1147742 - esr31 appears to be 'behind' on vcs-sync maintained version on github

Status: RESOLVED WORKSFORME

(Developer Services :: General, task)

Description

[Hal Wine [:hwine] (leaving Jan 17, ping on slack if needed)]

reported by mikeperry on #vcs

Since blocking tor's release, Callek paged hwine

Comment 1

[Hal Wine [:hwine] (leaving Jan 17, ping on slack if needed)]

log from #vcs:

20:48 <  mikeperry> | hwine-ooo, et al: It seems that the script that syncs the hg repo to https://github.com/mozilla/gecko-dev/commits/esr31 is currently down. can someone give it a kick?
20:49 <   mikeperry>| it's missing the last 5 commits or so, as well as the 31.6.0-build2 checkin
20:50 <      Callek>| mikeperry: btw, is there an urgency behind this, as in do you feel its "page a human" worthy?  (and if you do, whats the justification for it)
20:51              * | Callek is asking in earnest, since if I'm the one doing a page I like to be able to justify it myself ;-)
20:51 <   mikeperry>| Callek: Tor Browser builds from the git repo. we need those commits in order to kick off our builds based on 31.6.0-build2
20:52 <   mikeperry>| it's near the end of my day, so a tag now means I could start a build and let it run overnight, saving us roughly a day of wallclock time
20:54 <   mikeperry>| probably not worth waking somebody up over, but if they are awake but just afk, probably?
20:57 <       hwine> | mikeperry: how long ago were the commits made, and what is the hg link?
20:57 <   mikeperry>| https://hg.mozilla.org/releases/mozilla-esr31 vs https://github.com/mozilla/gecko-dev/commits/esr31
20:58 <   mikeperry>| it seems like it died about 5 hours ago? there have been commits arriving regularly at the hg branch since then, but not on git
21:00 <       hwine> | so, last hg commit is about 30 minutes ago -- that may not have propigated yet
21:01 <   mikeperry>| some of them were much more than 30 min ago though.
21:02 <   mikeperry>| the gstreamer ones are at least a couple hours old
21:02 <     Callek> | hwine: so yea Ryan's commit was ~4 minutes after jordan's https://hg.mozilla.org/releases/mozilla-esr31/pushloghtml
21:02 <      Callek>| then 5 hours passes, then the tagging for the firefox release
21:04 <       hwine> | yeah - checking -- this is modern vcs-sync
21:07 <       hwine> | no email alerts for that repo
21:19 <       hwine> | so, it looks correct in conversion - unsure why it's not pushed to github
21:20 <       hwine> | mikeperry: it's all there on github
21:23 <       hwine> | mikeperry: what may be confusing you is that those commits are not on default of esr31 -- they are only on the rel branch
21:24 <  mikeperry> | hwine: ah, yes, it looks like if I git pull I get all of them except for the version bump, which is good enough
21:24 <   mikeperry>| wonder why the webui is stale. anyways, thanks so much for taking the time to look into this!

Comment 2

[Hal Wine [:hwine] (leaving Jan 17, ping on slack if needed)]

As mikeperry noted in comment 1 it was the github webui that was not current.

modern vcs-sync was working properly, commits were in the repo on github, just not displayed.

Comment 3

[Hal Wine [:hwine] (leaving Jan 17, ping on slack if needed)]

Attachment: github-esr31-lag.png

From conversion repo on vcssync1:
(venv)[[e-mail] beagle]$ git show-ref esr31
489560a4cb3240efc2753c180e41a1641503d254 refs/heads/esr31

Attached is screen shot of github at same time, showing most recent commit to be df23c797eb4b3c550304f53dd50710ad0f7f0514

Comment 4

[Hal Wine [:hwine] (leaving Jan 17, ping on slack if needed)]

A non-gui check returns the correct value on github:

$ git ls-remote https://github.com/mozilla/gecko-dev esr31
489560a4cb3240efc2753c180e41a1641503d254        refs/heads/esr31