--
WIN32_LEAN_AND_MEAN is required to avoid definitions of grp1, grp2,
and grp3 in dlgs.h, which is included by windows.h.
Fixes-commit: fd595c9d36
Signed-off-by: NIIBE Yutaka <gniibe@fsij.org>
* g10/keydb.h (KEYDB_HANDLE): Move typedef to ...
* g10/gpg.h: here.
(struct server_control_s): Add field 'cached_getkey_kdb'.
* g10/gpg.c (gpg_deinit_default_ctrl): Release that keydb handle.
* g10/getkey.c (getkey_end): Cache keydb handle.
(get_pubkey): Use cached keydb handle.
* kbx/keybox-search.c (keybox_search_reset): Use lseek instead of
closing the file.
--
Before this patch a "gpg --check-sigs" opened and closed the keybox
file for almost every signature check. By caching the keydb handle
and using lseek(2) this can be limited to just 2 times. This might
speed up things on Windows.
Signed-off-by: Werner Koch <wk@gnupg.org>
--
kbnode_t has replaced KBNODE for new code years ago, but that should
be documented. No bulk changes please to keep git blame easy to read.
Signed-off-by: Werner Koch <wk@gnupg.org>
* g10/gpg.h (tofu): Rename field batch_update_ref to
batch_updated_wanted.
* g10/tofu.c (struct tofu_dbs_s): Rename field batch_update to
in_batch_transaction.
(begin_transaction): Only end an extant batch transaction if we are
not in a normal transaction. When ending a batch transaction, really
end it. Update ctrl->tofu.batch_update_started when starting a batch
transaction.
(end_transaction): Only release a batch transaction if ONLY_BATCH is
true. When releasing a batch transaction, assert that there is no
open normal transaction. Only allow DBS to be NULL if ONLY_BATCH is
true.
(tofu_begin_batch_update): Don't update
ctrl->tofu.batch_update_started.
(opendbs): Call end_transaction unconditionally.
--
Signed-off-by: Neal H. Walfield <neal@g10code.com>
* g10/gpg.h (struct server_control_s): New field in_transaction.
* g10/tofu.c (struct tofu_dbs_s): Remove fields savepoint_inner and
savepoint_inner_commit.
(begin_transaction): Increment CTRL->TOFU.IN_TRANSACTION. Name the
savepoint according to the nesting level.
(end_transaction): Name the savepoint according to the nesting level.
Decrement CTRL->TOFU.IN_TRANSACTION.
(rollback_transaction): Likewise. Only ever rollback a non-batch
transaction.
(opendbs): Assert that there are no outstanding transactions.
--
Signed-off-by: Neal H. Walfield <neal@g10code.com>
* g10/options.h (struct opt): Remove field tofu_db_format.
* g10/gpg.h (server_control_s): Add fields tofu.batch_update_ref and
tofu.batch_update_started.
* g10/gpg.c (parse_tofu_db_format): Remove.
(main): Make option --tofu-db-format obsolete.
* g10/tofu.c: Major rework. Remove the pretty complicated and slower
split format and with that all the caching. Use the dbs struct
directly. Move global vars for batch update into CTRL. Change
calling conventions of some function to take CTRL or DBS pointers
instead of the former low-level database pointer.
--
The split database format might have been nice for use with Unison but
it bypasses the concept of a relational database by doing parts of
this itself and also risking deadlocks. Working with the Tofu
database for debugging or experiments is also not possible with parts
of the database logic implemented in gpg.
The Tofu support is quite new and we can assume that it is not in real
use now. Thus we better remove that now so that we do not need to
maintain it for all future.
Signed-off-by: Werner Koch <wk@gnupg.org>
* g10/gpg.h (struct tofu_dbs_s, tofu_dbs_t): New declarations.
(struct server_control_s): Add field tofu.dbs.
* g10/tofu.c (struct dbs): Rename to tofu_dbs_s. Replace all users by
by tofu_dbs_t.
(opendbs): Add arg CTRL. Cache the DBS in CTRL.
(closedbs): Rename to tofu_closedbs and make global. Add arg CTRL.
(tofu_register): Add arg CTRL. Change all callers. Do not call
closedbs.
(tofu_get_validity): Ditto.
(tofu_set_policy): Ditto.
(tofu_get_policy): Ditto.
(tofu_set_policy_by_keyid): Add arg CTRL.
* g10/gpg.c (gpg_deinit_default_ctrl): Call tofu_closedbs.
Signed-off-by: Werner Koch <wk@gnupg.org>
* g10/tofu.c (fingerprint_str): Remove.
(tofu_register): Take a public key instead of a fingerprint as arg.
Use hexfingerprint() to get a fpr from the PK.
(tofu_get_validity): Ditto.
(tofu_set_policy, tofu_get_policy): Simplify by using hexfingerprint.
* g10/trustdb.c (tdb_get_validity_core): Pass the primary key PK to
instead of the fingerprint to the tofu functions.
--
This change has the advantage that we are not bound to a specific
fingerprint length and will thus helps us to implement rfc4880bis.
Signed-off-by: Werner Koch <wk@gnupg.org>
* g10/gpg.h (g10_errstr): Remove macro and change all occurrences by
gpg_strerror.
(G10ERR_): Remove all macros and change all occurrences by their
GPG_ERR_ counterparts.
Signed-off-by: Werner Koch <wk@gnupg.org>
* g10/getkey.c (keyid_list): Add field "fpr".
(cache_user_id): Store fpr and check for dups only by fpr.
(get_pubkey_byfpr): New.
(get_user_id_string): Make static and use xasprintf.
(get_long_user_id_string): Use xasprintf.
(get_user_id_byfpr): New.
(get_user_id_byfpr_native): New.
* g10/keyid.c (fingerprint_from_pk): Make arg RET_LEN optional.
* g10/import.c (import_one): Use get_user_id_byfpr_native.
--
We now cache the userids using the fingerprint. This allows to print
the correct user id for keys with a duplicated key id. We should
eventually start to retire the use of all the old keyid based
functions. However, at some places we only have the keyid and thus
some of them will need to be kept (maybe changed with an indication to
show that more than several user ids are matching).
Signed-off-by: Werner Koch <wk@gnupg.org>
We better do this once and for all instead of cluttering all future
commits with diffs of trailing white spaces. In the majority of cases
blank or single lines are affected and thus this change won't disturb
a git blame too much. For future commits the pre-commit scripts
checks that this won't happen again.
to libgcrypt functions, using shared error codes from libgpg-error,
replacing the old functions we used to have in ../util by those in
../jnlib and ../common, renaming the malloc functions and a couple of
types. Note, that not all changes are listed below becuause they are
too similar and done at far too many places. As of today the code
builds using the current libgcrypt from CVS but it is very unlikely
that it actually works.
