daaf1305f641dab34099e98cef38cb4a0bf01fa3: ansible: allow basic robots.txt indexing (bug 1234971)
Gregory Szorc <gps@mozilla.com> - Mon, 04 Jan 2016 12:29:37 -0800 - rev 362039
Push 16998 by rwood@mozilla.com at Mon, 02 May 2016 19:42:03 +0000
ansible: allow basic robots.txt indexing (bug 1234971) Our previous robots.txt disallowed all indexing. We now open it up slightly. We should keep an eye on robot requests, as I have a feeling that policy is still a bit too liberal and some expensive page loads could creep in. reviewboard-hg didn't have a robots.txt policy. The server started getting indexed today and it was querying expensive URLs like annotate. Deploy the same robot policy to hopefully curtail this.
3ae14cdbabf3e771ac7ca5a0896b6f8cf97b2f4b: mozreview: disable autopublish in selenium tests (Bug 1231159). r=gps
Steven MacLeod <smacleod@mozilla.com> - Sun, 03 Jan 2016 17:26:26 -0500 - rev 362038
Push 16998 by rwood@mozilla.com at Mon, 02 May 2016 19:42:03 +0000
mozreview: disable autopublish in selenium tests (Bug 1231159). r=gps Recently we started prompting to publish review requests even if they didn't have reviewers. This caused breakages in the selenium tests as we would always respond yes. Disabling autopublishing fixes the test breakages.
7c045be0008475204f9cde64952b1d64ab65219c: mozreview: fixup tests for RB 2.5.2 (Bug 1231159). r=gps
Steven MacLeod <smacleod@mozilla.com> - Wed, 30 Dec 2015 12:49:35 -0500 - rev 362037
Push 16998 by rwood@mozilla.com at Mon, 02 May 2016 19:42:03 +0000
mozreview: fixup tests for RB 2.5.2 (Bug 1231159). r=gps
f79b251db6d4fb9f06583445fcd8703da6bcc1d5: mozreview: upgrade to RB 2.5.2 (Bug 1231159). r=gps
Steven MacLeod <smacleod@mozilla.com> - Tue, 08 Dec 2015 13:48:23 -0500 - rev 362036
Push 16998 by rwood@mozilla.com at Mon, 02 May 2016 19:42:03 +0000
mozreview: upgrade to RB 2.5.2 (Bug 1231159). r=gps Due to changes in Review Board's static media packaging code we now run with DEBUG=False. This is required for Review Board to gather the static media from our extensions. We also switch to using supported means for installing and loading static media which may be slower but works and should continue to work with future versions.
6621bd075be45c5c42b83886e95f998caae999b6: ansible/hg-reviewboard: render git diffs (bug 1235254); r=smacleod
Gregory Szorc <gps@mozilla.com> - Thu, 31 Dec 2015 15:12:58 -0800 - rev 362035
Push 16998 by rwood@mozilla.com at Mon, 02 May 2016 19:42:03 +0000
ansible/hg-reviewboard: render git diffs (bug 1235254); r=smacleod This is necessary for binary file changes to be rendered in the diff output. I suppose this has the potential to break ReviewBoard's parsing of Mercurial diffs, since I /think/ ReviewBoard fetches raw-rev URLs when expanding diff context. It should be smart enough to parse git style diffs.
d6a1e14868f98e6381db85941fba0251d1616159: firefoxtree: check for firefoxtree attribute before accessing (bug 1234396); r=smacleod
Gregory Szorc <gps@mozilla.com> - Mon, 21 Dec 2015 16:55:50 -0800 - rev 362034
Push 16998 by rwood@mozilla.com at Mon, 02 May 2016 19:42:03 +0000
firefoxtree: check for firefoxtree attribute before accessing (bug 1234396); r=smacleod I'm not sure what's causing the attribute to get lost. I can reproduce it on my machine but I can't reproduce it in the test environment. Let's work around it for now until a better solution can be implemented. (The quick work around is important because this prevents people from cloning.)
8e1fc3cd3668fa084f6c2996dccfcb6b8c4b8851: vcsreplicator: fix typo
Gregory Szorc <gps@mozilla.com> - Mon, 04 Jan 2016 11:23:52 -0800 - rev 362033
Push 16998 by rwood@mozilla.com at Mon, 02 May 2016 19:42:03 +0000
vcsreplicator: fix typo
c8b6fba8c40eed8a9d0e5fbf7807ac058e1cb989: docs: fix typo
Gregory Szorc <gps@mozilla.com> - Mon, 04 Jan 2016 11:23:20 -0800 - rev 362032
Push 16998 by rwood@mozilla.com at Mon, 02 May 2016 19:42:03 +0000
docs: fix typo
ac87b427e57cfbe79d617eb3ee7e08397e211786: rbbz: Include full commit message when posting attachments (bug 1232894); r=smacleod
Mark Cote <mcote@mozilla.com> - Wed, 30 Dec 2015 20:03:35 -0500 - rev 362031
Push 16998 by rwood@mozilla.com at Mon, 02 May 2016 19:42:03 +0000
rbbz: Include full commit message when posting attachments (bug 1232894); r=smacleod The old code used to use changedesc, but that doesn't seem to be set nowadays, or at least not consistently. Use the 'description' field now, stripping off the first line, since it's in the attachment summary.
