bug 694432 has the details, but Win64 builds have been crashing on startup for a while now. We should stop updating users to broken builds.

We've had Win64 nightlies for mozilla-central, ux, and jaegermonkey in the last 5 days. I've renamed the $branch/WINNT_x86_64-msvc directories to WINNT_x86_64-msvc.disabled_bug695161 to 'unpublish' those builds. Patch coming up to keep them disabled when we do more builds.

Keep publishing the snippets, just in the location on the side so they're not given to users.

Comment on attachment 567586 [details] [diff] [review] [buildbot-configs] Change upload location I'll give review here if you'll take it. To save coop some trouble. This looks good.

Comment on attachment 567586 [details] [diff] [review] [buildbot-configs] Change upload location Landed http://hg.mozilla.org/build/buildbot-configs/rev/aa306c9ef734 and deployed to the build masters.

And now that I look at bug 694432 I'm confused by it talking about startup failing on Windows 8, both 32 and 64bit. And our tests are still running green on Win7, eg https://tbpl.mozilla.org/?noignore=1&rev=1318d1bbc15a What am I missing ?

This has nothing to do with 32 or 64 bit builds. The builds crash on Windows 8.

Can we turn the updates on please? The referring bug 694432 is already fixed. FWIW, current Nightly builds do not crash on start, regardless of Windows version.

Would updates be enabled? I've been on Mozilla/5.0 (Windows NT 6.1; Win64; x64; rv:10.0a1) Gecko/20111017 Firefox/10.0a1 ID:20111017031010 for months now with mostly no crashes. about:crashes shows: Report ID Date Submitted bp-373370e2-3eaf-4e75-98ec-efe092110817 17/08/201109:03 a.m. bp-6afcae4d-5adb-4630-a958-54aca2110817 17/08/201109:03 a.m. bp-7909af87-8427-4222-8160-264242110525 25/05/201109:45 a.m.

I guess we can re-enable updates if bug 694432 is fixed, so I'll do that today. Would still love to have an answer to comment #5. Seems strange that we disabled on 64bit only for a ultra-minority platform in Windows 8.

To comment #5 What I do not understand is why would Mozilla even accept a bug report dealing with an OS that is not even in Alpha stage yet?

Comment on attachment 567586 [details] [diff] [review] [buildbot-configs] Change upload location Backed out: http://hg.mozilla.org/build/buildbot-configs/rev/0ac1539bba0d Still need to reconfig the masters for that to take effect for new nightlies.

(In reply to Nick Thomas [:nthomas] from comment #1) > I've renamed the $branch/WINNT_x86_64-msvc directories to > WINNT_x86_64-msvc.disabled_bug695161 to 'unpublish' those builds. This has effectively been undone by symlinking WINNT_x86_64-msvc -> WINNT_x86_64-msvc.disabled_bug695161 Once we reconfig the masters I'll remove the symlink and rename instead, but in the meantime this makes sure we get the latest builds offered.

FWIW I just got updated to Mozilla/5.0 (Windows NT 6.1; Win64; x64; rv:10.0a1) Gecko/20111021 Firefox/10.0a1 ID:20111021031012 moments ago. Thanks Nick!

(In reply to Nick Thomas [:nthomas] from comment #12) > Once we reconfig the masters I'll remove the symlink and rename instead, but > in the meantime this makes sure we get the latest builds offered. A reconfig that included this happened today.

Sorry, I filed this on behalf of another user, and it wasn't made clear to me that it was only broken on Windows 8. (In reply to Gary King from comment #10) > To comment #5 > What I do not understand is why would Mozilla even accept a bug report > dealing with an OS that is not even in Alpha stage yet? Just because it's an alpha or pre-alpha does not mean that all bugs related to it are invalid. In this case, we appear to have had a valid crashing bug on Windows 8 builds. We fixed it.

All done here.