a0c2ee2f9c71850c65f8464a8936f0ddffbd5201: ansible/mozreview-dev: fix indentation and path reference
Gregory Szorc <gps@mozilla.com> - Tue, 01 Mar 2016 14:26:28 -0800 - rev 362347
Push 16998 by rwood@mozilla.com at Mon, 02 May 2016 19:42:03 +0000
ansible/mozreview-dev: fix indentation and path reference Indentation was causing a YAML failure. The path to virtualenv.yml was also wrong. This unblocks deploying to mozreview-dev.
807656c730ee92a79308ae2c38c6abbd8164bade: ansible/mozreview-dev: use proper python27-devel package name
Gregory Szorc <gps@mozilla.com> - Tue, 01 Mar 2016 14:25:54 -0800 - rev 362346
Push 16998 by rwood@mozilla.com at Mon, 02 May 2016 19:42:03 +0000
ansible/mozreview-dev: use proper python27-devel package name
238a4f5dd46d802cb350a40d95969682481d7163: mozreview: update Django and Djblets for security releases (bug 1251969)
Gregory Szorc <gps@mozilla.com> - Tue, 01 Mar 2016 14:15:33 -0800 - rev 362345
Push 16998 by rwood@mozilla.com at Mon, 02 May 2016 19:42:03 +0000
mozreview: update Django and Djblets for security releases (bug 1251969)
12e31fae88f4705036de3c657464737088ef9787: hghooks: allow pushing to try without a try syntax (bug 1247072), r=gps
Andrew Halberstadt <ahalberstadt@mozilla.com> - Thu, 25 Feb 2016 09:33:31 -0500 - rev 362344
Push 16998 by rwood@mozilla.com at Mon, 02 May 2016 19:42:03 +0000
hghooks: allow pushing to try without a try syntax (bug 1247072), r=gps Try syntax is no longer required to schedule jobs on try. This can now be accomplished directly in treeherder (or via other mozci based tools). This changes the try hook to allow pushes with no try syntax. For now, print a verbose WARNING so people understand what's happening. We may want to tone the warning down a bit if this workflow becomes more standard in the future. MozReview-Commit-ID: AKgkBpSxHdd
bdb954cc31aa180a1e0b46a60629b68cfc81b8d6: ansible/hg-ssh: produce bundles for mozilla-esr45 (bug 1250609)
Gregory Szorc <gps@mozilla.com> - Tue, 01 Mar 2016 11:04:36 -0800 - rev 362343
Push 16998 by rwood@mozilla.com at Mon, 02 May 2016 19:42:03 +0000
ansible/hg-ssh: produce bundles for mozilla-esr45 (bug 1250609)
df1f27cb9bea819d4a3bc6e5f2badb7da0a2e982: testing: add hash for setuptools
byron jones <glob@mozilla.com> - Tue, 01 Mar 2016 14:31:55 +0800 - rev 362342
Push 16998 by rwood@mozilla.com at Mon, 02 May 2016 19:42:03 +0000
testing: add hash for setuptools
08dae6a7bf9171df00a3152dace8904fc1598f08: testing: bump 'futures' version to fix requirement parsing error
byron jones <glob@mozilla.com> - Tue, 01 Mar 2016 14:30:54 +0800 - rev 362341
Push 16998 by rwood@mozilla.com at Mon, 02 May 2016 19:42:03 +0000
testing: bump 'futures' version to fix requirement parsing error
df66bc766888bf7c52b57fca6cbe5de342116d0a: bundleclone: watch for premature server termination in tests
Gregory Szorc <gps@mozilla.com> - Mon, 29 Feb 2016 18:31:29 -0800 - rev 362340
Push 16998 by rwood@mozilla.com at Mon, 02 May 2016 19:42:03 +0000
bundleclone: watch for premature server termination in tests Without this, the test can spin forever if the server fails to start. MozReview-Commit-ID: 5Vap62ddAhA
