Closed Bug 779907 Opened 12 years ago Closed 12 years ago

xulrunner builds don't pass configure on mac

Categories

(Release Engineering :: General, defect)

defect
Not set
normal

Tracking

(Not tracked)

RESOLVED FIXED

People

(Reporter: bhearsum, Assigned: rail)

References

Details

Attachments

(1 file, 1 obsolete file)

Adding configure options from /builds/slave/m-cen-osx64-xr/build/.mozconfig:
  --target=i386-apple-darwin10
  --with-macos-sdk=/Developer/SDKs/MacOSX10.6.sdk
  --enable-stdcxx-compat
  --with-ccache
  --enable-application=xulrunner
  --disable-tests
creating cache ./config.cache
checking host system type... /builds/slave/m-cen-osx64-xr/build/build/autoconf/config.guess: line 1251: -arch: command not found
i386-apple-darwin11.2.0
checking target system type... i386-apple-darwin10
checking build system type... i386-apple-darwin11.2.0
checking for mawk... no
checking for gawk... no
checking for nawk... no
checking for awk... awk
checking for perl5... no
checking for perl... /usr/bin/perl
cross compiling from i386-apple-darwin11.2.0 to i386-apple-darwin10
checking for host c compiler... checking for gcc... gcc
gcc
checking for host c++ compiler... checking for c++... c++
c++
checking for ranlib... no
checking for ar... no
checking whether the host c compiler (gcc  ) works... yes
checking whether the host c++ compiler (c++  ) works... yes
checking for -arch...  -arch i386
checking for gcc...  -arch i386
checking whether the C compiler ( -arch i386  ) works... no
configure: error: installation or configuration problem: C compiler cannot create executables.
*** Fix above errors and then restart with               "make -f client.mk build"
http://mxr.mozilla.org/mozilla-central/source/build/macosx/common

Which is sourced from the universal mozconfig, looks like it can get into a situation where it doesn't set CC/CXX, which is what's happening here. If we want these builds to use clang we should fix them.
Blocks: 733905
I this a case of tooltool not running on a build? bhearsum, can you include a link to the where you got the log?
Yes, that is it. It is not even trying to run tooltool.
Rail, do I need to create a manifest for this or can we just reuse the regular releng.manifest?
I suspect that we're probably not passing those options along to the factory in misc.py. In any case, this seems like primarily a RelEng issue. Thanks for your help Rafael!
Component: XULRunner → Release Engineering: Automation (General)
Product: Toolkit → mozilla.org
QA Contact: catlee
Version: unspecified → other
Assignee: nobody → rail
Attachment #648495 - Flags: review?(bhearsum)
Attachment #648495 - Flags: review?(bhearsum) → review+
Attached patch release.py (obsolete) — Splinter Review
Comment on attachment 648495 [details] [diff] [review]
enable tooltool for xulrunner builds

http://hg.mozilla.org/build/buildbotcustom/rev/125fb8df2570
Attachment #648495 - Flags: checked-in+
We should probably also fix that mozconfig so that it doesn't set CC=" -i386" when we hit this situation. Even just erroring out sooner would be better.
/builds/slave/m-cen-osx64-xr/build/obj-firefox/i386/config/nsinstall: cannot make symbolic link /builds/slave/m-cen-osx64-xr/build/obj-firefox/i386/dist/bin/xulrunner: File exists
make[7]: *** [libs] Error 1

Could be a missing clobber?
(In reply to Rafael Ávila de Espíndola (:espindola) from comment #11)
> Could be a missing clobber?

Bah, I forgot about that this time. Now the build is green! Wooo!
Comment on attachment 648509 [details] [diff] [review]
release.py

Moved to bug 780298
Attachment #648509 - Attachment is obsolete: true
All done here.
Status: NEW → RESOLVED
Closed: 12 years ago
Resolution: --- → FIXED
(In reply to Ted Mielczarek [:ted] from comment #9)
> We should probably also fix that mozconfig so that it doesn't set CC="
> -i386" when we hit this situation. Even just erroring out sooner would be
> better.

This was already filed as bug 778035.
BTW, why don't we show xulrunner in tbpl?
Because it's a tier-2 platform.
(In reply to Benjamin Smedberg  [:bsmedberg] [away 27-July until 7-Aug] from comment #17)
> Because it's a tier-2 platform.

Interesting. The original email from Ben Hearsum about the breakage stated:

------------------------------------------
Because they don't show up TBPL, there is little visibility into them, but these
*are* shipping products and need to be fixed:
------------------------------------------

From the wording I got the impression that they were tier 1.
They are not. We do ship them (in a most untested state) for each release because they are what produce the SDK for extension authors, which is why we don't want to break them. But we don't hold the tree closed if they break, we just expect committers to help fix them afterwards.
Product: mozilla.org → Release Engineering
Component: General Automation → General
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: