There seem to be many issues building on 10.9 (At the time of this writing, Developer Preview 3). Some issues may be our fault, others seem to be based on the third-party tools we depend on. This bug will track both cases, to keep tabs on any problems. Problems on our end we should fix, the other bugs may just need to wait until the dependent tools are updated.
My build issues went away after I added the following to my .mozconfig ac_add_options --with-macos-sdk=/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.8.sdk
(In reply to Stefan Arentz [:st3fan] from comment #1) > My build issues went away after I added the following to my .mozconfig > > ac_add_options > --with-macos-sdk=/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX. > platform/Developer/SDKs/MacOSX10.8.sdk Interesting. Did you happen to install 10.9 as an upgrade or on a fresh partition? Bootstrap appeared to have many problems, along with hg, linking, permissions, and more. The SDK seemed to be one of the simplier issues. Regardless, I plan to reinstall 10.9 and start reporting every issue I encounter in a few days. We could create a wiki page with temporary fixes until we can update mach and bootstrap to function correctly.
Not really a build setup / mach error, but Bug 884014 is possibly related.
Xcode 5.0 (I've tested DP1 and DP3) appear to have some sort of compiler bug resulting in very odd linker errors. You can see them here: https://gist.github.com/CorgiDude/6003830 -- disabling jemalloc allows the build to progress further but it still fails with a very similar error https://gist.github.com/CorgiDude/6003919 inside the js tier. Using the compiler toolchain from Xcode 4.6 solves this issue, even using the 10.9 SDK - so I believe this is a legitimate LLVM bug. I have not yet filed a bug because I would like to isolate a better test case than "build Mozilla" for the LLVM team. My only issue is that I cannot seem to find any common thread between the files that are failing that is special only to them. I've looked at compilation options, linker options, even headers they include, and can't find a single thing that is unique only to the files that it falls with. Will probably file a bug with them Friday even if I can't figure out a good trigger.
I think the duplicate symbols error you're seeing is a result of some code doing "#define inline" in an attempt to disable inlining. There's an occurrence of this at http://mxr.mozilla.org/mozilla-aurora/source/memory/jemalloc/src/include/jemalloc/internal/jemalloc_internal.h.in#235 which could explain the jemalloc failure.
By the way, this bug is a meta-bug, which is meant to block multiple bugs of the same kind, and thereby make it easier to keep track of them. Meta-bugs usually have minimal content. Clearly identifiable problems using mach to build Mozilla on OS X 10.9 should be reported in new bugs, which are made to block this one.
> which is meant to block multiple bugs which is meant to *depend on* multiple bugs ...
I'm redefining this meta-bug to reflect what bugs have been made to depend on it.
Summary: [meta] Problems with OS X 10.9 build setup and 'mach' → [meta] Problems with OS X 10.9 builds
Depends on: 932443
Hi, https://bugzilla.mozilla.org/show_bug.cgi?id=936977 is also the problem of mac os 10.9 mavericks. So, does anyone add bug 936977 to "Depends on"? I have no permission to modified "Depends on" column.
Ryan, can you uplift this to esr24 as well? Combined with those other two bugs, I can finally get a build. :-)
Ok, got it working again by doing brew uninstall autoconf213, then brew install autoconf213.
Status: NEW → RESOLVED
Last Resolved: 5 years ago
Resolution: --- → WORKSFORME
You need to log in before you can comment on or make changes to this bug.