Re: [cfe-dev] LLVM / Clang 3.0 rc1 Binaries Available

[Sebastian Reitenbach]

On Thursday, October 27, 2011 10:14 CEST, Richard Frith-Macdonald 
<richard@tiptree.demon.co.uk> wrote: 
 
> 
> On 21 Oct 2011, at 11:02, David Chisnall wrote:
> 
> > FYI: Please test on platforms you care about and file bugs...
> 
> Using svn trunk for make/base/libobjc2 and using the 3.0 rc1 on debian 64bit 
> I get a failure ~(uncaught exception) in the NSURL tests in basic.m ...

I've seen it too on OpenBSD, but thought it was specific to OpenBSD, have tried 
to work it out with David, and ended up with a bug report here:
http://llvm.org/bugs/show_bug.cgi?id=11165

Sebastian

> 
> I don't know whether it's a compiler bug or a libobjc2 bug though ...
> 
> Passed test:     basic.m:96 ... resourceSpecifier of /usr is //localhost/usr/
> /home/richard/setups/clang/devmodules/core/base/Tests/base/NSURL/obj/basic: 
> Uncaught exception NSInvalidArgumentException, reason: -[NSURL 
> initFileURLWithPath:isDir:]: unrecognized selector sent to instance 0x9bed98
> 
> In gdb I get ...
> 
> (gdb) where
> #0  0x00007ffff6a3a165 in *__GI_raise (sig=<value optimized out>)
>     at ../nptl/sysdeps/unix/sysv/linux/raise.c:64
> #1  0x00007ffff6a3cf70 in *__GI_abort () at abort.c:92
> #2  0x00007ffff7783a74 in _terminate ()
>     at 
> /home/richard/setups/clang/devmodules/core/base/Source/NSException.m:751
> #3  0x00007ffff7783a2d in _NSFoundationUncaughtExceptionHandler (
>     exception=0x9ed858)
>     at 
> /home/richard/setups/clang/devmodules/core/base/Source/NSException.m:775
> #4  0x00007ffff7781d49 in callUncaughtHandler (value=0x9ed858)
>     at 
> /home/richard/setups/clang/devmodules/core/base/Source/NSException.m:795
> #5  0x00007ffff7214522 in objc_exception_throw (object=0x9ed858)
>     at eh_personality.c:93
> #6  0x00007ffff7782f19 in -[NSException raise] (self=0x9ed858, 
>     _cmd=0x7ffff7d2a250)
>     at 
> /home/richard/setups/clang/devmodules/core/base/Source/NSException.m:963
> #7  0x00007ffff7782110 in +[NSException raise:format:arguments:] (
>     self=0x7ffff7d2a0e0, _cmd=0x7ffff7d2a2f0, name=0x7ffff7d29a10, 
>     format=0x7ffff7d9a0d0, argList=0x7fffffff7b60)
>     at 
> /home/richard/setups/clang/devmodules/core/base/Source/NSException.m:849
> #8  0x00007ffff7781ff3 in +[NSException raise:format:] (self=0x7ffff7d2a0e0, 
>     _cmd=0x7ffff7d9a510, name=0x7ffff7d29a10, format=0x7ffff7d9a0d0)
>     at 
> /home/richard/setups/clang/devmodules/core/base/Source/NSException.m:835
> #9  0x00007ffff7937b1d in gs_objc_msg_forward2 (receiver=0x9bed98, 
>     sel=0x621960)
>     at 
> /home/richard/setups/clang/devmodules/core/base/Source/GSFFIInvocation.m:196
> #10 0x00007ffff7937818 in gs_objc_msg_forward3 (receiver=0x9bed98, 
> op=0x621960)
>     at 
> /home/richard/setups/clang/devmodules/core/base/Source/GSFFIInvocation.m:243
> #11 0x00007ffff7220d26 in objc_msg_lookup_internal (receiver=0x9bed98, 
>     selector=0x621960) at sendmsg2.c:100
> #12 objc_msg_lookup (receiver=0x9bed98, selector=0x621960) at sendmsg2.c:375
> #13 0x0000000000407656 in main () at basic.m:99
> (gdb) 
> 
> Stepping up through the stack I see it's trying to look up a value for a nul 
> receiver ...
> 
> 243     slot->method = gs_objc_msg_forward2(receiver, op);
> (gdb) print *receiver
> $3 = {isa = 0x0}
> (gdb) 
> 
> but stepping further up it seems we have the real receiver ...
> 
> (gdb) 
> #12 objc_msg_lookup (receiver=0x9bed98, selector=0x621960) at sendmsg2.c:375
> 375           Slot_t slot = objc_msg_lookup_internal(&self, selector, nil);
> (gdb) print self
> $4 = (id) 0x9bed98
> (gdb) print *self
> $5 = {isa = 0x7ffff7d7a1b0}
> (gdb) print *self.isa
> $6 = {isa = 0x7ffff7d79db0, super_class = 0x7ffff7d4d190, 
>   name = 0x7ffff79a84e2 "NSURL", version = 0, info = 93, instance_size = 40, 
>   ivars = 0x7ffff7d79d20, methods = 0x7ffff7db7ea0, dtable = 0x76f280, 
>   subclass_list = 0x0, sibling_class = 0x7ffff7d78680, 
>   protocols = 0x7ffff7dddbf0, extra_data = 0x76f750, abi_version = 1, 
>   ivar_offsets = 0x7ffff7d7a190, properties = 0x7ffff7dd5030, 
>   strong_pointers = 1, weak_pointers = 1}
> (gdb) 
> 
> Is this the compiler passing the wrong value, or possibly something wrong in 
> the objc library?
> _______________________________________________
> Discuss-gnustep mailing list
> Discuss-gnustep@gnu.org
> https://lists.gnu.org/mailman/listinfo/discuss-gnustep