bug-gnustep
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Installation of GNUstep Gui failed on Mac OS 10.4.2


From: Quentin Mathé
Subject: Re: Installation of GNUstep Gui failed on Mac OS 10.4.2
Date: Wed, 21 Sep 2005 11:14:52 +0200

Le 20 sept. 05 à 04:13, Adam Fedor a écrit :

On Sep 19, 2005, at 4:26 PM, Quentin Mathé wrote:

It just fails with -gui by crashing when make_services is run (I used FSF GCC 4.0.1).

[Thread 0 Crashed:
0   libobjc.A.dylib     0x909ad0f8 objc_msgSend + 24
1 libobjc.A.dylib 0x909ad044 _objc_fixup_protocol_objects_for_image + 180
2   libobjc.A.dylib     0x909a92e8 map_images + 728
3 dyld 0x8fe102d4 ImageLoaderMachO::doNotification (dyld_image_mode, unsigned, dyld_image_info const*) + 112 4 dyld 0x8fe062e8 dyld::notifyAdding (std::vector<ImageLoader*, std::allocator<ImageLoader*> >&) + 256 5 dyld 0x8fe0f5b8 ImageLoader::runInitializers (ImageLoader::LinkContext const&) + 280 6 dyld 0x8fe0381c dyld::initializeMainExecutable () + 96 7 make_services 0x00002138 _call_mod_init_funcs + 52 (crt.c:305)

Well, it's using the wrong libobjc. But I don't know why. This is after the patch? And both make and base were compiled again? Is the correct libobjc in DYLD_LIBRARY_PATH?

That's true, I didn't pay attention to this fact. Though everything has been compiled again this patch and I have used two days ago cvs version. DYLD_LIBRARY_PATH is correct too, it includes /gcc401/lib. I have the same problem when I compile every core packages one by one without relying on Startup facility.

The strange thing is I'm observing a bunch of warnings when linking against -base occurs, for example with autogsdoc :
 Compiling file AGSHtml.m ...
Linking tool autogsdoc ...
/usr/bin/ld: warning suggest use of -bind_at_load, as lazy binding may result in errors or different symbols being used symbol __objcInit used from dynamic library /gcc401/lib/libobjc-gnu. 1.dylib(linking.o) not from earlier dynamic library /usr/lib/ libobjc.A.dylib(libobjc.ppc.o) symbol _object_copy used from dynamic library /gcc401/lib/libobjc-gnu. 1.dylib(objects.o) not from earlier dynamic library /usr/lib/ libobjc.A.dylib(libobjc.ppc.o) symbol _object_dispose used from dynamic library /gcc401/lib/libobjc- gnu.1.dylib(objects.o) not from earlier dynamic library /usr/lib/ libobjc.A.dylib(libobjc.ppc.o)

With otool, I can see autogsdoc is linking the wrong objc lib too :
macmini:/cvs/gnustep/core/base/Tools/obj qmathe$ otool -L autogsdoc
autogsdoc:
libgnustep-base.dylib.1.11 (compatibility version 0.0.0, current version 1.11.1) /usr/lib/libobjc.A.dylib (compatibility version 1.0.0, current version 227.0.0) /usr/lib/libxslt.1.dylib (compatibility version 3.0.0, current version 3.11.0) /usr/lib/libxml2.2.dylib (compatibility version 9.0.0, current version 9.16.0) /usr/lib/libz.1.dylib (compatibility version 1.0.0, current version 1.2.3) /usr/lib/libSystem.B.dylib (compatibility version 1.0.0, current version 88.0.0) /sw/lib/libiconv.2.dylib (compatibility version 5.0.0, current version 5.0.0) /usr/lib/libmx.A.dylib (compatibility version 1.0.0, current version 92.0.0)

-base itself is correctly linked however :
macmini:/cvs/gnustep/core/base/Source/obj qmathe$ otool -L libgnustep- base.dylib
libgnustep-base.dylib:
libgnustep-base.dylib.1.11 (compatibility version 0.0.0, current version 1.11.1) /gcc401/lib/libobjc-gnu.1.dylib (compatibility version 2.0.0, current version 2.0.0) /usr/lib/libxslt.1.dylib (compatibility version 3.0.0, current version 3.11.0) /usr/lib/libxml2.2.dylib (compatibility version 9.0.0, current version 9.16.0) /usr/lib/libz.1.dylib (compatibility version 1.0.0, current version 1.2.3) /usr/lib/libSystem.B.dylib (compatibility version 1.0.0, current version 88.0.0) /sw/lib/libiconv.2.dylib (compatibility version 5.0.0, current version 5.0.0) /gcc401/lib/libgcc_s.1.0.dylib (compatibility version 1.0.0, current version 1.0.0) /usr/lib/libmx.A.dylib (compatibility version 1.0.0, current version 92.0.0)

May be there is some problem with tool.make ?

Thanks,
Quentin.

--
Quentin Mathé
qmathe@club-internet.fr





reply via email to

[Prev in Thread] Current Thread [Next in Thread]