[webkit-gtk] Compiling webkitgtk on OSX (10.9.4)

Žan Doberšek zandobersek at gmail.com
Sun Aug 31 23:33:23 PDT 2014


Hi,

On Sun, Aug 31, 2014 at 3:15 PM, Devon Stewart <blast at hardchee.se> wrote:

> I'm getting a strange compilation failure trying to build webkitgtk 2.4.5
> with gcc-4.7 (which is stated as the low-end of the compiler requirement,
> and is already packaged in Homebrew). The error is as follows:
>
> > rawrbookpro:webkitgtk-2.4.5 blast$ make
> >   CXX      Source/WTF/wtf/libWTF_la-PrintStream.lo
> >   CXX      Source/WTF/wtf/libWTF_la-RandomNumber.lo
> >   CXX      Source/WTF/wtf/libWTF_la-FastMalloc.lo
> >   CXX      Source/WTF/wtf/libWTF_la-RefCountedLeakCounter.lo
> >   CXX      Source/WTF/wtf/libWTF_la-RunLoop.lo
> >   CXX      Source/WTF/wtf/libWTF_la-SHA1.lo
> >   CXX      Source/WTF/wtf/libWTF_la-SixCharacterHash.lo
> >   CXX      Source/WTF/wtf/libWTF_la-StackBounds.lo
> >   CXX      Source/WTF/wtf/libWTF_la-StringPrintStream.lo
> >   CXX      Source/WTF/wtf/libWTF_la-TCSystemAlloc.lo
> >   CXX      Source/WTF/wtf/libWTF_la-ThreadIdentifierDataPthreads.lo
> > Source/WTF/wtf/FastMalloc.cpp: In member function 'void
> WTF::TCMalloc_PageHeap::initializeScavenger()':
> > Source/WTF/wtf/FastMalloc.cpp:1953:56: error: expected
> primary-expression before '^' token
> > Source/WTF/wtf/FastMalloc.cpp:1953:57: error: expected
> primary-expression before '{' token
> > Source/WTF/wtf/FastMalloc.cpp:1953:80: error:
> 'dispatch_source_set_event_handler' was not declared in this scope
> >   CXX      Source/WTF/wtf/libWTF_la-ThreadSpecificWin.lo
> >   CXX      Source/WTF/wtf/libWTF_la-Threading.lo
> > make: *** [Source/WTF/wtf/libWTF_la-FastMalloc.lo] Error 1
> > make: *** Waiting for unfinished jobs....
>
> Now, I know dispatch.h is being found and used, because line 1952 is
>
> >     dispatch_source_set_timer(m_scavengeTimer, startTime,
> scavengeDelayInNanoseconds, scavengeDelayInNanoseconds / 10);
>
> and dispatch_source_set_timer is defined in the same header as
> dispatch_source_set_event_handler.
>
> jfpoole on Freenode's #macdev suggested gcc-4.7 might not understand
> blocks, but there doesn't seem to be anything in gcc-4.7's configure script
> that references blocks at all. I'm out of my element here, looking for
> other ideas.
>

Does GCC also require the -fblocks compiler flag to be specified? The few
examples that I found use that flag when compiling code with Clang.


>
> Thanks!
> -Devon
>
> PS: Apple's version of clang fails the clang >= 3.3 version test, despite
> being based on clang 3.4... kind of strange.
>
> > rawrbookpro:~ blast$ clang --version
> > Apple LLVM version 5.1 (clang-503.0.40) (based on LLVM 3.4svn)
> > Target: x86_64-apple-darwin13.3.0
> > Thread model: posix
>

The __clang_major__ and __clang_minor__ macros are used to check the Clang
version in Source/autotools/CheckSystemAndBasicDependencies.m4.
The logic there is really specific to upstream Clang versions, so it's
obvious why the 5.1 version would fail the check.

Cheers,
Zan


>
> _______________________________________________
> webkit-gtk mailing list
> webkit-gtk at lists.webkit.org
> https://lists.webkit.org/mailman/listinfo/webkit-gtk
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.webkit.org/pipermail/webkit-gtk/attachments/20140831/bf5bb43e/attachment.html>


More information about the webkit-gtk mailing list