Routinely pushing (and often hitting) the 5.5hr timeout limit now: https://tbpl.mozilla.org/php/getParsedLog.php?id=44724884&tree=Mozilla-Inbound Regression range: https://tbpl.mozilla.org/?jobname=WINNT%205.2%20mozilla-central%20pgo-build&fromchange=6c0971104909&tochange=e07264876182 Lots of build system changes in that inbound mega-merge. Will try to narrow it down on inbound next.

Tentative inbound regression range: https://tbpl.mozilla.org/?tree=Mozilla-Inbound&jobname=WINNT%205.2%20mozilla-inbound%20pgo-build&fromchange=4563a79a7968&tochange=0300f04572b6 The run on rev 51ff7710ccbc tells me that the regression was before glandium's build system changes landed. Below that, bug 1038537 stands out as a definite maybe. Also, bug 1043534 looks like a possibility.

Backout bug 1038537: https://tbpl.mozilla.org/?tree=Try&rev=fbdfa49e3b0c Backout bug 1043534: https://tbpl.mozilla.org/?tree=Try&rev=d50998766017 Backout glandium's push: https://tbpl.mozilla.org/?tree=Try&rev=46e80adcc5e2

Reading the build log, it appears for some reason, libxul is rebuilt during the libs phase. Which means bug 1043865 is going to fix this.

Retriggers appear to be pointing at Nate's push as the regressor (the push prior ran in 197min): https://tbpl.mozilla.org/?tree=Mozilla-Inbound&jobname=WINNT%205.2%20mozilla-inbound%20pgo-build&fromchange=4563a79a7968&tochange=0300f04572b6&rev=2246b59a134a (In reply to Mike Hommey [:glandium] from comment #3) > Reading the build log, it appears for some reason, libxul is rebuilt during > the libs phase. Which means bug 1043865 is going to fix this. Is that going to land soon? This regression is absolutely brutal from a sheriff merge standpoint. Otherwise I'll just back bug 1043534 out and let you two sort it out.

This was fixed by bug 1043865.