g10/keyring.c (keyring_search): Only mark the cache as completely
filled if we start the scan from the beginning of the keyring.
--
Signed-off-by: Neal H. Walfield <neal@g10code.com>
Reported-by: NIIBE Yutaka <gniibe@fsij.org>
A new feature (e8c53fc) turned up a bug whereby checking if a search
term matches multiple keys in the keyring causes the cache to be
inconsistent.
When we look for a key on the keyring, we iterate over each of the
keyblocks starting with the keyblock following the last result. For
each keyblock, we iterate over the public key and any subkeys. As we
iterate over each key, we first insert it into the cache and then
check if the key matches. If so, we are done.
In pseudo code:
for (i = last_result + 1; i < num_records; i ++)
keyblock = get_keyblock (i)
for (j = 1; j < len(keyblock); j ++)
key = keyblock[j]
update_cache (key)
if (compare (key, search_terms))
return ok
cache_filled = true
return ENOFOUND
When we look for the next match, we start with the following keyblock.
The result is that any subkeys following the key that matched are not
added to the cache (in other words, when a keyblock matches, the inner
loop did not necessarily complete and the subsequent search doesn't
resume it).
This patch includes a straightforward fix: only indicate the cache as
complete if we started the scan from the beginning of the keyring and
really didn't find anything.
* g10/trustdb.c (init_trustdb): If we can't read the trust model from
the trust DB, default to TM_PGP, not TM_TOFU_PGP.
--
Signed-off-by: Neal H. Walfield <neal@g10code.com>
* g10/tofu.c (opendbs): If the TOFU DB format is set to auto and there
is no TOFU DB, default to the flat format.
--
Signed-off-by: Neal H. Walfield <neal@g10code.com>
* g10/options.h (opt): Add field only_sign_text_ids.
* g10/gpg.c (enum cmd_and_opt_values): Add value oOnlySignTextIDs.
(opts): Handle oOnlySignTextIDs.
(main): Likewise.
* g10/keyedit.c (sign_uids): If OPT.ONLY_SIGN_TEXT_IDS is set, don't
select non-text based IDs automatically.
(keyedit_menu): Adapt the prompt asking to sign all user ids according
to OPT.ONLY_SIGN_TEXT_IDS.
* doc/gpg.texi: Document the new option --only-sign-text-ids.
--
Signed-off-by: Neal H. Walfield <neal@g10code.com>
GnuPG-bug-id: 1241
Debian-bug-id: 569702
* g10/gpg.c (check_user_ids): New function.
(main): Check that any user id specifications passed to --local-user
and --remote-user correspond to exactly 1 user. Check that any user
id specifications passed to --default-key correspond to at most 1
user. Warn if any user id specifications passed to --local-user or
--default-user are possible ambiguous (are not specified by long keyid
or fingerprint).
* g10/getkey.c (parse_def_secret_key): Don't warn about possible
ambiguous key descriptions here.
--
Signed-off-by: Neal H. Walfield <neal@g10code.com>
GnuPG-bug-id: 1128
Debian-debug-id: 544490
* g10/decrypt-data.c (decrypt_data): If OPT.UNWRAP_ENCRYPTION is set,
copy the data to the output file instead of continuing to process it.
* g10/gpg.c (enum cmd_and_opt_values): Add new value oUnwrap.
(opts): Handle oUnwrap.
(main): Likewise.
* g10/options.h (opt): Add field unwrap_encryption.
* g10/plaintext.c (handle_plaintext): Break the output file selection
functionality into ...
(get_output_file): ... this new function.
--
Signed-off-by: Neal H. Walfield <neal@g10code.com>
GnuPG-bug-id: 1060
Debian-bug-id: 282061
* g10/tofu.c (fingerprint_str): Die with the error code returned by
the failed function.
(time_ago_str): Ditto. Do not make a comma translatable.
(fingerprint_format): Use "%zu" for a size_t.
--
Also wrapped some long strings.
In general we should not use log_fatal or use xmalloc functions but
properly return an error code and use xtrymalloc like functions.
Signed-off-by: Werner Koch <wk@gnupg.org>
* g10/tofu.c (show_statistics): Also show when the most recently
signed message was observed.
--
Signed-off-by: Neal H. Walfield <neal@g10code.com>.
Suggested-by: MFPA <2014-667rhzu3dc-lists-groups@riseup.net>
* g10/tofu.c (show_statistics): Break the time delta to string code
into...
(time_ago_str): ... this new function.
--
Signed-off-by: Neal H. Walfield <neal@g10code.com>
* g10/tofu.c (fingerprint_pp): Split this function into...
(fingerprint_str): ... this function...
(fingerprint_format): ... and this function.
(record_binding): Store the unformatted fingerprint in the DB. Only
use the formatting fingerprint when displaying a message to the user.
(get_trust): Likewise.
(show_statistics): Likewise.
(tofu_register): Likewise.
(tofu_get_validity): Likewise.
(tofu_set_policy): Likewise.
(tofu_get_policy): Likewise.
--
Signed-off-by: Neal H. Walfield <neal@g10code.com>
* g10/import.c (transfer_secret_keys): Return GPG_ERR_NOT_PROCESSED
when stub_key_skipped.
(import_secret_one): Notify a user, suggesting --card-status.
--
Migration to 2.1 might be confusing with smartcard. With this patch,
a user can learn to run gpg ---card-status.
Thanks to intrigeri for the report.
Debian-bug-id: 795881
* g10/sqlite.c (sqlite3_stepx): When making sure that there is no
second SQL statement, ignore newlines.
--
Signed-off-by: Neal H. Walfield <neal@g10code.com>
* common/util.h (zb32_encode): Move prototype to ...
* common/zb32.h: new. Include this for all callers of zb32_encode.
Signed-off-by: Werner Koch <wk@gnupg.org>
* g10/trustdb.c (validate_keys): If tdbio_update_version_record fails,
RC does not contain the error code. Save the error code in rc2 and
use that.
--
Signed-off-by: Neal H. Walfield <neal@g10code.com>
* g10/keydb.c (keydb_rebuild_caches): Only mark the cached as prepared
if it is actually prepared, which it only is if the resource is a
keybox.
--
Signed-off-by: Neal H. Walfield <neal@g10code.com>
* g10/sqlite.h (enum sqlite_arg_type): Add SQLITE_ARG_BLOB.
(sqlite3_stepx_callback): New declaration.
(sqlite3_stepx): Change the callback's type to sqlite3_stepx_callback,
which passes an additional parameter, the sqlite3_stmt *. Update
users.
--
Signed-off-by: Neal H. Walfield <neal@g10code.com>
* g10/tofu.c (sqlite3_exec_printf): Move from here...
* g10/sqlite.c (sqlite3_exec_printf): ... to this new file. Don't
mark as static.
* g10/tofu.c (sqlite3_stepx): Move from here...
* g10/sqlite.c (sqlite3_stepx): ... to this new file. Don't
mark as static.
* g10/tofu.c (enum sqlite_arg_type): Move from here...
* g10/sqlite.h (enum sqlite_arg_type): ... to this new file.
--
Signed-off-by: Neal H. Walfield <neal@g10code.com>
* g10/mainproc.c (check_sig_and_print): Do not call the informational
get_validity if we are not going to use it.
Signed-off-by: Werner Koch <wk@gnupg.org>
* g10/main.h: Add rejection_shown flag to each weakhash struct
* g10/misc.c (print_digest_algo_note, additional_weak_digest): Do not
treat MD5 separately; (print_digest_rejected_note): Use
weakhash.rejection_shown instead of static shown.
* g10/options.h (opt): Change from additional_weak_digests to
weak_digests.
* g10/sig-check.c: Do not treat MD5 separately.
* g10/gpg.c (main): Explicitly set MD5 as weak.
* g10/gpgv.c (main): Explicitly set MD5 as weak.
--
Previously, only one weak digest rejection message was shown, of
whichever was the first type encountered. This meant that if "gpg
--weak-digest SHA224" encountered both an MD5 digest and a SHA224
digest, it would only show the user that the MD5 digest was rejected.
In order to let the user know which algorithms were rejected, we
needed to move the "shown" flag into a per-weak-algorithm location.
Given this additional complication, it made no sense to continue to
treat MD5 specially, so it is added as a default weak algorithm in the
same opt.weak_digests data structure as any other.
Signed-Off-By: Daniel Kahn Gillmor <dkg@fifthhorseman.net>
* g10/tofu.c: Include <sched.h>.
(batch_update_started): New variable.
(begin_transaction): If we've been in batch mode for a while, then
commit any extant batch transactions.
(tofu_begin_batch_update): If we are not in batch mode, initialize
batch_update_started.
--
Signed-off-by: Neal H. Walfield <neal@g10code.com>
* g10/tofu.c (struct db): Rename begin_transaction to savepoint_batch.
Rename end_transaction to savepoint_batch_commit. Update users.
Remove field rollback. Add fields savepoint_inner and
savepoint_inner_commit. Add field batch_update.
(dump_cache): New function.
(batch_update): New variable.
(begin_transaction). New function.
(end_transaction): New function.
(rollback_transaction): New function.
(tofu_begin_batch_update): New function.
(tofu_end_batch_update): New function.
(closedb): End any pending batch transaction.
(closedbs): Assert that none of the DBs have a started batch
transaction if we not in batch mode.
(record_binding): Use the begin_transaction, end_transaction and
rollback_transaction functions instead of including the SQL inline.
Also start a batch mode transaction if we are using the flat format.
(tofu_register): Use the begin_transaction, end_transaction and
rollback_transaction functions instead of including the SQL inline.
* g10/gpgv.c (tofu_begin_batch_update): New function.
(tofu_end_batch_update): New function.
* g10/test-stubs.c (tofu_begin_batch_update): New function.
(tofu_end_batch_update): New function.
--
Signed-off-by: Neal H. Walfield <neal@g10code.com>
* g10/tofu.c: Include <stdarg.h>.
(prepares_saved) [DEBUG_TOFU_CACHE]: New variable.
