- 12 8月, 2015 3 次提交
-
-
由 Ansgar Burchardt 提交于
-
由 Ansgar Burchardt 提交于
-
由 Ansgar Burchardt 提交于
I really would like to get rid of mapped components on security-master...
-
- 11 8月, 2015 1 次提交
-
-
由 Ansgar Burchardt 提交于
-
- 18 6月, 2015 3 次提交
-
-
由 Niels Thykier 提交于
Signed-off-by: NNiels Thykier <niels@thykier.net>
-
由 Ansgar Burchardt 提交于
Letting Python close the file will also release the lock after having flushed all buffers. Reported-by: NNiels Thykier <niels@thykier.net>
-
由 Niels Thykier 提交于
Signed-off-by: NNiels Thykier <niels@thykier.net>
-
- 16 6月, 2015 1 次提交
-
-
由 Ansgar Burchardt 提交于
session_wrapper gets confused by generators: let "f" be a generator making use of a database session: def wrapped_f(): session = DBConn().session() try: return f(session) finally: session.close() for item in wrapped_f(): do_something(item) will leak database sessions as the implementation of "f" will only be invoked from the for-loop. However at this time the "finally" block closing the session has already run. SQLAlchemy will helpfully reopen the session, but it will not be closed in the end.
-
- 14 6月, 2015 3 次提交
-
-
由 Ansgar Burchardt 提交于
-
由 Ansgar Burchardt 提交于
-
由 Ansgar Burchardt 提交于
Bug: https://bugs.debian.org/788669
-
- 12 6月, 2015 4 次提交
-
-
由 Ansgar Burchardt 提交于
-
由 Ansgar Burchardt 提交于
-
由 Ansgar Burchardt 提交于
This would need addition of the codename to the "package_list" view. This reverts commit 87087596.
-
由 Ansgar Burchardt 提交于
-
- 11 6月, 2015 3 次提交
-
-
由 Niels Thykier 提交于
Signed-off-by: NNiels Thykier <niels@thykier.net>
-
由 Niels Thykier 提交于
Signed-off-by: NNiels Thykier <niels@thykier.net>
-
由 Niels Thykier 提交于
Signed-off-by: NNiels Thykier <niels@thykier.net>
-
- 10 6月, 2015 1 次提交
-
-
由 Ansgar Burchardt 提交于
This patch removes the old implementation of generate-packages-sources that used apt-ftparchive instead of the metadata in the Postgres database.
-
- 09 6月, 2015 5 次提交
-
-
由 Ansgar Burchardt 提交于
-
由 Ansgar Burchardt 提交于
-
由 Niels Thykier 提交于
Which doubles as a fitting disclaimer too. Signed-off-by: NNiels Thykier <niels@thykier.net>
-
由 Niels Thykier 提交于
Add a ReverseDependencyChecker class for bulk testing breakage in reverse dependencies and use it in the auto-decrufter. At this point, disable the NBS removal - it will be re-added in the next commit. Signed-off-by: NNiels Thykier <niels@thykier.net>
-
由 Niels Thykier 提交于
Signed-off-by: NNiels Thykier <niels@thykier.net>
-
- 08 6月, 2015 2 次提交
-
-
由 Niels Thykier 提交于
Signed-off-by: NNiels Thykier <niels@thykier.net>
-
由 Niels Thykier 提交于
Avoid creating collections that are thrown away after a single (or even zero) uses. Signed-off-by: NNiels Thykier <niels@thykier.net>
-
- 06 6月, 2015 3 次提交
-
-
由 Niels Thykier 提交于
Signed-off-by: NNiels Thykier <niels@thykier.net>
-
由 Niels Thykier 提交于
Signed-off-by: NNiels Thykier <niels@thykier.net>
-
由 Niels Thykier 提交于
Signed-off-by: NNiels Thykier <niels@thykier.net>
-
- 29 5月, 2015 2 次提交
-
-
由 Ansgar Burchardt 提交于
-
由 Ansgar Burchardt 提交于
-
- 19 5月, 2015 1 次提交
-
-
由 Ansgar Burchardt 提交于
CACERTFILE now must point to a CA, using service certificates directly doesn't work. Also use per-connection settings for TLS. Note that the order of set_option calls matters (setting ldap.OPT_X_TLS_NEWCTX first fails)...
-
- 16 5月, 2015 1 次提交
-
-
由 Ansgar Burchardt 提交于
apt_pkg.parse_depends does not handle all features allowed in the Build-Depends(-Indep) fields. In particular it gives an error when Build-Depends contains restriction formulas as used for build profiles.
-
- 18 3月, 2015 1 次提交
-
-
由 Ansgar Burchardt 提交于
-
- 04 2月, 2015 1 次提交
-
-
由 Ansgar Burchardt 提交于
Also inherit from Reject so a proper rejection message is shown instead of a backtrace.
-
- 06 12月, 2014 1 次提交
-
-
由 Joerg Jaspert 提交于
Signed-off-by: NJoerg Jaspert <joerg@debian.org>
-
- 27 11月, 2014 4 次提交
-
-
由 Ansgar Burchardt 提交于
-
由 Ansgar Burchardt 提交于
-
由 Ansgar Burchardt 提交于
-
由 Ansgar Burchardt 提交于
-