int val = hextobyte (src + 1);
if (val < 0)
gc_error (1, 0, "malformed end of string %s", src);
*(dst++) = (char) val;
src += 3;
}
else
*(dst++) = *(src++);
}
*dst = '\0';
return str;
}
/* List all components that are available. */
void
gc_component_list_components (estream_t out)
{
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
gc_component_id_t component;
const char *desc;
const char *pgmname;
for (component = 0; component < GC_COMPONENT_NR; component++)
{
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
if (!gc_component[component].program)
continue;
if (gc_component[component].module_name)
pgmname = gnupg_module_name (gc_component[component].module_name);
else
pgmname = "";
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
desc = gc_component[component].desc;
desc = my_dgettext (gc_component[component].desc_domain, desc);
es_fprintf (out, "%s:%s:",
gc_component[component].program, gc_percent_escape (desc));
es_fprintf (out, "%s\n", gc_percent_escape (pgmname));
}
}
static int
all_digits_p (const char *p, size_t len)
{
if (!len)
return 0; /* No. */
for (; len; len--, p++)
if (!isascii (*p) || !isdigit (*p))
return 0; /* No. */
return 1; /* Yes. */
}
/* Collect all error lines from stream FP. Only lines prefixed with
TAG are considered. Returns a list of error line items (which may
be empty). There is no error return. */
static error_line_t
collect_error_output (estream_t fp, const char *tag)
{
char buffer[1024];
char *p, *p2, *p3;
int c, cont_line;
unsigned int pos;
error_line_t eitem, errlines, *errlines_tail;
size_t taglen = strlen (tag);
errlines = NULL;
errlines_tail = &errlines;
pos = 0;
cont_line = 0;
while ((c=es_getc (fp)) != EOF)
{
buffer[pos++] = c;
if (pos >= sizeof buffer - 5 || c == '\n')
{
buffer[pos - (c == '\n')] = 0;
if (cont_line)
; /*Ignore continuations of previous line. */
else if (!strncmp (buffer, tag, taglen) && buffer[taglen] == ':')
{
/* "gpgsm: foo:4: bla" */
/* Yep, we are interested in this line. */
p = buffer + taglen + 1;
while (*p == ' ' || *p == '\t')
p++;
trim_trailing_spaces (p); /* Get rid of extra CRs. */
if (!*p)
; /* Empty lines are ignored. */
else if ( (p2 = strchr (p, ':')) && (p3 = strchr (p2+1, ':'))
&& all_digits_p (p2+1, p3 - (p2+1)))
{
/* Line in standard compiler format. */
p3++;
while (*p3 == ' ' || *p3 == '\t')
p3++;
eitem = xmalloc (sizeof *eitem + strlen (p));
eitem->next = NULL;
strcpy (eitem->buffer, p);
eitem->fname = eitem->buffer;
eitem->buffer[p2-p] = 0;
eitem->errtext = eitem->buffer + (p3 - p);
/* (we already checked that there are only ascii
digits followed by a colon) */
eitem->lineno = 0;
for (p2++; isdigit (*p2); p2++)
eitem->lineno = eitem->lineno*10 + (*p2 - '0');
*errlines_tail = eitem;
errlines_tail = &eitem->next;
}
else
{
/* Other error output. */
eitem = xmalloc (sizeof *eitem + strlen (p));
eitem->next = NULL;
strcpy (eitem->buffer, p);
eitem->fname = NULL;
eitem->errtext = eitem->buffer;
eitem->lineno = 0;
*errlines_tail = eitem;
errlines_tail = &eitem->next;
}
}
pos = 0;
/* If this was not a complete line mark that we are in a
continuation. */
cont_line = (c != '\n');
}
}
/* We ignore error lines not terminated by a LF. */
return errlines;
}
/* Check the options of a single component. If CONF_FILE is NULL the
* standard config file is used. If OUT is not NULL the output is
* written to that stream. Returns 0 if everything is OK. */
int
gc_component_check_options (int component, estream_t out, const char *conf_file)
{
gpg_error_t err;
unsigned int result;
const char *pgmname;
const char *argv[5];
int i;
pid_t pid;
int exitcode;
estream_t errfp;
error_line_t errlines;
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
log_assert (component >= 0 && component < GC_COMPONENT_NR);
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
if (!gc_component[component].program)
return 0;
if (!gc_component[component].module_name)
return 0;
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
pgmname = gnupg_module_name (gc_component[component].module_name);
i = 0;
if (!gnupg_default_homedir_p ()
&& component != GC_COMPONENT_PINENTRY)
{
argv[i++] = "--homedir";
argv[i++] = gnupg_homedir ();
}
if (conf_file)
{
argv[i++] = "--options";
argv[i++] = conf_file;
}
if (component == GC_COMPONENT_PINENTRY)
argv[i++] = "--version";
else
argv[i++] = "--gpgconf-test";
argv[i++] = NULL;
result = 0;
errlines = NULL;
err = gnupg_spawn_process (pgmname, argv, NULL, NULL, 0,
NULL, NULL, &errfp, &pid);
if (err)
result |= 1; /* Program could not be run. */
else
{
errlines = collect_error_output (errfp,
gc_component[component].name);
2010-06-24 12:51:30 +02:00
if (gnupg_wait_process (pgmname, pid, 1, &exitcode))
{
if (exitcode == -1)
result |= 1; /* Program could not be run or it
terminated abnormally. */
result |= 2; /* Program returned an error. */
}
gnupg_release_process (pid);
es_fclose (errfp);
}
/* If the program could not be run, we can't tell whether
the config file is good. */
if (result & 1)
result |= 2;
if (out)
{
const char *desc;
error_line_t errptr;
desc = gc_component[component].desc;
desc = my_dgettext (gc_component[component].desc_domain, desc);
es_fprintf (out, "%s:%s:",
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
gc_component[component].program, gc_percent_escape (desc));
es_fputs (gc_percent_escape (pgmname), out);
es_fprintf (out, ":%d:%d:", !(result & 1), !(result & 2));
for (errptr = errlines; errptr; errptr = errptr->next)
{
if (errptr != errlines)
es_fputs ("\n:::::", out); /* Continuation line. */
if (errptr->fname)
es_fputs (gc_percent_escape (errptr->fname), out);
es_putc (':', out);
if (errptr->fname)
es_fprintf (out, "%u", errptr->lineno);
es_putc (':', out);
es_fputs (gc_percent_escape (errptr->errtext), out);
es_putc (':', out);
}
es_putc ('\n', out);
}
while (errlines)
{
error_line_t tmp = errlines->next;
xfree (errlines);
errlines = tmp;
}
return result;
}
/* Check all components that are available. */
void
gc_check_programs (estream_t out)
{
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
gc_component_id_t component;
for (component = 0; component < GC_COMPONENT_NR; component++)
gc_component_check_options (component, out, NULL);
}
/* Find the component with the name NAME. Returns -1 if not
found. */
int
gc_component_find (const char *name)
{
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
gc_component_id_t idx;
for (idx = 0; idx < GC_COMPONENT_NR; idx++)
{
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
if (gc_component[idx].program
&& !strcmp (name, gc_component[idx].program))
return idx;
}
return -1;
}
/* List the option OPTION. */
static void
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
list_one_option (gc_component_id_t component,
const gc_option_t *option, estream_t out)
{
const char *desc = NULL;
char *arg_name = NULL;
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
unsigned long flags;
const char *desc_domain = gc_component[component].desc_domain;
if (option->desc)
{
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
desc = my_dgettext (desc_domain, option->desc);
if (*desc == '|')
{
const char *arg_tail = strchr (&desc[1], '|');
if (arg_tail)
{
int arg_len = arg_tail - &desc[1];
arg_name = xmalloc (arg_len + 1);
memcpy (arg_name, &desc[1], arg_len);
arg_name[arg_len] = '\0';
desc = arg_tail + 1;
}
}
}
/* YOU MUST NOT REORDER THE FIELDS IN THIS OUTPUT, AS THEIR ORDER IS
PART OF THE EXTERNAL INTERFACE. YOU MUST NOT REMOVE ANY
FIELDS. */
/* The name field. */
es_fprintf (out, "%s", option->name);
/* The flags field. */
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
flags = 0;
if (option->is_header) flags |= GC_OPT_FLAG_GROUP;
if (option->is_list) flags |= GC_OPT_FLAG_LIST;
if (option->runtime) flags |= GC_OPT_FLAG_RUNTIME;
if (option->has_default) flags |= GC_OPT_FLAG_DEFAULT;
if (option->def_in_desc) flags |= GC_OPT_FLAG_DEF_DESC;
if (option->no_arg_desc) flags |= GC_OPT_FLAG_NO_ARG_DESC;
if (option->no_change) flags |= GC_OPT_FLAG_NO_CHANGE;
es_fprintf (out, ":%lu", flags);
if (opt.verbose)
{
es_putc (' ', out);
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
if (!flags)
es_fprintf (out, "none");
else
{
unsigned long flag = 0;
unsigned long first = 1;
while (flags)
{
if (flags & 1)
{
if (first)
first = 0;
else
es_putc (',', out);
es_fprintf (out, "%s", gc_flag[flag].name);
}
flags >>= 1;
flag++;
}
}
}
/* The level field. */
es_fprintf (out, ":%u", option->level);
if (opt.verbose)
es_fprintf (out, " %s", gc_level[option->level].name);
/* The description field. */
es_fprintf (out, ":%s", desc ? gc_percent_escape (desc) : "");
/* The type field. */
es_fprintf (out, ":%u", option->arg_type);
if (opt.verbose)
es_fprintf (out, " %s", gc_arg_type[option->arg_type].name);
/* The alternate type field. */
es_fprintf (out, ":%u", gc_arg_type[option->arg_type].fallback);
if (opt.verbose)
es_fprintf (out, " %s",
gc_arg_type[gc_arg_type[option->arg_type].fallback].name);
/* The argument name field. */
es_fprintf (out, ":%s", arg_name ? gc_percent_escape (arg_name) : "");
xfree (arg_name);
/* The default value field. */
es_fprintf (out, ":%s", option->default_value ? option->default_value : "");
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
/* The default argument field. This was never used and is thus empty. */
es_fprintf (out, ":");
/* The value field. */
if (gc_arg_type[option->arg_type].fallback == GC_ARG_TYPE_NONE
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
&& option->is_list && option->value)
{
/* The special format "1,1,1,1,...,1" is converted to a number
here. */
es_fprintf (out, ":%u", (unsigned int)((strlen (option->value) + 1) / 2));
}
else
es_fprintf (out, ":%s", option->value ? option->value : "");
/* ADD NEW FIELDS HERE. */
es_putc ('\n', out);
}
/* List all options of the component COMPONENT. */
void
gc_component_list_options (int component, estream_t out)
{
const gc_option_t *option = gc_component[component].options;
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
for ( ; option && option->name; option++)
{
/* Do not output unknown or internal options. */
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
if (!option->is_header