a6cba09c935e694588bb619e40c73b69bbd16073: Bug 1228358 - Release version pushlog_client 0.3.0
Armen Zambrano Gasparnian <armenzg@mozilla.com> - Mon, 04 Jan 2016 12:43:21 -0500 - rev 362030
Push 16998 by rwood@mozilla.com at Mon, 02 May 2016 19:42:03 +0000
Bug 1228358 - Release version pushlog_client 0.3.0
1fe80bd2bd583a6407f6cbb021766a1f9fa32bbd: hghooks: use underscores in verify css image only approvals hook
Gregory Szorc <gps@mozilla.com> - Thu, 31 Dec 2015 15:27:03 -0800 - rev 362029
Push 16998 by rwood@mozilla.com at Mon, 02 May 2016 19:42:03 +0000
hghooks: use underscores in verify css image only approvals hook For consistency. And underscores are more appropriate for Python modules. I should have spotted this during review.
acb5c03328a4c04b66087e4525c76decea264d58: ansible/hg-web: enable ExtendedStatus
Gregory Szorc <gps@mozilla.com> - Thu, 31 Dec 2015 12:31:41 -0800 - rev 362028
Push 16998 by rwood@mozilla.com at Mon, 02 May 2016 19:42:03 +0000
ansible/hg-web: enable ExtendedStatus So we can see which URLs are being requested.
c3bddd76912f81dc5e21a256d3c57ce14d9ed078: check_kafka_hg_heartbeat: rewrite to not use `hg sendheartbeat`
Gregory Szorc <gps@mozilla.com> - Thu, 31 Dec 2015 11:26:32 -0800 - rev 362027
Push 16998 by rwood@mozilla.com at Mon, 02 May 2016 19:42:03 +0000
check_kafka_hg_heartbeat: rewrite to not use `hg sendheartbeat` This check has been failing periodically in production for unknown reasons. The solution for unwedging a partition is often to restart the leader node. We desire to print broker information in check output to help with diagnosing and correcting the issue. While we could put this in `hg sendheartbeat`, it feels out of place there. `hg sendheartbeat` was really just a convenience to avoid having to reimplement some code. This commit reimplements `hg sendheartbeat` in the check_kafka_hg_heartbeat Nagios check. By doing so, we can formulate output to contain more details. We now print the failing partition and its leader host in the failure output, allowing easier diagnosis of failing partitions. There is still room to improve the output. This will be done later.
a90496b82b78450c94c8fa67cf4d9a4e242d359a: vcsreplicator: mark as compatible against Mercurial 3.6
Gregory Szorc <gps@mozilla.com> - Thu, 31 Dec 2015 09:56:12 -0800 - rev 362026
Push 16998 by rwood@mozilla.com at Mon, 02 May 2016 19:42:03 +0000
vcsreplicator: mark as compatible against Mercurial 3.6
525d1a8751c8868c0751c698714d10c2948590c8: check_zookeeper: catch ZookeeperUnavailable
Gregory Szorc <gps@mozilla.com> - Thu, 31 Dec 2015 09:54:08 -0800 - rev 362025
Push 16998 by rwood@mozilla.com at Mon, 02 May 2016 19:42:03 +0000
check_zookeeper: catch ZookeeperUnavailable This has occurred a few times in production. Not sure how to test for it.
834cf197846c11e655bd06b7144e2897ae0f6afa: Bug 1231768 - add hook for css-image-only-approvals, r=gps
Gijs Kruitbosch <gijskruitbosch@gmail.com> - Tue, 15 Dec 2015 16:59:01 +0000 - rev 362024
Push 16998 by rwood@mozilla.com at Mon, 02 May 2016 19:42:03 +0000
Bug 1231768 - add hook for css-image-only-approvals, r=gps
c1fa09f2886de8f435cce2575c392b21efe8b9f0: testing: refactor Review Board user creation order (bug 1228002)
Gregory Szorc <gps@mozilla.com> - Wed, 23 Dec 2015 17:53:11 -0800 - rev 362023
Push 16998 by rwood@mozilla.com at Mon, 02 May 2016 19:42:03 +0000
testing: refactor Review Board user creation order (bug 1228002) I think there was a race condition between creating users and configuring URLs in Review Board that caused the mozreview user to sometimes not get created. Reshuffling the order of operations to force user creation to run first, before URL setting, should hopefully make the problem go away.
2814ab81d93873829a25c1f839bd571b8bb46d72: testing: make query argument to `reviewboard get-users` optional
Gregory Szorc <gps@mozilla.com> - Wed, 23 Dec 2015 17:38:05 -0800 - rev 362022
Push 16998 by rwood@mozilla.com at Mon, 02 May 2016 19:42:03 +0000
testing: make query argument to `reviewboard get-users` optional This allows us to dump the state of the user database. This will enable me to track down intermitten failures with test-auth.t.
dd3a6f0eab5b54ac8b0bd9e123ede7a6b8b2974b: reviewboard: drop support for Mercurial 3.1 and 3.2
Gregory Szorc <gps@mozilla.com> - Tue, 22 Dec 2015 12:08:50 -0800 - rev 362021
Push 16998 by rwood@mozilla.com at Mon, 02 May 2016 19:42:03 +0000
reviewboard: drop support for Mercurial 3.1 and 3.2 Our policy is to only support the last 4 releases of Mercurial.
c1713e8b39809dc715556dda04e9bde3cf153bed: ansible: install Mercurial 3.6.2 RPM
Gregory Szorc <gps@mozilla.com> - Tue, 22 Dec 2015 11:55:13 -0800 - rev 362020
Push 16998 by rwood@mozilla.com at Mon, 02 May 2016 19:42:03 +0000
ansible: install Mercurial 3.6.2 RPM While we run all production processes out of virtualenvs, it is useful to have `hg` in the system PATH so people can SSH in and do useful things without having to muck about with paths. Let's upgrade the RPM package to something modern.
(0) -300000 -100000 -30000 -10000 -3000 -1000 -300 -100 -50 -20 +20 +50 +100 +300 +1000 +3000 +10000 +30000 +100000 +300000 tip