1a974e3e4f236c44d6c589a81ac3edab5e2e9685: testing: add hghave checks for hg37+
Gregory Szorc <gps@mozilla.com> - Mon, 29 Feb 2016 18:18:10 -0800 - rev 362339
Push 16998 by rwood@mozilla.com at Mon, 02 May 2016 19:42:03 +0000
testing: add hghave checks for hg37+ MozReview-Commit-ID: INK0YCEdlsS
b41681c581ea80620d2e7368decfffd30f2601f6: testing: upgrade run-tests.py
Gregory Szorc <gps@mozilla.com> - Mon, 29 Feb 2016 17:30:52 -0800 - rev 362338
Push 16998 by rwood@mozilla.com at Mon, 02 May 2016 19:42:03 +0000
testing: upgrade run-tests.py Imported run-tests.py from changeset c7f89ad87bae from upstream. This corresponds to a version partially through the 3.8 release cycle. This gets us a number of bug fixes. Our fixes to increase port allocation and perform DOCKER_HOSTNAME substitution have been added back in. I should really spend time to make these configurable upstream so we don't need to modify run-tests.py... We had to revert a change to _hghave() regarding locating the hghave binary. Upstream changed it to look for hghave next to run-tests.py. We could refactor our tree to put our custom hghave in the same directory. This is probably for the better. For now, it is scope bloat.
79fa6ea6a9a7981a0576763e427f525e04b18936: testing: securely install Python packages
Gregory Szorc <gps@mozilla.com> - Mon, 29 Feb 2016 14:27:09 -0800 - rev 362337
Push 16998 by rwood@mozilla.com at Mon, 02 May 2016 19:42:03 +0000
testing: securely install Python packages Before, we relied on the virtualenv on the system. We also upgraded pip and setuptools via normal pip foo, implicitly trusting the server. We also didn't pin hashes in test-requirements.txt. This commit changes all that. We now download virtualenv, setuptools, and pip from our S3 package server. The hashes are pinned so we can detect tampering. Hashes have also been added to test-requirements.txt. Some packages have multiple hashes because OS X and Linux will download different archives. lxml has been upgraded to 3.5.0 due to issues building the existing version on Ubuntu.
7b9b58e346dff827a146de78edf4a626725d5989: git-mozreview: treat commit messages as byte sequences (bug 1251602); r=dminor
Gregory Szorc <gps@mozilla.com> - Mon, 29 Feb 2016 11:47:52 -0800 - rev 362336
Push 16998 by rwood@mozilla.com at Mon, 02 May 2016 19:42:03 +0000
git-mozreview: treat commit messages as byte sequences (bug 1251602); r=dminor More Unicode foo. We have unicode_literals enabled in git-mozreview. We were opening the commit message file in binary mode but when we performed some string operations with '' literals, Python would try to convert the unicode literal to str using the ascii encoding and the world would blow up. We use b'' literals to prevent implicit type conversion. A test demonstrating bytes preservation has been added. MozReview-Commit-ID: CTL5sXUkd1W
f07e02773a0a03a4bfbd2bf41907ea55809f7dc7: ansible/mozreview: Install npm in virtualenv.
Mark Cote <mcote@mozilla.com> - Mon, 29 Feb 2016 12:15:05 -0500 - rev 362335
Push 16998 by rwood@mozilla.com at Mon, 02 May 2016 19:42:03 +0000
ansible/mozreview: Install npm in virtualenv.