gpgconf: Further simplify the gpgconf option processing. * common/gc-opt-flags.h (GC_OPT_FLAG_RUNTIME): Move to ... * tools/gpgconf-comp.c: here. (known_options_scdaemon): Remove "options". (known_options_dirmngr): Remove "options". (known_options_gpgsm): Remove "options". (known_options_gpg): Remove "options" and "keyserver". (struct gc_option_s): Rename active t gpgconf_list. (gc_component_list_options): Do not act upon active. (option_check_validity): Ditto. (is_known_option): Make it work correctly for unknown options. (retrieve_options_from_program): Use renamed flag gpgconf_list only to detect duplicated items from --gpgconf-list. Do not set runtime. Only e set the options if set by --gpgconf-list; never clear them. * agent/gpg-agent.c: Simplify the --gpgconf-list output. * dirmngr/dirmngr.c: Ditto. * g10/gpg.c: Ditto. * kbx/keyboxd.c: Ditto. * scd/scdaemon.c: Ditto. * sm/gpgsm.c: Ditto. * tests/openpgp/gpgconf.scm: Use "compliance" instead of "keyserver" for the string arg test. -- There is no need to read the list of options from the components unless they convey a default value. It is better to consult only the list we have in gpgconf-comp.c to decide on whether an option should be displayed. Right, this might mess up thing if a newer gpgconf version is used with an older component, but we already print warnings in this case and in general we do not want to support this anymore - the times of gpg 1.4. and 2.0 are long over now. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-14 18:53:43 +01:00
&& option->level == GC_LEVEL_INTERNAL)
continue;
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
if (option->is_header)
{
const gc_option_t *group_option = option + 1;
gc_expert_level_t level = GC_LEVEL_NR;
/* The manual states that the group level is always the
minimum of the levels of all contained options. Due to
different active options, and because it is hard to
maintain manually, we calculate it here. The value in
the global static table is ignored. */
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
for ( ; group_option->name; group_option++)
{
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
if (group_option->is_header)
break;
if (group_option->level < level)
level = group_option->level;
}
/* Check if group is empty. */
if (level != GC_LEVEL_NR)
{
gc_option_t opt_copy;
/* Fix up the group level. */
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
opt_copy = *option;
opt_copy.level = level;
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
list_one_option (component, &opt_copy, out);
}
}
else
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
list_one_option (component, option, out);
}
}
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
/* Return true if the option NAME is known and that we want it as
* gpgconf managed option. */
static known_option_t *
is_known_option (gc_component_id_t component, const char *name)
{
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
known_option_t *option = gc_component[component].known_options;
if (option)
{
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
for (; option->name; option++)
if (!strcmp (option->name, name))
break;
}
gpgconf: Further simplify the gpgconf option processing. * common/gc-opt-flags.h (GC_OPT_FLAG_RUNTIME): Move to ... * tools/gpgconf-comp.c: here. (known_options_scdaemon): Remove "options". (known_options_dirmngr): Remove "options". (known_options_gpgsm): Remove "options". (known_options_gpg): Remove "options" and "keyserver". (struct gc_option_s): Rename active t gpgconf_list. (gc_component_list_options): Do not act upon active. (option_check_validity): Ditto. (is_known_option): Make it work correctly for unknown options. (retrieve_options_from_program): Use renamed flag gpgconf_list only to detect duplicated items from --gpgconf-list. Do not set runtime. Only e set the options if set by --gpgconf-list; never clear them. * agent/gpg-agent.c: Simplify the --gpgconf-list output. * dirmngr/dirmngr.c: Ditto. * g10/gpg.c: Ditto. * kbx/keyboxd.c: Ditto. * scd/scdaemon.c: Ditto. * sm/gpgsm.c: Ditto. * tests/openpgp/gpgconf.scm: Use "compliance" instead of "keyserver" for the string arg test. -- There is no need to read the list of options from the components unless they convey a default value. It is better to consult only the list we have in gpgconf-comp.c to decide on whether an option should be displayed. Right, this might mess up thing if a newer gpgconf version is used with an older component, but we already print warnings in this case and in general we do not want to support this anymore - the times of gpg 1.4. and 2.0 are long over now. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-14 18:53:43 +01:00
return (option && option->name)? option : NULL;
}
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
/* Find the option NAME in component COMPONENT. Returns pointer to
* the option descriptor or NULL if not found. */
static gc_option_t *
find_option (gc_component_id_t component, const char *name)
{
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
gc_option_t *option = gc_component[component].options;
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
if (option)
{
for (; option->name; option++)
{
if (!option->is_header
&& !strcmp (option->name, name))
return option;
}
}
return NULL;
}
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
/* Retrieve the options for the component COMPONENT. With
* ONLY_INSTALLED set components which are not installed are silently
* ignored. */
static void
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
retrieve_options_from_program (gc_component_id_t component, int only_installed)
{
gpg_error_t err;
const char *pgmname;
const char *argv[2];
estream_t outfp;
int exitcode;
pid_t pid;
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
known_option_t *known_option;
gc_option_t *option;
char *line = NULL;
size_t line_len = 0;
ssize_t length;
const char *config_name;
gpgrt_argparse_t pargs;
int dummy_argc;
char *twopartconfig_name = NULL;
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
gpgrt_opt_t *opt_table = NULL; /* A malloced option table. */
size_t opt_table_used = 0; /* Its current length. */
size_t opt_table_size = 0; /* Its allocated length. */
gc_option_t *opt_info = NULL; /* A malloced options table. */
size_t opt_info_used = 0; /* Its current length. */
size_t opt_info_size = 0; /* Its allocated length. */
int i;
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
pgmname = (gc_component[component].module_name
? gnupg_module_name (gc_component[component].module_name)
: gc_component[component].program );
if (only_installed && gnupg_access (pgmname, X_OK))
{
return; /* The component is not installed. */
}
/* First we need to read the option table from the program. */
argv[0] = "--dump-option-table";
argv[1] = NULL;
err = gnupg_spawn_process (pgmname, argv, NULL, NULL, 0,
NULL, &outfp, NULL, &pid);
if (err)
{
gc_error (1, 0, "could not gather option table from '%s': %s",
pgmname, gpg_strerror (err));
}
while ((length = es_read_line (outfp, &line, &line_len, NULL)) > 0)
{
const char *fields[4];
const char *optname, *optdesc;
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
unsigned int optflags;
int short_opt;
gc_arg_type_t arg_type;
/* Strip newline and carriage return, if present. */
while (length > 0
&& (line[length - 1] == '\n' || line[length - 1] == '\r'))
line[--length] = '\0';
if (split_fields_colon (line, fields, DIM (fields)) < 4)
{
gc_error (0,0, "WARNING: invalid line in option table of '%s'\n",
pgmname);
continue;
}
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
optname = fields[0];
short_opt = atoi (fields[1]);
if (short_opt < 1)
{
gc_error (0,0, "WARNING: bad short option in option table of '%s'\n",
pgmname);
continue;
}
optflags = strtoul (fields[2], NULL, 10);
if ((optflags & ARGPARSE_OPT_HEADER))
known_option = NULL; /* We want all header-only options. */
else if ((known_option = is_known_option (component, optname)))
; /* Yes we want this one. */
else
continue; /* No need to store this option description. */
/* The +1 here is to make sure that we will have a zero item at
* the end of the table. */
if (opt_table_used + 1 >= opt_table_size)
{
/* Note that this also does the initial allocation. */
opt_table_size += 128;
opt_table = xreallocarray (opt_table,
opt_table_used,
opt_table_size,
sizeof *opt_table);
}
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
/* The +1 here is to make sure that we will have a zero item at
* the end of the table. */
if (opt_info_used + 1 >= opt_info_size)
{
/* Note that this also does the initial allocation. */
opt_info_size += 128;
opt_info = xreallocarray (opt_info,
opt_info_used,
opt_info_size,
sizeof *opt_info);
}
/* The +1 here accounts for the two items we are going to add to
* the global string table. */
if (string_array_used + 1 >= string_array_size)
{
string_array_size += 256;
string_array = xreallocarray (string_array,
string_array_used,
string_array_size,
sizeof *string_array);
}
optname = string_array[string_array_used++] = xstrdup (fields[0]);
optdesc = string_array[string_array_used++] = xstrdup (fields[3]);
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
/* Create an option table which can then be supplied to
* gpgrt_parser. Unfortunately there is no private pointer in
* the public option table struct so that we can't add extra
* data we need here. Thus we need to build up another table
* for such info and for ease of use we also copy the tehre the
* data from the option table. It is not possible to use the
* known_option_s for this because that one does not carry
* header lines and it might also be problematic to use such
* static tables for caching options and default values. */
opt_table[opt_table_used].long_opt = optname;
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
opt_table[opt_table_used].short_opt = short_opt;
opt_table[opt_table_used].description = optdesc;
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
opt_table[opt_table_used].flags = optflags;
opt_table_used++;
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
/* Note that as per argparser specs the opt_table uses "@" to
* specifify an empty description. In the DESC script of
* options (opt_info_t) we want to have a real empty string. */
opt_info[opt_info_used].name = optname;
if (*optdesc == '@' && !optdesc[1])
opt_info[opt_info_used].desc = optdesc+1;
else
opt_info[opt_info_used].desc = optdesc;
/* Unfortunately we need to remap the types. */
switch ((optflags & ARGPARSE_TYPE_MASK))
{
case ARGPARSE_TYPE_INT: arg_type = GC_ARG_TYPE_INT32; break;
case ARGPARSE_TYPE_LONG: arg_type = GC_ARG_TYPE_INT32; break;
case ARGPARSE_TYPE_ULONG: arg_type = GC_ARG_TYPE_UINT32; break;
case ARGPARSE_TYPE_STRING: arg_type = GC_ARG_TYPE_STRING; break;
default: arg_type = GC_ARG_TYPE_NONE; break;
}
opt_info[opt_info_used].arg_type = arg_type;
if ((optflags & ARGPARSE_OPT_HEADER))
opt_info[opt_info_used].is_header = 1;
if (known_option)
{
if ((known_option->flags & GC_OPT_FLAG_LIST))
opt_info[opt_info_used].is_list = 1;
/* FIXME: The next can also be taken from opt_table->flags.
* We need to check the code whether both specifications match. */
if ((known_option->flags & GC_OPT_FLAG_ARG_OPT))
opt_info[opt_info_used].opt_arg = 1;
gpgconf: Further simplify the gpgconf option processing. * common/gc-opt-flags.h (GC_OPT_FLAG_RUNTIME): Move to ... * tools/gpgconf-comp.c: here. (known_options_scdaemon): Remove "options". (known_options_dirmngr): Remove "options". (known_options_gpgsm): Remove "options". (known_options_gpg): Remove "options" and "keyserver". (struct gc_option_s): Rename active t gpgconf_list. (gc_component_list_options): Do not act upon active. (option_check_validity): Ditto. (is_known_option): Make it work correctly for unknown options. (retrieve_options_from_program): Use renamed flag gpgconf_list only to detect duplicated items from --gpgconf-list. Do not set runtime. Only e set the options if set by --gpgconf-list; never clear them. * agent/gpg-agent.c: Simplify the --gpgconf-list output. * dirmngr/dirmngr.c: Ditto. * g10/gpg.c: Ditto. * kbx/keyboxd.c: Ditto. * scd/scdaemon.c: Ditto. * sm/gpgsm.c: Ditto. * tests/openpgp/gpgconf.scm: Use "compliance" instead of "keyserver" for the string arg test. -- There is no need to read the list of options from the components unless they convey a default value. It is better to consult only the list we have in gpgconf-comp.c to decide on whether an option should be displayed. Right, this might mess up thing if a newer gpgconf version is used with an older component, but we already print warnings in this case and in general we do not want to support this anymore - the times of gpg 1.4. and 2.0 are long over now. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-14 18:53:43 +01:00
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
if ((known_option->flags & GC_OPT_FLAG_RUNTIME))
opt_info[opt_info_used].runtime = 1;
opt_info[opt_info_used].level = known_option->level;
/* Override the received argtype by a complex type. */
if (known_option->arg_type)
opt_info[opt_info_used].arg_type = known_option->arg_type;
}
opt_info_used++;
}
if (length < 0 || es_ferror (outfp))
gc_error (1, errno, "error reading from %s", pgmname);
if (es_fclose (outfp))
gc_error (1, errno, "error closing %s", pgmname);
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
log_assert (opt_table_used == opt_info_used);
err = gnupg_wait_process (pgmname, pid, 1, &exitcode);
if (err)
gc_error (1, 0, "running %s failed (exitcode=%d): %s",
pgmname, exitcode, gpg_strerror (err));
gnupg_release_process (pid);
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
/* Make the gpgrt option table and the internal option table available. */
gc_component[component].opt_table = opt_table;
gc_component[component].options = opt_info;
/* Now read the default options. */
argv[0] = "--gpgconf-list";
argv[1] = NULL;
err = gnupg_spawn_process (pgmname, argv, NULL, NULL, 0,
NULL, &outfp, NULL, &pid);
if (err)
{
gc_error (1, 0, "could not gather active options from '%s': %s",
pgmname, gpg_strerror (err));
}
while ((length = es_read_line (outfp, &line, &line_len, NULL)) > 0)
{
char *linep;
unsigned long flags = 0;
char *default_value = NULL;
/* Strip newline and carriage return, if present. */
while (length > 0
&& (line[length - 1] == '\n' || line[length - 1] == '\r'))
line[--length] = '\0';
linep = strchr (line, ':');
if (linep)
*(linep++) = '\0';
/* Extract additional flags. Default to none. */
if (linep)
{
char *end;
char *tail;
end = strchr (linep, ':');
if (end)
*(end++) = '\0';
2010-03-29 14:57:11 +02:00
gpg_err_set_errno (0);
flags = strtoul (linep, &tail, 0);
if (errno)
gc_error (1, errno, "malformed flags in option %s from %s",
line, pgmname);
if (!(*tail == '\0' || *tail == ':' || *tail == ' '))
gc_error (1, 0, "garbage after flags in option %s from %s",
line, pgmname);
linep = end;
}
/* Extract default value, if present. Default to empty if
not. */
if (linep)
{
char *end;
end = strchr (linep, ':');
if (end)
*(end++) = '\0';
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
if ((flags & GC_OPT_FLAG_DEFAULT))
default_value = linep;
linep = end;
}
/* Look up the option in the component and install the
configuration data. */
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
option = find_option (component, line);
if (option)
{
gpgconf: Further simplify the gpgconf option processing. * common/gc-opt-flags.h (GC_OPT_FLAG_RUNTIME): Move to ... * tools/gpgconf-comp.c: here. (known_options_scdaemon): Remove "options". (known_options_dirmngr): Remove "options". (known_options_gpgsm): Remove "options". (known_options_gpg): Remove "options" and "keyserver". (struct gc_option_s): Rename active t gpgconf_list. (gc_component_list_options): Do not act upon active. (option_check_validity): Ditto. (is_known_option): Make it work correctly for unknown options. (retrieve_options_from_program): Use renamed flag gpgconf_list only to detect duplicated items from --gpgconf-list. Do not set runtime. Only e set the options if set by --gpgconf-list; never clear them. * agent/gpg-agent.c: Simplify the --gpgconf-list output. * dirmngr/dirmngr.c: Ditto. * g10/gpg.c: Ditto. * kbx/keyboxd.c: Ditto. * scd/scdaemon.c: Ditto. * sm/gpgsm.c: Ditto. * tests/openpgp/gpgconf.scm: Use "compliance" instead of "keyserver" for the string arg test. -- There is no need to read the list of options from the components unless they convey a default value. It is better to consult only the list we have in gpgconf-comp.c to decide on whether an option should be displayed. Right, this might mess up thing if a newer gpgconf version is used with an older component, but we already print warnings in this case and in general we do not want to support this anymore - the times of gpg 1.4. and 2.0 are long over now. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-14 18:53:43 +01:00
if (option->gpgconf_list)
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
gc_error (1, errno,
"option %s returned twice from \"%s --gpgconf-list\"",
line, pgmname);
gpgconf: Further simplify the gpgconf option processing. * common/gc-opt-flags.h (GC_OPT_FLAG_RUNTIME): Move to ... * tools/gpgconf-comp.c: here. (known_options_scdaemon): Remove "options". (known_options_dirmngr): Remove "options". (known_options_gpgsm): Remove "options". (known_options_gpg): Remove "options" and "keyserver". (struct gc_option_s): Rename active t gpgconf_list. (gc_component_list_options): Do not act upon active. (option_check_validity): Ditto. (is_known_option): Make it work correctly for unknown options. (retrieve_options_from_program): Use renamed flag gpgconf_list only to detect duplicated items from --gpgconf-list. Do not set runtime. Only e set the options if set by --gpgconf-list; never clear them. * agent/gpg-agent.c: Simplify the --gpgconf-list output. * dirmngr/dirmngr.c: Ditto. * g10/gpg.c: Ditto. * kbx/keyboxd.c: Ditto. * scd/scdaemon.c: Ditto. * sm/gpgsm.c: Ditto. * tests/openpgp/gpgconf.scm: Use "compliance" instead of "keyserver" for the string arg test. -- There is no need to read the list of options from the components unless they convey a default value. It is better to consult only the list we have in gpgconf-comp.c to decide on whether an option should be displayed. Right, this might mess up thing if a newer gpgconf version is used with an older component, but we already print warnings in this case and in general we do not want to support this anymore - the times of gpg 1.4. and 2.0 are long over now. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-14 18:53:43 +01:00
option->gpgconf_list = 1;
gpgconf: Further simplify the gpgconf option processing. * common/gc-opt-flags.h (GC_OPT_FLAG_RUNTIME): Move to ... * tools/gpgconf-comp.c: here. (known_options_scdaemon): Remove "options". (known_options_dirmngr): Remove "options". (known_options_gpgsm): Remove "options". (known_options_gpg): Remove "options" and "keyserver". (struct gc_option_s): Rename active t gpgconf_list. (gc_component_list_options): Do not act upon active. (option_check_validity): Ditto. (is_known_option): Make it work correctly for unknown options. (retrieve_options_from_program): Use renamed flag gpgconf_list only to detect duplicated items from --gpgconf-list. Do not set runtime. Only e set the options if set by --gpgconf-list; never clear them. * agent/gpg-agent.c: Simplify the --gpgconf-list output. * dirmngr/dirmngr.c: Ditto. * g10/gpg.c: Ditto. * kbx/keyboxd.c: Ditto. * scd/scdaemon.c: Ditto. * sm/gpgsm.c: Ditto. * tests/openpgp/gpgconf.scm: Use "compliance" instead of "keyserver" for the string arg test. -- There is no need to read the list of options from the components unless they convey a default value. It is better to consult only the list we have in gpgconf-comp.c to decide on whether an option should be displayed. Right, this might mess up thing if a newer gpgconf version is used with an older component, but we already print warnings in this case and in general we do not want to support this anymore - the times of gpg 1.4. and 2.0 are long over now. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-14 18:53:43 +01:00
if ((flags & GC_OPT_FLAG_DEFAULT))
option->has_default = 1;
if ((flags & GC_OPT_FLAG_DEF_DESC))
option->def_in_desc = 1;
if ((flags & GC_OPT_FLAG_NO_ARG_DESC))
option->no_arg_desc = 1;
if ((flags & GC_OPT_FLAG_NO_CHANGE))
option->no_change = 1;
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
if (default_value && *default_value)
option->default_value = xstrdup (default_value);
}
}
if (length < 0 || es_ferror (outfp))
gc_error (1, errno, "error reading from %s", pgmname);
2010-08-23 21:26:05 +02:00
if (es_fclose (outfp))
gc_error (1, errno, "error closing %s", pgmname);
2010-06-24 12:51:30 +02:00
err = gnupg_wait_process (pgmname, pid, 1, &exitcode);
if (err)
gc_error (1, 0, "running %s failed (exitcode=%d): %s",
pgmname, exitcode, gpg_strerror (err));
gnupg_release_process (pid);
/* At this point, we can parse the configuration file. */
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