(queries) [DEBUG_TOFU_CACHE]: New variable.
(struct db): Add fields prevp, begin_transaction, end_transaction,
rollback, record_binding_get_old_policy, record_binding_update,
record_binding_update2, get_policy_select_policy_and_conflict,
get_trust_bindings_with_this_email, get_trust_gather_other_user_ids,
get_trust_gather_other_keys, register_already_seen, and
register_insert.
[DEBUG_TOFU_CACHE]: Add field hits.
(STRINGIFY): New macro.
(STRINGIFY2): New macro.
(enum sqlite_arg_type): New enum.
(sqlite3_stepx): New function.
(combined_db): Remove variable.
(opendb): Don't cache the combined db.
(struct dbs): New struct. Update users to use this as the head of the
local DB list rather than overloading struct db.
(unlink_db): New function.
(link_db): New function.
(db_cache): New variable.
(db_cache_count): New variable.
(DB_CACHE_ENTRIES): Define.
(getdb): If the dbs specific cache doesn't include the DB, look at
DB_CACHE. Only if that also doesn't include the DB open the
corresponding DB.
(closedb): New function.
(opendbs): Don't open the combined DB. Just return an initialized
struct dbs.
(closedbs): Don't close the dbs specific dbs. Attach them to the
front of DB_CACHE. If DB_CACHE contains more than DB_CACHE_ENTRIES,
close enough dbs from the end of the DB_CACHE list such that DB_CACHE
only contains DB_CACHE_ENTRIES. Don't directly close the dbs, instead
use the new closedb function.
[DEBUG_TOFU_CACHE]: Print out some statistics.
(record_binding): Use sqlite3_stepx instead of sqlite3_exec or
sqlite3_exec_printf.
(get_policy): Likewise.
(get_trust): Likewise.
(tofu_register): Likewise.
--
Signed-off-by: Neal H. Walfield <neal@g10code.com>
* common/srv.c: Merge into dirmngr/dns-stuff.c. Delete file.
* common/srv.h: Merge into dirmngr/dns-stuff.h. Delete file.
* common/Makefile.am (common_sources): Remove srv.c and srv.h.
* g10/keyserver.c: Do not include srv.h. The code using it is anyway
disabled.
* dirmngr/http.c: Remove header srv.h and stubs.
* dirmngr/t-dns-stuff.c: Add option --srv.
Signed-off-by: Werner Koch <wk@gnupg.org>
* g10/trustdb.c (init_trustdb): If the saved trust model is unknown,
default to tofu+pgp instead of pgp.
--
Signed-off-by: Neal H. Walfield <neal@g10code.com>
* g10/trustdb.c (validate_one_keyblock): When checking trust regular
expressions, treat the tofu+pgp trust model the same as the pgp trust
model.
--
Signed-off-by: Neal H. Walfield <neal@g10code.com>
* g10/tofu.c (get_trust): If the policy is auto or none, check if the
key is ultimately trusted. If so, return that.
(tofu_register): If the key is ultimately trusted, don't show any
statistics.
(tofu_get_validity): Likewise.
--
Signed-off-by: Neal H. Walfield <neal@g10code.com>
Suggested-by: Andre Heinecke <aheinecke@intevation.de>
* g10/trustdb.c (init_trustdb): Recognize tofu and tofu+pgp as
possibly saved trust models. Also register the ultimately trusted
keys if the trust model is tofu or tofu+pgp.
(check_trustdb): Don't skip if the trust model is tofu or tofu+pgp.
(update_trustdb): Likewise.
(tdb_check_trustdb_stale): Likewise.
(validate_keys): If the trust model is TOFU, just write out the
ultimately trusted keys.
--
Signed-off-by: Neal H. Walfield <neal@g10code.com>
* g10/tofu.c (initdb): Make the version check and the database
initialization atomic.
--
Signed-off-by: Neal H. Walfield <neal@g10code.com>
Co-authored-by: Andre Heinecke <aheinecke@intevation.de>
* configure.ac: Add option --dsiable-tofu and --disable-sqlite.
(NEED_SQLITE_VERSION): New var.
(USE_TOFU): New ac_define and am_conditional.
* autogen.sh (build-w32): Add PKG_CONFIG_LIBDIR to configure so that
pkg-config find the correct .pc file.
* g10/Makefile.am (tofu_source): New. Build only if enabled.
* g10/gpg.c (parse_trust_model)[!USE_TOFU]: Disable tofu models.
(parse_tofu_policy)[!USE_TOFU]: Disable all.
(parse_tofu_db_format)[!USE_TOFU]: Disable all.
(main) <aTOFUPolicy>[!USE_TOFU]: Skip.
* g10/keyedit.c (show_key_with_all_names_colon)[!USE_TOFU]: Do not
call tofu functions.
* g10/keylist.c (list_keyblock_colon)[!USE_TOFU]: Ditto.
* g10/trustdb.c (tdb_get_validity_core)[!USE_TOFU]: Skip tofu
processing.
--
This allows to build a minimal version of GnuPG. It is also currently
required to build for Windows.
Signed-off-by: Werner Koch <wk@gnupg.org>