2d5d156dcb24ac07619745115b80a77c64b68a26: mozreview: draw attention to review comments (bug 1101611) r=mdoglio
byron jones <glob@mozilla.com> - Tue, 23 Feb 2016 15:34:50 +0800 - rev 362334
Push 16998 by rwood@mozilla.com at Mon, 02 May 2016 19:42:03 +0000
mozreview: draw attention to review comments (bug 1101611) r=mdoglio Review comments appear only as a tiny number in the left margin. This makes them easy to overlook, especially when viewing side-by-side diffs where your focus is on the right column. This adds a dotted line across the diff view to draw attention to the comments. MozReview-Commit-ID: EohHRRTY9pB
e1c7901963e78bba147e1dcf1338c1082a049ae2: mozreview: fix indentation (bug 1101611) r=mdoglio
byron jones <glob@mozilla.com> - Tue, 23 Feb 2016 13:31:07 +0800 - rev 362333
Push 16998 by rwood@mozilla.com at Mon, 02 May 2016 19:42:03 +0000
mozreview: fix indentation (bug 1101611) r=mdoglio MozReview-Commit-ID: 3V2VdzlT0AZ
ae18684189a0f3c6045f5a8aa08ee579a8d715c2: docker/bmoweb: use socket for connecting to MySQL (bug 1246634)
Gregory Szorc <gps@mozilla.com> - Fri, 26 Feb 2016 14:09:38 -0800 - rev 362332
Push 16998 by rwood@mozilla.com at Mon, 02 May 2016 19:42:03 +0000
docker/bmoweb: use socket for connecting to MySQL (bug 1246634) We run MySQL inside the container. We don't need to involve TCP sockets. Let's connect to the server over the UNIX socket.
aca609c335630243fbac9d5540bb9aa99544e3de: docker/bmoweb: explicitly wait on mysqld process to exit (bug 1246634)
Gregory Szorc <gps@mozilla.com> - Fri, 26 Feb 2016 13:58:42 -0800 - rev 362331
Push 16998 by rwood@mozilla.com at Mon, 02 May 2016 19:42:03 +0000
docker/bmoweb: explicitly wait on mysqld process to exit (bug 1246634) I /think/ there is a race condition here since terminate() only sends a signal and doesn't wait on the process to actually exit.
52aaee1436b7d4cf07bce337287600135396ea9c: docker/bmoweb: start mysql after cpan module installation (bug 1246634)
Gregory Szorc <gps@mozilla.com> - Fri, 26 Feb 2016 13:58:01 -0800 - rev 362330
Push 16998 by rwood@mozilla.com at Mon, 02 May 2016 19:42:03 +0000
docker/bmoweb: start mysql after cpan module installation (bug 1246634) For reasons I can't explain, failure in cpanm will result in the mysqld process terminating. I suspect a signal is being sent to mysqld somehow. Perhaps subprocess.check_call doesn't trap signal handlers like I thought it did...
8e7b9b7313633d975aff16ccd2cd9e92c0d87627: docker/bmoweb: use mysqld instead of mysqld_safe (bug 1246634)
Gregory Szorc <gps@mozilla.com> - Fri, 26 Feb 2016 13:40:32 -0800 - rev 362329
Push 16998 by rwood@mozilla.com at Mon, 02 May 2016 19:42:03 +0000
docker/bmoweb: use mysqld instead of mysqld_safe (bug 1246634)
28a39ce8a3b75a654e77ad62a0c2a40d55d249a9: ansible/mozreview: install Python 2.7
Gregory Szorc <gps@mozilla.com> - Fri, 26 Feb 2016 11:35:08 -0800 - rev 362328
Push 16998 by rwood@mozilla.com at Mon, 02 May 2016 19:42:03 +0000
ansible/mozreview: install Python 2.7 The MozReview web heads are still running Python 2.6 from the system Python installation. We want to run Python 2.7 everywhere. The first step towards transitioning rbweb to Python 2.7 is to get Python 2.7 installed. We do this by establishing a new "rbweb" role and have it install Python 2.7. We add this role in various locations. It's worth noting that the admin and web machines are currently more under the control of Puppet than Ansible. We'll need to slowly bring more things under Ansible control in order to switch to actually running things from Python 2.7.
(0) -300000 -100000 -30000 -10000 -3000 -1000 -300 -100 -50 -20 +20 +50 +100 +300 +1000 +3000 +10000 +30000 +100000 +300000 tip