config_name = gc_component[component].option_config_filename;
if (!config_name)
gc_error (1, 0, "name of config file for %s is not known\n", pgmname);
if (!gnupg_default_homedir_p ())
{
/* This is not the default homedir. We need to take an absolute
* config name for the user config file; gpgrt_argparser
* fortunately supports this. */
char *tmp = make_filename (gnupg_homedir (), config_name, NULL);
twopartconfig_name = xstrconcat (config_name, PATHSEP_S, tmp, NULL);
xfree (tmp);
config_name = twopartconfig_name;
}
memset (&pargs, 0, sizeof pargs);
dummy_argc = 0;
pargs.argc = &dummy_argc;
pargs.flags = (ARGPARSE_FLAG_KEEP
| ARGPARSE_FLAG_SYS
| ARGPARSE_FLAG_USER
| ARGPARSE_FLAG_WITHATTR);
if (opt.verbose)
pargs.flags |= ARGPARSE_FLAG_VERBOSE;
while (gpgrt_argparser (&pargs, opt_table, config_name))
{
char *opt_value;
if (pargs.r_type & ARGPARSE_OPT_IGNORE)
{
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
/* log_debug ("ignored\n"); */
continue;
}
if (pargs.r_opt == ARGPARSE_CONFFILE)
{
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
/* log_debug ("current conffile='%s'\n", */
/* pargs.r_type? pargs.r.ret_str: "[cmdline]"); */
continue;
}
/* We only have the short option. Search in the option table
* for the long option name. */
for (i=0; opt_table[i].short_opt; i++)
if (opt_table[i].short_opt == pargs.r_opt)
break;
if (!opt_table[i].short_opt || !opt_table[i].long_opt)
continue; /* No or only a short option - ignore. */
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
/* Look up the option from the config file in our list of
* supported options. */
option= find_option (component, opt_table[i].long_opt);
if (!option)
continue; /* We don't want to handle this option. */
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
option->attr_ignore = !!(pargs.r_type & ARGPARSE_ATTR_IGNORE);
option->attr_force = !!(pargs.r_type & ARGPARSE_ATTR_FORCE);
switch ((pargs.r_type & ARGPARSE_TYPE_MASK))
{
case ARGPARSE_TYPE_INT:
opt_value = xasprintf ("%d", pargs.r.ret_int);
break;
case ARGPARSE_TYPE_LONG:
opt_value = xasprintf ("%ld", pargs.r.ret_long);
break;
case ARGPARSE_TYPE_ULONG:
opt_value = xasprintf ("%lu", pargs.r.ret_ulong);
break;
case ARGPARSE_TYPE_STRING:
opt_value = xasprintf ("\"%s", gc_percent_escape (pargs.r.ret_str));
break;
default: /* ARGPARSE_TYPE_NONE or any unknown type. */
opt_value = xstrdup ("1"); /* Make sure we have some value. */
break;
}
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
/* Now enter the value read from the config file into the table. */
if (!option->is_list)
{
xfree (option->value);
option->value = opt_value;
}
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
else if (!option->value) /* LIST but first item. */
option->value = opt_value;
else
{
char *old = option->value;
option->value = xstrconcat (old, ",", opt_value, NULL);
xfree (old);
xfree (opt_value);
}
}
xfree (line);
xfree (twopartconfig_name);
}
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
/* Retrieve the currently active options and their defaults for this
component. Using -1 for component will retrieve all options from
all installed components. */
void
gc_component_retrieve_options (int component)
{
int process_all = 0;
if (component == -1)
{
process_all = 1;
component = 0;
}
do
{
if (component == GC_COMPONENT_PINENTRY)
continue; /* Skip this dummy component. */
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
if (gc_component[component].program)
retrieve_options_from_program (component, process_all);
}
while (process_all && ++component < GC_COMPONENT_NR);
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
}
/* Perform a simple validity check based on the type. Return in
* NEW_VALUE_NR the value of the number in NEW_VALUE if OPTION is of
* type GC_ARG_TYPE_NONE. If VERBATIM is set the profile parsing mode
* is used. */
static void
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
option_check_validity (gc_component_id_t component,
gc_option_t *option, unsigned long flags,
char *new_value, unsigned long *new_value_nr,
int verbatim)
{
char *arg;
gpgconf: Further simplify the gpgconf option processing. * common/gc-opt-flags.h (GC_OPT_FLAG_RUNTIME): Move to ... * tools/gpgconf-comp.c: here. (known_options_scdaemon): Remove "options". (known_options_dirmngr): Remove "options". (known_options_gpgsm): Remove "options". (known_options_gpg): Remove "options" and "keyserver". (struct gc_option_s): Rename active t gpgconf_list. (gc_component_list_options): Do not act upon active. (option_check_validity): Ditto. (is_known_option): Make it work correctly for unknown options. (retrieve_options_from_program): Use renamed flag gpgconf_list only to detect duplicated items from --gpgconf-list. Do not set runtime. Only e set the options if set by --gpgconf-list; never clear them. * agent/gpg-agent.c: Simplify the --gpgconf-list output. * dirmngr/dirmngr.c: Ditto. * g10/gpg.c: Ditto. * kbx/keyboxd.c: Ditto. * scd/scdaemon.c: Ditto. * sm/gpgsm.c: Ditto. * tests/openpgp/gpgconf.scm: Use "compliance" instead of "keyserver" for the string arg test. -- There is no need to read the list of options from the components unless they convey a default value. It is better to consult only the list we have in gpgconf-comp.c to decide on whether an option should be displayed. Right, this might mess up thing if a newer gpgconf version is used with an older component, but we already print warnings in this case and in general we do not want to support this anymore - the times of gpg 1.4. and 2.0 are long over now. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-14 18:53:43 +01:00
(void)component;
if (option->new_flags || option->new_value)
gc_error (1, 0, "option %s already changed", option->name);
if (flags & GC_OPT_FLAG_DEFAULT)
{
if (*new_value)
gc_error (1, 0, "argument %s provided for deleted option %s",
new_value, option->name);
return;
}
/* GC_ARG_TYPE_NONE options have special list treatment. */
if (gc_arg_type[option->arg_type].fallback == GC_ARG_TYPE_NONE)
{
char *tail;
2010-03-29 14:57:11 +02:00
gpg_err_set_errno (0);
*new_value_nr = strtoul (new_value, &tail, 0);
if (errno)
gc_error (1, errno, "invalid argument for option %s",
option->name);
if (*tail)
gc_error (1, 0, "garbage after argument for option %s",
option->name);
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
if (!option->is_list)
{
if (*new_value_nr != 1)
gc_error (1, 0, "argument for non-list option %s of type 0 "
"(none) must be 1", option->name);
}
else
{
if (*new_value_nr == 0)
gc_error (1, 0, "argument for option %s of type 0 (none) "
"must be positive", option->name);
}
return;
}
arg = new_value;
do
{
if (*arg == '\0' || (*arg == ',' && !verbatim))
{
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
if (!option->opt_arg)
gc_error (1, 0, "argument required for option %s", option->name);
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
if (*arg == ',' && !verbatim && !option->is_list)
gc_error (1, 0, "list found for non-list option %s", option->name);
}
else if (gc_arg_type[option->arg_type].fallback == GC_ARG_TYPE_STRING)
{
if (*arg != '"' && !verbatim)
gc_error (1, 0, "string argument for option %s must begin "
"with a quote (\") character", option->name);
/* FIXME: We do not allow empty string arguments for now, as
we do not quote arguments in configuration files, and
thus no argument is indistinguishable from the empty
string. */
if (arg[1] == '\0' || (arg[1] == ',' && !verbatim))
gc_error (1, 0, "empty string argument for option %s is "
"currently not allowed. Please report this!",
option->name);
}
else if (gc_arg_type[option->arg_type].fallback == GC_ARG_TYPE_INT32)
{
long res;
2010-03-29 14:57:11 +02:00
gpg_err_set_errno (0);
res = strtol (arg, &arg, 0);
(void) res;
if (errno)
gc_error (1, errno, "invalid argument for option %s",
option->name);
if (*arg != '\0' && (*arg != ',' || verbatim))
gc_error (1, 0, "garbage after argument for option %s",
option->name);
}
else if (gc_arg_type[option->arg_type].fallback == GC_ARG_TYPE_UINT32)
{
unsigned long res;
2010-03-29 14:57:11 +02:00
gpg_err_set_errno (0);
res = strtoul (arg, &arg, 0);
(void) res;
if (errno)
gc_error (1, errno, "invalid argument for option %s",
option->name);
if (*arg != '\0' && (*arg != ',' || verbatim))
gc_error (1, 0, "garbage after argument for option %s",
option->name);
}
arg = verbatim? strchr (arg, ',') : NULL;
if (arg)
arg++;
}
while (arg && *arg);
}
#ifdef HAVE_W32_SYSTEM
int
copy_file (const char *src_name, const char *dst_name)
{
#define BUF_LEN 4096
char buffer[BUF_LEN];
int len;
gpgrt_stream_t src;
gpgrt_stream_t dst;
src = gpgrt_fopen (src_name, "r");
if (src == NULL)
return -1;
dst = gpgrt_fopen (dst_name, "w");
if (dst == NULL)
{
int saved_err = errno;
gpgrt_fclose (src);
2010-03-29 14:57:11 +02:00
gpg_err_set_errno (saved_err);
return -1;
}
do
{
int written;
len = gpgrt_fread (buffer, 1, BUF_LEN, src);
if (len == 0)
break;
written = gpgrt_fwrite (buffer, 1, len, dst);
if (written != len)
break;
}
while (! gpgrt_feof (src) && ! gpgrt_ferror (src) && ! gpgrt_ferror (dst));
if (gpgrt_ferror (src) || gpgrt_ferror (dst) || ! gpgrt_feof (src))
{
int saved_errno = errno;
gpgrt_fclose (src);
gpgrt_fclose (dst);
unlink (dst_name);
2010-03-29 14:57:11 +02:00
gpg_err_set_errno (saved_errno);
return -1;
}
if (gpgrt_fclose (dst))
gc_error (1, errno, "error closing %s", dst_name);
if (gpgrt_fclose (src))
gc_error (1, errno, "error closing %s", src_name);
return 0;
}
#endif /* HAVE_W32_SYSTEM */
/* Create and verify the new configuration file for the specified
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
* component. Returns 0 on success and -1 on error. If
* VERBATIM is set the profile mode is used. This function may store
* pointers to malloced strings in SRC_FILENAMEP, DEST_FILENAMEP, and
* ORIG_FILENAMEP. Those must be freed by the caller. The strings
* refer to three versions of the configuration file:
*
* SRC_FILENAME: The updated configuration is written to this file.
