Sat, 28 Nov 2020 19:58:37 +0100 phabricator: introduce a `phabricator.retry` option
Pierre-Yves David <pierre-yves.david@octobus.net> [Sat, 28 Nov 2020 19:58:37 +0100] rev 45969
phabricator: introduce a `phabricator.retry` option For the past 2 days, my connection to phab.mercurial-scm.org became extremely poor. In practice this mean that any conduit call has a fairly high change to hang and die. Giving the amount of call done by the phabricator extension, it means the when I am lucky I can get 1 or 2 Diff to update after a few try, but anything sizeable doesn't have any hope to get through. This changesets introduce a new option for the fabricator extension to try retry failed command itself. So that I can get Diff through. As you can guess, this changeset managed to reach Phabricator thanks to itself. Differential Revision: https://phab.mercurial-scm.org/D9449
Tue, 24 Nov 2020 16:17:16 -0500 packaging: drop Disco (19.04) and add Focal (20.04)
Matt Harbison <matt_harbison@yahoo.com> [Tue, 24 Nov 2020 16:17:16 -0500] rev 45968
packaging: drop Disco (19.04) and add Focal (20.04) Disco support ended in January 2020, and Focal does not have an announced EOL. Something is now installing and configuring `tzdata`, which was throwing up an interactive prompt to configure the timezone. Aside from being hostile to automation, the prompt didn't actually accept input and hung the process. This propagates the host's timezone into the image via environment variable in order to skip the prompt, and avoid hardcoding a value. Differential Revision: https://phab.mercurial-scm.org/D9396
Tue, 24 Nov 2020 14:47:24 -0500 make: drop Trusty and Artful targets
Matt Harbison <matt_harbison@yahoo.com> [Tue, 24 Nov 2020 14:47:24 -0500] rev 45967
make: drop Trusty and Artful targets These were removed from the packaging makefile in 0363bb086c57. Differential Revision: https://phab.mercurial-scm.org/D9395
Tue, 24 Nov 2020 14:03:19 -0500 packaging: add `HG_DOCKER_OWN_USER` to `dockerdeb` like exists in `dockerrpm`
Matt Harbison <matt_harbison@yahoo.com> [Tue, 24 Nov 2020 14:03:19 -0500] rev 45966
packaging: add `HG_DOCKER_OWN_USER` to `dockerdeb` like exists in `dockerrpm` I was getting build failures when it was trying to write to the working directory on CentOS 7 without this. It is basically the same as was added to the RPM builder in 4c0d4bbdc395. For some reason, this doesn't work with Xenial, and the only solution I found was to invoke it with UID 1000 on the host side. It doesn't EOL until April 2024, but it also has python 3.5.2, so this build complication is the least of the problems with supporting it after py2 is dropped. Differential Revision: https://phab.mercurial-scm.org/D9394
Sun, 29 Nov 2020 15:17:14 +0100 heptapod-ci: do not publish changeset when doing the local clone
Pierre-Yves David <pierre-yves.david@octobus.net> [Sun, 29 Nov 2020 15:17:14 +0100] rev 45965
heptapod-ci: do not publish changeset when doing the local clone Otherwise, checks and script relying on some changeset being draft fails. Differential Revision: https://phab.mercurial-scm.org/D9461
Mon, 02 Nov 2020 19:25:26 +0100 copies-rust: pre-indent some code to clarify the next changeset
Pierre-Yves David <pierre-yves.david@octobus.net> [Mon, 02 Nov 2020 19:25:26 +0100] rev 45964
copies-rust: pre-indent some code to clarify the next changeset The next changeset will massively rewrite the next function. However having a clear diff on the core semantic of the function will help making the next changesets clearer. So we do most of the churn beforehand. Differential Revision: https://phab.mercurial-scm.org/D9301
(0) -30000 -10000 -3000 -1000 -300 -100 -30 -10 -6 +6 +10 +30 +100 +300 +1000 +3000 tip