- 06 5月, 2017 7 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
-
由 Ansgar Burchardt 提交于
-
由 Ansgar Burchardt 提交于
-
由 Ansgar Burchardt 提交于
-
- 25 4月, 2017 2 次提交
-
-
由 Ansgar Burchardt 提交于
-
由 Ansgar Burchardt 提交于
Use OpenSSH's `ControlPersist` option instead of implementing an approximation ourselves. This avoids an infinite sleep in case the `ssh` call fails and the socket never appears.
-
- 08 3月, 2017 1 次提交
-
-
由 Ansgar Burchardt 提交于
-
- 05 3月, 2017 1 次提交
-
-
由 Joerg Jaspert 提交于
-
- 04 3月, 2017 2 次提交
-
-
由 Joerg Jaspert 提交于
-
由 Joerg Jaspert 提交于
-
- 23 2月, 2017 5 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
-
由 Ansgar Burchardt 提交于
-
- 21 2月, 2017 2 次提交
-
-
由 Julien Cristau 提交于
Rather than keying on the byhash boolean, make Release always a symlink to zzz-dists. Signed-off-by: NJulien Cristau <jcristau@debian.org>
-
由 Julien Cristau 提交于
The byhash boolean now controls: - whether Release lives in zzz-dists (should we change this too?) - whether Acquire-by-hash is advertised in Release This should help speed up the second stage of mirroring. Signed-off-by: NJulien Cristau <jcristau@debian.org>
-
- 14 2月, 2017 1 次提交
-
-
由 Ansgar Burchardt 提交于
-
- 13 2月, 2017 2 次提交
-
-
由 Joerg Jaspert 提交于
-
由 Joerg Jaspert 提交于
-
- 24 1月, 2017 3 次提交
-
-
由 Ansgar Burchardt 提交于
-
由 Ansgar Burchardt 提交于
-
由 Ansgar Burchardt 提交于
-
- 15 1月, 2017 2 次提交
-
-
由 Ansgar Burchardt 提交于
This silences a warning from newer SQLAlchemy releases: SAWarning: Textual SQL expression '[...]' should be explicitly declared as text('[...]')
-
由 Ansgar Burchardt 提交于
-
- 14 1月, 2017 1 次提交
-
-
由 Joerg Jaspert 提交于
-
- 13 1月, 2017 2 次提交
-
-
由 Ansgar Burchardt 提交于
Making the same error when adding the file locally in /usr/share/dpkg for testing doesn't help with noticing this.
-
由 Ansgar Burchardt 提交于
-
- 07 1月, 2017 1 次提交
-
-
由 Joerg Jaspert 提交于
That doesn't store original name/timestamp, with the latter meaning a change even if there wasn't one in the input. Ought to help systems like snapshot.
-
- 05 1月, 2017 3 次提交
-
-
由 Ansgar Burchardt 提交于
-
由 Ansgar Burchardt 提交于
-
由 Ansgar Burchardt 提交于
While not really relevant, this change stops `check-override` from trying to sync the source package's override with every run (as the priority continued to differ).
-
- 22 12月, 2016 2 次提交
-
-
由 Chris Lamb 提交于
-
由 Chris Lamb 提交于
-
- 16 12月, 2016 1 次提交
-
-
由 Chris Lamb 提交于
Signed-off-by: NChris Lamb <lamby@debian.org>
-
- 08 12月, 2016 1 次提交
-
-
由 Ansgar Burchardt 提交于
-
- 07 12月, 2016 1 次提交
-
-
由 Ansgar Burchardt 提交于
When uploading using scp, the mtime might be set to the time from the uploader's host. Especially for the upstream tarballs this might result in a mtime a long time in the past and queued might remove the uploaded files right away as stray files. Note that this removal only happens when the file is not processed: for example when the upstream tarball gets uploaded (with an old mtime) and queued processes uploads and removes stray files before the .changes for the upload including the "old" tarball arrives.
-