* DEST_FILENAME: Name of the configuration file read by the
* component.
* ORIG_FILENAME: A backup of the previous configuration file.
*
* To apply the configuration change, rename SRC_FILENAME to
* DEST_FILENAME. To revert to the previous configuration, rename
* ORIG_FILENAME to DEST_FILENAME. */
static int
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
change_options_program (gc_component_id_t component,
char **src_filenamep, char **dest_filenamep,
char **orig_filenamep,
int verbatim)
{
static const char marker[] = "###+++--- " GPGCONF_DISP_NAME " ---+++###";
/* True if we are within the marker in the config file. */
int in_marker = 0;
gc_option_t *option;
char *line = NULL;
size_t line_len;
ssize_t length;
int res;
int fd;
gpgrt_stream_t src_file = NULL;
gpgrt_stream_t dest_file = NULL;
char *src_filename;
char *dest_filename;
char *orig_filename;
/* Special hack for gpg, see below. */
int utf8strings_seen = 0;
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
/* FIXME. Throughout the function, do better error reporting. */
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
if (!gc_component[component].option_config_filename)
gc_error (1, 0, "name of config file for %s is not known\n",
gc_component[component].name);
dest_filename = make_absfilename
(gnupg_homedir (), gc_component[component].option_config_filename, NULL);
src_filename = xasprintf ("%s.%s.%i.new",
dest_filename, GPGCONF_NAME, (int)getpid ());
orig_filename = xasprintf ("%s.%s.%i.bak",
dest_filename, GPGCONF_NAME, (int)getpid ());
#ifdef HAVE_W32_SYSTEM
res = copy_file (dest_filename, orig_filename);
#else
res = link (dest_filename, orig_filename);
#endif
if (res < 0 && errno != ENOENT)
{
xfree (dest_filename);
xfree (src_filename);
xfree (orig_filename);
return -1;
}
if (res < 0)
{
xfree (orig_filename);
orig_filename = NULL;
}
/* We now initialize the return strings, so the caller can do the
cleanup for us. */
*src_filenamep = src_filename;
*dest_filenamep = dest_filename;
*orig_filenamep = orig_filename;
/* Use open() so that we can use O_EXCL.
* FIXME: gpgrt has an x flag for quite some time now - use that. */
fd = gnupg_open (src_filename, O_CREAT | O_EXCL | O_WRONLY, 0644);
if (fd < 0)
return -1;
src_file = gpgrt_fdopen (fd, "w");
res = errno;
if (!src_file)
{
2010-03-29 14:57:11 +02:00
gpg_err_set_errno (res);
return -1;
}
/* Only if ORIG_FILENAME is not NULL did the configuration file
exist already. In this case, we will copy its content into the
new configuration file, changing it to our liking in the
process. */
if (orig_filename)
{
dest_file = gpgrt_fopen (dest_filename, "r");
if (!dest_file)
goto change_one_err;
while ((length = gpgrt_read_line (dest_file, &line, &line_len, NULL)) > 0)
{
int disable = 0;
char *start;
if (!strncmp (marker, line, sizeof (marker) - 1))
{
if (!in_marker)
in_marker = 1;
else
break;
}
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
else if (component == GC_COMPONENT_GPG && in_marker
&& ! strcmp ("utf8-strings\n", line))
{
/* Strip duplicated entries. */
if (utf8strings_seen)
disable = 1;
else
utf8strings_seen = 1;
}
start = line;
while (*start == ' ' || *start == '\t')
start++;
if (*start && *start != '\r' && *start != '\n' && *start != '#')
{
char *end;
char saved_end;
end = start;
while (*end && *end != ' ' && *end != '\t'
&& *end != '\r' && *end != '\n' && *end != '#')
end++;
saved_end = *end;
*end = '\0';
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
option = find_option (component, start);
*end = saved_end;
if (option && ((option->new_flags & GC_OPT_FLAG_DEFAULT)
|| option->new_value))
disable = 1;
}
if (disable)
{
if (!in_marker)
{
gpgrt_fprintf (src_file,
"# %s disabled this option here at %s\n",
GPGCONF_DISP_NAME, asctimestamp (gnupg_get_time ()));
if (gpgrt_ferror (src_file))
goto change_one_err;
gpgrt_fprintf (src_file, "# %s", line);
if (gpgrt_ferror (src_file))
goto change_one_err;
}
}
else
{
gpgrt_fprintf (src_file, "%s", line);
if (gpgrt_ferror (src_file))
goto change_one_err;
}
}
if (length < 0 || gpgrt_ferror (dest_file))
goto change_one_err;
}
if (!in_marker)
{
/* There was no marker. This is the first time we edit the
file. We add our own marker at the end of the file and
proceed. Note that we first write a newline, this guards us
against files which lack the newline at the end of the last
line, while it doesn't hurt us in all other cases. */
gpgrt_fprintf (src_file, "\n%s\n", marker);
if (gpgrt_ferror (src_file))
goto change_one_err;
}
/* At this point, we have copied everything up to the end marker
into the new file, except for the options we are going to change.
Now, dump the changed options (except for those we are going to
revert to their default), and write the end marker, possibly
followed by the rest of the original file. */
/* We have to turn on UTF8 strings for GnuPG. */
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
if (component == GC_COMPONENT_GPG && ! utf8strings_seen)
gpgrt_fprintf (src_file, "utf8-strings\n");
option = gc_component[component].options;
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
for ( ; option->name; option++)
{
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
if (!option->is_header && option->new_value)
{
char *arg = option->new_value;
do
{
if (*arg == '\0' || *arg == ',')
{
gpgrt_fprintf (src_file, "%s\n", option->name);
if (gpgrt_ferror (src_file))
goto change_one_err;
}
else if (gc_arg_type[option->arg_type].fallback
== GC_ARG_TYPE_NONE)
{
log_assert (*arg == '1');
gpgrt_fprintf (src_file, "%s\n", option->name);
if (gpgrt_ferror (src_file))
goto change_one_err;
arg++;
}
else if (gc_arg_type[option->arg_type].fallback
== GC_ARG_TYPE_STRING)
{
char *end;
if (!verbatim)
{
log_assert (*arg == '"');
arg++;
end = strchr (arg, ',');
if (end)
*end = '\0';
}
else
end = NULL;
gpgrt_fprintf (src_file, "%s %s\n", option->name,
verbatim? arg : percent_deescape (arg));
if (gpgrt_ferror (src_file))
goto change_one_err;
if (end)
*end = ',';
arg = end;
}
else
{
char *end;
end = strchr (arg, ',');
if (end)
*end = '\0';
gpgrt_fprintf (src_file, "%s %s\n", option->name, arg);
if (gpgrt_ferror (src_file))
goto change_one_err;
if (end)
*end = ',';
arg = end;
}
log_assert (arg == NULL || *arg == '\0' || *arg == ',');
if (arg && *arg == ',')
arg++;
}
while (arg && *arg);
}
}
gpgrt_fprintf (src_file, "%s %s\n", marker, asctimestamp (gnupg_get_time ()));
if (gpgrt_ferror (src_file))
goto change_one_err;
if (!in_marker)
{
gpgrt_fprintf (src_file, "# %s edited this configuration file.\n",
GPGCONF_DISP_NAME);
if (gpgrt_ferror (src_file))
goto change_one_err;
gpgrt_fprintf (src_file, "# It will disable options before this marked "
"block, but it will\n");
if (gpgrt_ferror (src_file))
goto change_one_err;
gpgrt_fprintf (src_file, "# never change anything below these lines.\n");
if (gpgrt_ferror (src_file))
goto change_one_err;
}
if (dest_file)
{
while ((length = gpgrt_read_line (dest_file, &line, &line_len, NULL)) > 0)
{
gpgrt_fprintf (src_file, "%s", line);
if (gpgrt_ferror (src_file))
goto change_one_err;
}
if (length < 0 || gpgrt_ferror (dest_file))
goto change_one_err;
}
xfree (line);
line = NULL;
res = gpgrt_fclose (src_file);
if (res)
{
res = errno;
close (fd);
if (dest_file)
gpgrt_fclose (dest_file);
2010-03-29 14:57:11 +02:00
gpg_err_set_errno (res);
return -1;
}
close (fd);
if (dest_file)
{
res = gpgrt_fclose (dest_file);
if (res)
return -1;
}
return 0;
change_one_err:
xfree (line);
res = errno;
if (src_file)
{
gpgrt_fclose (src_file);
close (fd);
}
if (dest_file)
gpgrt_fclose (dest_file);
2010-03-29 14:57:11 +02:00
gpg_err_set_errno (res);
return -1;
}
/* Common code for gc_component_change_options and
* gc_process_gpgconf_conf. If VERBATIM is set the profile parsing
* mode is used. */
static void
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
change_one_value (gc_component_id_t component,
gc_option_t *option, int *r_runtime,
unsigned long flags, char *new_value, int verbatim)
{
unsigned long new_value_nr = 0;
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
option_check_validity (component, option,
flags, new_value, &new_value_nr, verbatim);
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
if (option->runtime)
*r_runtime = 1;
option->new_flags = flags;
if (!(flags & GC_OPT_FLAG_DEFAULT))
{
if (gc_arg_type[option->arg_type].fallback == GC_ARG_TYPE_NONE
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
&& option->is_list)
{
char *str;
/* We convert the number to a list of 1's for convenient
list handling. */
log_assert (new_value_nr > 0);
option->new_value = xmalloc ((2 * (new_value_nr - 1) + 1) + 1);
str = option->new_value;
*(str++) = '1';
while (--new_value_nr > 0)
{
*(str++) = ',';
*(str++) = '1';
}
*(str++) = '\0';
}
else
option->new_value = xstrdup (new_value);
}
}
/* Read the modifications from IN and apply them. If IN is NULL the
modifications are expected to already have been set to the global
table. If VERBATIM is set the profile mode is used. */
void
gc_component_change_options (int component, estream_t in, estream_t out,
int verbatim)
{
int err = 0;
int block = 0;
