437143 - try servers need to support building from a mercurial repository with a patch

Status: VERIFIED FIXED

(Release Engineering :: General, defect, P2)

Description

[Robert O'Callahan (:roc) (email my personal email if necessary)]

The try servers are still pulling from CVS. That means that patches against mozilla-central don't apply anymore.

This wouldn't be so bad if the "build from a Mercurial repository" had an option to take an additional patch to apply. But it doesn't.

Comment 1

[[e-mail] (:bhearsum)]

Yeah, that needs to be fixed. I can't give priority to this right now since there's a workaround.

Comment 2

[[e-mail] (:bhearsum)]

To be clear, I think we _do_ want to maintain CVS try server (at least until the build environment changes incompatibly), but we _should_ allow a patch to be applied when pulling from an hg repo.

Comment 3

[Vladimir Vukicevic [:vlad] [:vladv] (needinfo me, slow to respond)]

Attachment: Add pulling from hg repo with a patch

Untested, but I think this'll do what's needed.  Ben, what's the easiest way to test this?

Comment 4

[[e-mail] (:bhearsum)]

Comment on attachment 328425 [details] [diff] [review]
Add pulling from hg repo with a patch

I don't have time to give this a full review right now but I scanned over it and it seems OK.

We've got a try server staging environment that we can test it on.

Comment 5

[Vladimir Vukicevic [:vlad] [:vladv] (needinfo me, slow to respond)]

Any progress on getting this set up on the staging environment?

Comment 6

[[e-mail] (:bhearsum)]

Sorry if there was some confusion, but I have not even tried to get this going in staging yet. I *might* get a chance later this week.

Comment 7

[[e-mail] (:bhearsum)]

OK, testing this out today.

Comment 8

[[e-mail] (:bhearsum)]

With a few tweaks I've successfully tested this patch with cvs+patch and w/ an alternative hg repo. I did a quick test of m-c+patch which failed. It appears that the patch is never written to a file on the webserver. I don't have time to figure out why tonight, but I'll tackle that tomorrow.

Thanks for doing the initial work here, Vlad, give that I've already tweaked it some I may as well finish it up.

Comment 9

[[e-mail] (:bhearsum)]

Attachment: [checked in] try server support for m-c builds w/ patch - cleaned up + tested

Rob, are you up to reviewing this?

Comment 10

[[e-mail] (:bhearsum)]

Comment on attachment 329070 [details] [diff] [review]
[checked in] try server support for m-c builds w/ patch - cleaned up + tested

Oh, I meant to comment about the processchange.pl removal from buildbot-configs/tryserver. I originally imported it there as a local version (ie, with the global vars set properly) of the webtools/buildbot-try script. Looking at it now, it seems more confusing than necessary, so I just got rid of the buildbot-configs/tryserver one and set-up the webtools/buildbot-try one properly.

Comment 11

[Robert Helmer [:rhelmer]]

Comment on attachment 329070 [details] [diff] [review]
[checked in] try server support for m-c builds w/ patch - cleaned up + tested

Looks ok to me. 

It'd be nice to use the native bbot try features, Try_Jobdir, etc. instead of rolling this all ourselves, once we drop support for CVS/client.mk builds.. I see no reason for the web interface to not call "buildbot try" directly and use bbot's queuing feature, in the glorious trouble-free future.

Comment 12

[[e-mail] (:bhearsum)]

(In reply to comment #11)
> (From update of attachment 329070 [details] [diff] [review])
> Looks ok to me. 
> 
> It'd be nice to use the native bbot try features, Try_Jobdir, etc. instead of
> rolling this all ourselves, once we drop support for CVS/client.mk builds.. I
> see no reason for the web interface to not call "buildbot try" directly and use
> bbot's queuing feature, in the glorious trouble-free future.
> 

Yeah, I totally agree. I've been wanting to do that for awhile..maybe I'll get time some day ;-).

Comment 13

[[e-mail] (:bhearsum)]

Comment on attachment 329070 [details] [diff] [review]
[checked in] try server support for m-c builds w/ patch - cleaned up + tested

Checking in master.cfg;
/cvsroot/mozilla/tools/buildbot-configs/tryserver/master.cfg,v  <--  master.cfg
new revision: 1.16; previous revision: 1.15
done
Removing processchanges.pl;
/cvsroot/mozilla/tools/buildbot-configs/tryserver/processchanges.pl,v  <--  processchanges.pl
new revision: delete; previous revision: 1.1
done
Checking in steps.py;
/cvsroot/mozilla/tools/buildbotcustom/tryserver/steps.py,v  <--  steps.py
new revision: 1.4; previous revision: 1.3
done
Checking in processchanges.pl;
/cvsroot/mozilla/webtools/buildbot-try/processchanges.pl,v  <--  processchanges.pl
new revision: 1.4; previous revision: 1.3
done
Checking in sendchange-ui.pm;
/cvsroot/mozilla/webtools/buildbot-try/sendchange-ui.pm,v  <--  sendchange-ui.pm
new revision: 1.5; previous revision: 1.4
done
Checking in sendchange.cgi;
/cvsroot/mozilla/webtools/buildbot-try/sendchange.cgi,v  <--  sendchange.cgi
new revision: 1.3; previous revision: 1.2
done

Comment 14

[[e-mail] (:bhearsum)]

OK, I've updated the web server and buildbot master. Thanks again for the patch, Vlad.

Comment 15

[Henrik Skupin [:whimboo][⌚️UTC+2]]

Verfied with a tryserver build from dao yesterday:
https://build.mozilla.org/tryserver-builds/2008-08-05_14:[e-mail]-prefs_apps_search/