Closed
Bug 1095679
Opened 10 years ago
Closed 10 years ago
Mulet builds on !trunk are broken from today's reconfig
Categories
(Infrastructure & Operations Graveyard :: CIDuty, task)
Infrastructure & Operations Graveyard
CIDuty
Tracking
(firefox34 fixed, firefox35 fixed, firefox36 unaffected, b2g-v2.1 fixed, b2g-v2.2 unaffected)
RESOLVED
FIXED
Tracking | Status | |
---|---|---|
firefox34 | --- | fixed |
firefox35 | --- | fixed |
firefox36 | --- | unaffected |
b2g-v2.1 | --- | fixed |
b2g-v2.2 | --- | unaffected |
People
(Reporter: RyanVM, Assigned: janx)
References
Details
http://ftp.mozilla.org/pub/mozilla.org/b2g/tinderbox-builds/mozilla-b2g34_v2_1-linux64-mulet/1415387045/mozilla-b2g34_v2_1-linux64-mulet-bm73-build1-build13.txt.gz b2g/dev/config/tooltool-manifests/linux64/releng.manifest doesn't exist, skipping... Which leads to... checking for gcc... /builds/slave/m-b34_21-l64-mulet-00000000000/build/gcc/bin/gcc checking whether the C compiler (/builds/slave/m-b34_21-l64-mulet-00000000000/build/gcc/bin/gcc ) works... no configure: error: installation or configuration problem: C compiler cannot create executables. ------ config.log ------ This file contains any messages produced by compilers while running configure, to aid debugging if configure makes a mistake. configure:1184: checking host system type configure:1205: checking target system type configure:1223: checking build system type configure:1299: checking for gawk configure:1384: checking for python2.7 configure:1494: checking Python environment is Mozilla virtualenv configure:1715: checking for perl5 configure:1715: checking for perl configure:2191: checking for objcopy configure:3194: checking for gcc configure:3307: checking whether the C compiler (/builds/slave/m-b34_21-l64-mulet-00000000000/build/gcc/bin/gcc ) works configure:3323: /builds/slave/m-b34_21-l64-mulet-00000000000/build/gcc/bin/gcc -o conftest conftest.c 1>&5 /builds/slave/m-b34_21-l64-mulet-00000000000/build/configure: line 3322: /builds/slave/m-b34_21-l64-mulet-00000000000/build/gcc/bin/gcc: No such file or directory configure: failed program was: #line 3318 "configure" #include "confdefs.h" main(){return(0);} configure: error: installation or configuration problem: C compiler cannot create executables. *** Fix above errors and then restart with\ "make -f client.mk build" make[2]: *** [configure] Error 1 make[2]: Leaving directory `/builds/slave/m-b34_21-l64-mulet-00000000000/build' make[1]: *** [/builds/slave/m-b34_21-l64-mulet-00000000000/build/obj-firefox/Makefile] Error 2 make[1]: Leaving directory `/builds/slave/m-b34_21-l64-mulet-00000000000/build' make: *** [build] Error 2
Flags: needinfo?(janx)
Reporter | ||
Comment 1•10 years ago
|
||
Also affecting Gum (presumably any non-trunk trees)
Reporter | ||
Updated•10 years ago
|
Summary: Mulet builds on b2g34 are broken from today's reconfig → Mulet builds on !trunk are broken from today's reconfig
Comment 2•10 years ago
|
||
So I think we want a patch like this to use the older tooltool manifest for previous versions (untested): # Enable win32/macosx64 mulet in gecko 36+ for name, branch in items_before(BRANCHES, 'gecko_version', 36): + if 'linux64-mulet' in branch['platforms']: + branch['platforms']['linux64-mulet']['tooltool_manifest_src'] = 'browser/config/tooltool-manifests/linux64/releng.manifest' if 'win32-mulet' in branch['platforms']: del branch['platforms']['win32-mulet'] if 'macosx64-mulet' in branch['platforms']: del branch['platforms']['macosx64-mulet'] Though given that this logic is supposed to be disabling win32-mulet and macosx-64mulet for gecko_version < 36, I find it odd that we get win32 and macosx builds on Gum. (I was expecting either no win32/macosx builds, or for it to have m-c's tooltool manifests). According to http://mxr.mozilla.org/build/source/buildbot-configs/mozilla/project_branches.py#227 gum should be gecko_version==35, but I guess this is different for b2g: http://mxr.mozilla.org/build/source/buildbot-configs/mozilla/b2g_project_branches.py#61 Anyway, I'm not sure what the correct fix here is. I suspect we'll want something like this with the branch info sorted out, and possibly with additional tests for osx/win32 with their previous tooltool manifest locations.
Comment hidden (obsolete) |
Comment hidden (obsolete) |
Assignee | ||
Comment 5•10 years ago
|
||
Disregard my previous comments. The reconfig made buildbots look for new tooltool manifests for Mulet (files b2g/dev/config/tooltool-manifests/{win32,linux64,macosx64}/releng.manifest) that were landed in bug 1011562 as https://hg.mozilla.org/mozilla-central/rev/8a53b8b76456. My guess is that these manifests are not present in non-trunk trees, causing Mulet builds to fail. If that's the case, I see two solutions for non-trunk trees: - Patches https://bug1011562.bugzilla.mozilla.org/attachment.cgi?id=8518112 and https://bug1011562.bugzilla.mozilla.org/attachment.cgi?id=8518934 should be landed there, adding the right manifests. - Or make non-trunk trees use the old browser manifests for Mulet (files browser/config/tooltool-manifests/{win32,linux64,macosx64}/releng.manifest), as Michael suggests in comment 2. Not sure how to "back out" the reconfig on non-trunk trees brought by buildbot patch https://bug1011562.bugzilla.mozilla.org/attachment.cgi?id=8517399
Reporter | ||
Comment 6•10 years ago
|
||
Special-casing build configs sucks and backporting looks trivial. Let's go that way :)
Flags: needinfo?(ryanvm)
Reporter | ||
Comment 7•10 years ago
|
||
Only attachment 8518112 [details] [diff] [review] has landed on trunk so far, so I'm only uplifting that for now.
Flags: needinfo?(ryanvm)
Assignee | ||
Comment 8•10 years ago
|
||
Sure! Then again I'm not even sure you want the commented code in mozconfig files (might cause conflicts), you just need the manifests. I shouldn't have done both changes as one patch.
Reporter | ||
Comment 9•10 years ago
|
||
Victory!
Assignee: nobody → janx
Status: NEW → RESOLVED
Closed: 10 years ago
status-b2g-v2.1:
--- → fixed
status-b2g-v2.2:
--- → unaffected
status-firefox34:
--- → fixed
status-firefox35:
--- → fixed
status-firefox36:
--- → unaffected
Resolution: --- → FIXED
Assignee | ||
Comment 10•10 years ago
|
||
Yay! Ryan, thanks a lot for the heavy (up)lifting :)
Updated•6 years ago
|
Product: Release Engineering → Infrastructure & Operations
Updated•4 years ago
|
Product: Infrastructure & Operations → Infrastructure & Operations Graveyard
You need to log in
before you can comment on or make changes to this bug.
Description
•