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
int runtime = 0;
char *src_filename = NULL;
char *dest_filename = NULL;
char *orig_filename = NULL;
gc_option_t *option;
char *line = NULL;
size_t line_len = 0;
ssize_t length;
if (component == GC_COMPONENT_PINENTRY)
return; /* Dummy component for now. */
if (in)
{
/* Read options from the file IN. */
while ((length = es_read_line (in, &line, &line_len, NULL)) > 0)
{
char *linep;
unsigned long flags = 0;
char *new_value = "";
/* Strip newline and carriage return, if present. */
while (length > 0
&& (line[length - 1] == '\n' || line[length - 1] == '\r'))
line[--length] = '\0';
linep = strchr (line, ':');
if (linep)
*(linep++) = '\0';
/* Extract additional flags. Default to none. */
if (linep)
{
char *end;
char *tail;
end = strchr (linep, ':');
if (end)
*(end++) = '\0';
2010-03-29 14:57:11 +02:00
gpg_err_set_errno (0);
flags = strtoul (linep, &tail, 0);
if (errno)
gc_error (1, errno, "malformed flags in option %s", line);
if (!(*tail == '\0' || *tail == ':' || *tail == ' '))
gc_error (1, 0, "garbage after flags in option %s", line);
linep = end;
}
/* Don't allow setting of the no change flag. */
flags &= ~GC_OPT_FLAG_NO_CHANGE;
/* Extract default value, if present. Default to empty if not. */
if (linep)
{
char *end;
end = strchr (linep, ':');
if (end)
*(end++) = '\0';
new_value = linep;
linep = end;
}
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
option = find_option (component, line);
if (!option)
gc_error (1, 0, "unknown option %s", line);
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
if (option->no_change)
{
gc_error (0, 0, "ignoring new value for option %s",
option->name);
continue;
}
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
change_one_value (component, option, &runtime, flags, new_value, 0);
}
if (length < 0 || gpgrt_ferror (in))
gc_error (1, errno, "error reading stream 'in'");
}
/* Now that we have collected and locally verified the changes,
write them out to new configuration files, verify them
externally, and then commit them. */
option = gc_component[component].options;
2007-06-26 15:48:44 +02:00
while (option && option->name)
{
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
/* Go on if there is nothing to do. */
if (src_filename || !(option->new_flags || option->new_value))
{
option++;
continue;
}
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
if (gc_component[component].program)
{
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
err = change_options_program (component,
&src_filename,
&dest_filename,
&orig_filename,
verbatim);
if (! err)
{
/* External verification. */
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
err = gc_component_check_options (component, out, src_filename);
if (err)
{
gc_error (0, 0,
_("External verification of component %s failed"),
gc_component[component].name);
2010-03-29 14:57:11 +02:00
gpg_err_set_errno (EINVAL);
}
}
}
if (err)
break;
option++;
}
/* We are trying to atomically commit all changes. Unfortunately,
we cannot rely on gnupg_rename_file to manage the signals for us,
doing so would require us to pass NULL as BLOCK to any subsequent
call to it. Instead, we just manage the signal handling
manually. */
block = 1;
gnupg_block_all_signals ();
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
if (!err && !opt.dry_run)
{
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
if (src_filename)
{
/* FIXME: Make a verification here. */
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
log_assert (dest_filename);
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
if (orig_filename)
err = gnupg_rename_file (src_filename, dest_filename, NULL);
else
{
#ifdef HAVE_W32_SYSTEM
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
/* We skip the unlink if we expect the file not to be
* there. */
err = gnupg_rename_file (src_filename, dest_filename, NULL);
#else /* HAVE_W32_SYSTEM */
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
/* This is a bit safer than rename() because we expect
* DEST_FILENAME not to be there. If it happens to be
* there, this will fail. */
err = link (src_filename, dest_filename);
if (!err)
err = unlink (src_filename);
#endif /* !HAVE_W32_SYSTEM */
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
}
if (!err)
{
xfree (src_filename);
src_filename = NULL;
}
}
}
if (err || opt.dry_run)
{
int saved_errno = errno;
Fix typos found using codespell. * agent/cache.c: Fix typos. * agent/call-pinentry.c: Likewise. * agent/call-scd.c: Likewise. * agent/command-ssh.c: Likewise. * agent/command.c: Likewise. * agent/divert-scd.c: Likewise. * agent/findkey.c: Likewise. * agent/gpg-agent.c: Likewise. * agent/w32main.c: Likewise. * common/argparse.c: Likewise. * common/audit.c: Likewise. * common/audit.h: Likewise. * common/convert.c: Likewise. * common/dotlock.c: Likewise. * common/exechelp-posix.c: Likewise. * common/exechelp-w32.c: Likewise. * common/exechelp-w32ce.c: Likewise. * common/exechelp.h: Likewise. * common/helpfile.c: Likewise. * common/i18n.h: Likewise. * common/iobuf.c: Likewise. * common/iobuf.h: Likewise. * common/localename.c: Likewise. * common/logging.c: Likewise. * common/openpgp-oid.c: Likewise. * common/session-env.c: Likewise. * common/sexputil.c: Likewise. * common/sysutils.c: Likewise. * common/t-sexputil.c: Likewise. * common/ttyio.c: Likewise. * common/util.h: Likewise. * dirmngr/cdblib.c: Likewise. * dirmngr/certcache.c: Likewise. * dirmngr/crlcache.c: Likewise. * dirmngr/dirmngr-client.c: Likewise. * dirmngr/dirmngr.c: Likewise. * dirmngr/dirmngr_ldap.c: Likewise. * dirmngr/dns-stuff.c: Likewise. * dirmngr/http.c: Likewise. * dirmngr/ks-engine-hkp.c: Likewise. * dirmngr/ks-engine-ldap.c: Likewise. * dirmngr/ldap-wrapper.c: Likewise. * dirmngr/ldap.c: Likewise. * dirmngr/misc.c: Likewise. * dirmngr/ocsp.c: Likewise. * dirmngr/validate.c: Likewise. * g10/encrypt.c: Likewise. * g10/getkey.c: Likewise. * g10/gpg.c: Likewise. * g10/gpgv.c: Likewise. * g10/import.c: Likewise. * g10/keydb.c: Likewise. * g10/keydb.h: Likewise. * g10/keygen.c: Likewise. * g10/keyid.c: Likewise. * g10/keylist.c: Likewise. * g10/keyring.c: Likewise. * g10/mainproc.c: Likewise. * g10/misc.c: Likewise. * g10/options.h: Likewise. * g10/packet.h: Likewise. * g10/parse-packet.c: Likewise. * g10/pkclist.c: Likewise. * g10/pkglue.c: Likewise. * g10/plaintext.c: Likewise. * g10/server.c: Likewise. * g10/sig-check.c: Likewise. * g10/sqlite.c: Likewise. * g10/tdbio.c: Likewise. * g10/test-stubs.c: Likewise. * g10/tofu.c: Likewise. * g10/trust.c: Likewise. * g10/trustdb.c: Likewise. * g13/create.c: Likewise. * g13/mountinfo.c: Likewise. * kbx/keybox-blob.c: Likewise. * kbx/keybox-file.c: Likewise. * kbx/keybox-init.c: Likewise. * kbx/keybox-search-desc.h: Likewise. * kbx/keybox-search.c: Likewise. * kbx/keybox-update.c: Likewise. * scd/apdu.c: Likewise. * scd/app-openpgp.c: Likewise. * scd/app-p15.c: Likewise. * scd/app.c: Likewise. * scd/ccid-driver.c: Likewise. * scd/command.c: Likewise. * scd/iso7816.c: Likewise. * sm/base64.c: Likewise. * sm/call-agent.c: Likewise. * sm/call-dirmngr.c: Likewise. * sm/certchain.c: Likewise. * sm/gpgsm.c: Likewise. * sm/import.c: Likewise. * sm/keydb.c: Likewise. * sm/minip12.c: Likewise. * sm/qualified.c: Likewise. * sm/server.c: Likewise. * tools/gpg-check-pattern.c: Likewise. * tools/gpgconf-comp.c: Likewise. * tools/gpgkey2ssh.c: Likewise. * tools/gpgparsemail.c: Likewise. * tools/gpgtar.c: Likewise. * tools/rfc822parse.c: Likewise. * tools/symcryptrun.c: Likewise. Signed-off-by: Justus Winter <justus@g10code.com>
2015-11-16 12:41:46 +01:00
/* An error occurred or a dry-run is requested. */
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
if (src_filename)
{
/* The change was not yet committed. */
unlink (src_filename);
if (orig_filename)
unlink (orig_filename);
}
else
{
/* The changes were already committed. FIXME: This is a tad
dangerous, as we don't know if we don't overwrite a
version of the file that is even newer than the one we
just installed. */
if (orig_filename)
gnupg_rename_file (orig_filename, dest_filename, NULL);
else
unlink (dest_filename);
}
if (err)
gc_error (1, saved_errno, "could not commit changes");
/* Fall-through for dry run. */
goto leave;
}
/* If it all worked, notify the daemons of the changes. */
if (opt.runtime)
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
do_runtime_change (component, 0);
/* Move the per-process backup file into its place. */
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
if (orig_filename)
{
char *backup_filename;
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
log_assert (dest_filename);
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
backup_filename = xasprintf ("%s.%s.bak",
dest_filename, GPGCONF_NAME);
gnupg_rename_file (orig_filename, backup_filename, NULL);
xfree (backup_filename);
}
leave:
if (block)
gnupg_unblock_all_signals ();
xfree (line);
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
xfree (src_filename);
xfree (dest_filename);
xfree (orig_filename);
}
2020-04-07 16:01:40 +02:00
/* Check whether USER matches the current user or one of its group.
This function may change USER. Returns true is there is a
match. */
static int
key_matches_user_or_group (char *user)
{
char *group;
if (*user == '*' && user[1] == 0)
return 1; /* A single asterisk matches all users. */
group = strchr (user, ':');
if (group)
*group++ = 0;
2007-06-26 15:48:44 +02:00
#ifdef HAVE_W32_SYSTEM
/* Under Windows we don't support groups. */
2007-06-26 15:48:44 +02:00
if (group && *group)
gc_error (0, 0, _("Note that group specifications are ignored\n"));
2010-03-29 14:57:11 +02:00
#ifndef HAVE_W32CE_SYSTEM
2007-06-26 15:48:44 +02:00
if (*user)
{
static char *my_name;
if (!my_name)
{
char tmp[1];
DWORD size = 1;
GetUserNameA (tmp, &size);
my_name = xmalloc (size);
if (!GetUserNameA (my_name, &size))
gc_error (1,0, "error getting current user name: %s",
w32_strerror (-1));
}
if (!strcmp (user, my_name))
return 1; /* Found. */
}
2010-03-29 14:57:11 +02:00
#endif /*HAVE_W32CE_SYSTEM*/
2007-06-26 15:48:44 +02:00
#else /*!HAVE_W32_SYSTEM*/
/* First check whether the user matches. */
if (*user)
{
static char *my_name;
if (!my_name)
{
struct passwd *pw = getpwuid ( getuid () );
if (!pw)
gc_error (1, errno, "getpwuid failed for current user");
my_name = xstrdup (pw->pw_name);
}
if (!strcmp (user, my_name))
return 1; /* Found. */
}
/* If that failed, check whether a group matches. */
if (group && *group)
{
static char *my_group;
static char **my_supgroups;
2007-06-26 15:48:44 +02:00
int n;
if (!my_group)
{
struct group *gr = getgrgid ( getgid () );
if (!gr)
gc_error (1, errno, "getgrgid failed for current user");
my_group = xstrdup (gr->gr_name);
}
if (!strcmp (group, my_group))
return 1; /* Found. */
if (!my_supgroups)
{
int ngids;
gid_t *gids;
ngids = getgroups (0, NULL);
gids = xcalloc (ngids+1, sizeof *gids);
ngids = getgroups (ngids, gids);
if (ngids < 0)
gc_error (1, errno, "getgroups failed for current user");
my_supgroups = xcalloc (ngids+1, sizeof *my_supgroups);
for (n=0; n < ngids; n++)
{
struct group *gr = getgrgid ( gids[n] );
if (!gr)
gc_error (1, errno, "getgrgid failed for supplementary group");
my_supgroups[n] = xstrdup (gr->gr_name);
}
xfree (gids);
}
for (n=0; my_supgroups[n]; n++)
if (!strcmp (group, my_supgroups[n]))
return 1; /* Found. */
}
2007-06-26 15:48:44 +02:00
#endif /*!HAVE_W32_SYSTEM*/
return 0; /* No match. */
}
/* Read and process the global configuration file for gpgconf. This
optional file is used to update our internal tables at runtime and
may also be used to set new default values. If FNAME is NULL the
default name will be used. With UPDATE set to true the internal
tables are actually updated; if not set, only a syntax check is
done. If DEFAULTS is true the global options are written to the
2007-10-23 20:13:27 +02:00
configuration files. If LISTFP is set, no changes are done but the
configuration file is printed to LISTFP in a colon separated format.
Returns 0 on success or if the config file is not present; -1 is
returned on error. */
int
2007-10-23 20:13:27 +02:00
gc_process_gpgconf_conf (const char *fname_arg, int update, int defaults,
estream_t listfp)
{
int result = 0;
char *line = NULL;
size_t line_len = 0;
ssize_t length;
gpgrt_stream_t config;
int lineno = 0;
int in_rule = 0;
int got_match = 0;
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
int runtime[GC_COMPONENT_NR] = { 0 };
int component_id;
2007-10-23 20:13:27 +02:00
char *fname;
2007-10-23 20:13:27 +02:00
if (fname_arg)
fname = xstrdup (fname_arg);
else
fname = make_filename (gnupg_sysconfdir (), GPGCONF_NAME EXTSEP_S "conf",
NULL);
config = gpgrt_fopen (fname, "r");
if (!config)
{
/* Do not print an error if the file is not available, except
2007-10-23 20:13:27 +02:00
when running in syntax check mode. */
if (errno != ENOENT || !update)
{
gc_error (0, errno, "can't open global config file '%s'", fname);
result = -1;
}
xfree (fname);
return result;
}
while ((length = gpgrt_read_line (config, &line, &line_len, NULL)) > 0)
{
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
char *key, *compname, *option, *flags, *value;
char *empty;
gc_option_t *option_info = NULL;
char *p;
int is_continuation;
lineno++;
key = line;
while (*key == ' ' || *key == '\t')
key++;
if (!*key || *key == '#' || *key == '\r' || *key == '\n')
continue;
is_continuation = (key != line);
/* Parse the key field. */
if (!is_continuation && got_match)
break; /* Finish after the first match. */
else if (!is_continuation)
{
in_rule = 0;
for (p=key+1; *p && !strchr (" \t\r\n", *p); p++)
;
if (!*p)
{
gc_error (0, 0, "missing rule at '%s', line %d", fname, lineno);
result = -1;
gpgconf_write_status (STATUS_WARNING,
"gpgconf.conf %d file '%s' line %d "
"missing rule",
GPG_ERR_SYNTAX, fname, lineno);
continue;
}
*p++ = 0;
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
compname = p;
}
else if (!in_rule)
{
gc_error (0, 0, "continuation but no rule at '%s', line %d",
fname, lineno);
result = -1;
continue;
}
else
{
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
compname = key;
key = NULL;
}
in_rule = 1;
/* Parse the component. */
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
while (*compname == ' ' || *compname == '\t')
compname++;
for (p=compname; *p && !strchr (" \t\r\n", *p); p++)
;
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
if (p == compname)
{
gc_error (0, 0, "missing component at '%s', line %d",
fname, lineno);
gpgconf_write_status (STATUS_WARNING,
"gpgconf.conf %d file '%s' line %d "
" missing component",
GPG_ERR_NO_NAME, fname, lineno);
result = -1;
continue;
}
empty = p;
*p++ = 0;
option = p;
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
component_id = gc_component_find (compname);
if (component_id < 0)
{
gc_error (0, 0, "unknown component at '%s', line %d",
fname, lineno);
gpgconf_write_status (STATUS_WARNING,
"gpgconf.conf %d file '%s' line %d "
"unknown component",
GPG_ERR_UNKNOWN_NAME, fname, lineno);
result = -1;
}
/* Parse the option name. */
while (*option == ' ' || *option == '\t')
option++;
for (p=option; *p && !strchr (" \t\r\n", *p); p++)
;
if (p == option)
{
gc_error (0, 0, "missing option at '%s', line %d",
fname, lineno);
gpgconf_write_status (STATUS_WARNING,
"gpgconf.conf %d file '%s' line %d "
"missing option",
GPG_ERR_INV_NAME, fname, lineno);
result = -1;
continue;
}
*p++ = 0;
flags = p;
if ( component_id != -1)
{
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
/* We need to make sure that we got the option list for the
* component. */
if (!gc_component[component_id].options)
gc_component_retrieve_options (component_id);
option_info = find_option (component_id, option);
if (!option_info)
{
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
gc_error (0, 0, "unknown option '%s' at '%s', line %d",
option, fname, lineno);
gpgconf_write_status (STATUS_WARNING,
"gpgconf.conf %d file '%s' line %d "
"unknown option",
GPG_ERR_UNKNOWN_OPTION, fname, lineno);
result = -1;
}
}
/* Parse the optional flags. */
while (*flags == ' ' || *flags == '\t')
flags++;
if (*flags == '[')
{
flags++;
p = strchr (flags, ']');
if (!p)
{
gc_error (0, 0, "syntax error in rule at '%s', line %d",
fname, lineno);
gpgconf_write_status (STATUS_WARNING,
"gpgconf.conf %d file '%s' line %d "
"syntax error in rule",
GPG_ERR_SYNTAX, fname, lineno);
result = -1;
continue;
}
*p++ = 0;
value = p;
}
else /* No flags given. */
{
value = flags;
flags = NULL;
}
/* Parse the optional value. */
while (*value == ' ' || *value == '\t')
value++;
for (p=value; *p && !strchr ("\r\n", *p); p++)
;
if (p == value)
value = empty; /* No value given; let it point to an empty string. */
else
{
/* Strip trailing white space. */
*p = 0;
for (p--; p > value && (*p == ' ' || *p == '\t'); p--)
*p = 0;
}
/* Check flag combinations. */
if (!flags)
;
else if (!strcmp (flags, "default"))
{
if (*value)
{
gc_error (0, 0, "flag \"default\" may not be combined "
"with a value at '%s', line %d",
fname, lineno);
result = -1;
}
}
else if (!strcmp (flags, "change"))
;
else if (!strcmp (flags, "no-change"))
;
else
{
gc_error (0, 0, "unknown flag at '%s', line %d",
fname, lineno);
result = -1;
}
2007-10-23 20:13:27 +02:00
/* In list mode we print out all records. */
if (listfp && !result)
{
/* If this is a new ruleset, print a key record. */
if (!is_continuation)
{
char *group = strchr (key, ':');
if (group)
{
*group++ = 0;
if ((p = strchr (group, ':')))
*p = 0; /* We better strip any extra stuff. */
}
es_fprintf (listfp, "k:%s:", gc_percent_escape (key));
es_fprintf (listfp, "%s\n", group? gc_percent_escape (group):"");
2007-10-23 20:13:27 +02:00
}
/* All other lines are rule records. */
es_fprintf (listfp, "r:::%s:%s:%s:",
gc_component[component_id].name,
option_info->name? option_info->name : "",
flags? flags : "");
2007-10-23 20:13:27 +02:00
if (value != empty)
es_fprintf (listfp, "\"%s", gc_percent_escape (value));
es_putc ('\n', listfp);
2007-10-23 20:13:27 +02:00
}
/* Check whether the key matches but do this only if we are not
running in syntax check mode. */
if ( update
2007-10-23 20:13:27 +02:00
&& !result && !listfp
&& (got_match || (key && key_matches_user_or_group (key))) )
{
int newflags = 0;
got_match = 1;
/* Apply the flags from gpgconf.conf. */
if (!flags)
;
else if (!strcmp (flags, "default"))
newflags |= GC_OPT_FLAG_DEFAULT;
else if (!strcmp (flags, "no-change"))
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
option_info->no_change = 1;
else if (!strcmp (flags, "change"))
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
option_info->no_change = 0;
if (defaults)
{
/* Here we explicitly allow updating the value again. */
if (newflags)
{
option_info->new_flags = 0;
}
if (*value)
{
xfree (option_info->new_value);
option_info->new_value = NULL;
}
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
change_one_value (component_id, option_info,
runtime, newflags, value, 0);
}
}
}
if (length < 0 || gpgrt_ferror (config))
{
gc_error (0, errno, "error reading from '%s'", fname);
result = -1;
}
if (gpgrt_fclose (config))
gc_error (0, errno, "error closing '%s'", fname);
xfree (line);
/* If it all worked, process the options. */
2007-10-23 20:13:27 +02:00
if (!result && update && defaults && !listfp)
{
/* We need to switch off the runtime update, so that we can do
it later all at once. */
int save_opt_runtime = opt.runtime;
opt.runtime = 0;
for (component_id = 0; component_id < GC_COMPONENT_NR; component_id++)
{
gc_component_change_options (component_id, NULL, NULL, 0);
}
opt.runtime = save_opt_runtime;
if (opt.runtime)
{
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
for (component_id = 0; component_id < GC_COMPONENT_NR; component_id++)
if (runtime[component_id]
&& gc_component[component_id].runtime_change)
(*gc_component[component_id].runtime_change) (0);
}
}
xfree (fname);
return result;
}
/*
* Apply the profile FNAME to all known configure files.
*/
gpg_error_t
gc_apply_profile (const char *fname)
{
gpg_error_t err;
char *fname_buffer = NULL;
char *line = NULL;
size_t line_len = 0;
ssize_t length;
estream_t fp;
int lineno = 0;
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
int runtime[GC_COMPONENT_NR] = { 0 };
int component_id = -1;
int skip_section = 0;
int error_count = 0;
int newflags;
if (!fname)
fname = "-";
if (!(!strcmp (fname, "-")
|| strchr (fname, '/')
#ifdef HAVE_W32_SYSTEM
|| strchr (fname, '\\')
#endif
|| strchr (fname, '.')))
{
/* FNAME looks like a standard profile name. Check whether one
* is installed and use that instead of the given file name. */
fname_buffer = xstrconcat (gnupg_datadir (), DIRSEP_S,
fname, ".prf", NULL);
if (!gnupg_access (fname_buffer, F_OK))
fname = fname_buffer;
}
fp = !strcmp (fname, "-")? es_stdin : es_fopen (fname, "r");
if (!fp)
{
err = gpg_error_from_syserror ();
log_error ("can't open '%s': %s\n", fname, gpg_strerror (err));
return err;
}
if (opt.verbose)
log_info ("applying profile '%s'\n", fname);
err = 0;
while ((length = es_read_line (fp, &line, &line_len, NULL)) > 0)
{
char *name, *flags, *value;
gc_option_t *option_info = NULL;
char *p;
lineno++;
name = line;
while (*name == ' ' || *name == '\t')
name++;
if (!*name || *name == '#' || *name == '\r' || *name == '\n')
continue;
trim_trailing_spaces (name);
/* Check whether this is a new section. */
if (*name == '[')
{
name++;
skip_section = 0;
/* New section: Get the name of the component. */
p = strchr (name, ']');
if (!p)
{
error_count++;
log_info ("%s:%d:%d: error: syntax error in section tag\n",
fname, lineno, (int)(name - line));
skip_section = 1;
continue;
}
*p++ = 0;
if (*p)
log_info ("%s:%d:%d: warning: garbage after section tag\n",
fname, lineno, (int)(p - line));
trim_spaces (name);
component_id = gc_component_find (name);
if (component_id < 0)
{
log_info ("%s:%d:%d: warning: skipping unknown section '%s'\n",
fname, lineno, (int)(name - line), name );
skip_section = 1;
}
continue;
}
if (skip_section)
continue;
if (component_id < 0)
{
error_count++;
log_info ("%s:%d:%d: error: not in a valid section\n",
fname, lineno, (int)(name - line));
skip_section = 1;
continue;
}
/* Parse the option name. */
for (p = name; *p && !spacep (p); p++)
;
*p++ = 0;
value = p;
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
option_info = find_option (component_id, name);
if (!option_info)
{
error_count++;
log_info ("%s:%d:%d: error: unknown option '%s' in section '%s'\n",
fname, lineno, (int)(name - line),
name, gc_component[component_id].name);
continue;
}
/* Parse the optional flags. */
trim_spaces (value);
flags = value;
if (*flags == '[')
{
flags++;
p = strchr (flags, ']');
if (!p)
{
log_info ("%s:%d:%d: warning: invalid flag specification\n",
fname, lineno, (int)(p - line));
continue;
}
*p++ = 0;
value = p;
trim_spaces (value);
}
else /* No flags given. */
flags = NULL;
/* Set required defaults. */
if (gc_arg_type[option_info->arg_type].fallback == GC_ARG_TYPE_NONE
&& !*value)
value = "1";
/* Check and save this option. */
newflags = 0;
if (flags && !strcmp (flags, "default"))
newflags |= GC_OPT_FLAG_DEFAULT;
if (newflags)
option_info->new_flags = 0;
if (*value)
{
xfree (option_info->new_value);
option_info->new_value = NULL;
}
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
change_one_value (component_id, option_info, runtime, newflags, value, 1);
}
if (length < 0 || es_ferror (fp))
{
err = gpg_error_from_syserror ();
error_count++;
log_error (_("%s:%u: read error: %s\n"),
fname, lineno, gpg_strerror (err));
}
if (es_fclose (fp))
log_error (_("error closing '%s'\n"), fname);
if (error_count)
log_error (_("error parsing '%s'\n"), fname);
xfree (line);
/* If it all worked, process the options. */
if (!err)
{
/* We need to switch off the runtime update, so that we can do
it later all at once. */
int save_opt_runtime = opt.runtime;
opt.runtime = 0;
for (component_id = 0; component_id < GC_COMPONENT_NR; component_id++)
{
gc_component_change_options (component_id, NULL, NULL, 1);
}
opt.runtime = save_opt_runtime;
if (opt.runtime)
{
gpgconf: Rewrite the gpgconf-comp module. * tools/gpgconf.h (gc_component_t): Change type to ... (gc_component_id_t): this. (GC_COMPONENT_ANY): New, so that we can use that in gpgconf-comp.c directly. * tools/gpgconf-comp.c: Major rework. -- The primary reason for this rework is to support the global options. A second reason is to clean up the code and simplify it so that we do not anymore need to maintain a list of options in the components _and_ in gpgconf-comp.c. What we do now is to 1. Read the option tables directly from the components using the new generic --dump-option-table option. This includes the header (group) descriptions. 2. Read the default values from the components as before using --gpgconf-list and update gpgconf's internal tables with that info. 3. Read the options using gpgrt_argparser in the same way as we do this in the components. The changes also do away with the second level notion of backends; they were only used for dirmngr's extra dirmngr_ldapservers.conf file. We intend to remove that file and replace it with a regular option so that it will be similar on how OpenPGP keyservers are specified. The whole thing will currently be slower than before (in particular on Windows) but we can optimize that by keeping a cached version of the option tables and the default values in a per homedir cache file. There is also some work planned to remove most of the data returned by --gpgconf-list. What can also be done is to replace the internal tables, which list the gpgconf maintainable options, by a configuration file so that admins are able to maintain the list of these options. GnuPG-bug-id: 4788 Signed-off-by: Werner Koch <wk@gnupg.org>
2020-03-12 10:16:56 +01:00
for (component_id = 0; component_id < GC_COMPONENT_NR; component_id++)
if (runtime[component_id]
&& gc_component[component_id].runtime_change)
(*gc_component[component_id].runtime_change) (0);
}
}
xfree (fname_buffer);
return err;
}