FYI: large number of objc failures on i686-linux-gnu

[Jakub Jelinek]

On Sun, Jan 06, 2002 at 10:54:22AM -0800, H . J . Lu wrote:
> Well, the linker message had always been there before. It is just that
> the old objc test framework didn't catch the linker message:
> 
> Executing on host: /export/build/gnu/gcc-mips/build-i686-linux/gcc/xgcc -B/export/build/gnu/gcc-mips/build-i686-linux/gcc/ /home/hjl/work/gnu/src/gcc-mips/gcc/gcc/testsuite/objc/execute/accessing_ivars.m  -w  -O -I/home/hjl/work/gnu/src/gc c-mips/gcc/gcc/testsuite/../../libobjc -L/export/build/gnu/gcc-mips/build-i686-l inux/i686-pc-linux-gnu/libobjc/.libs -lobjc  -lm   -o /export/build/gnu/gcc-mips /build-i686-linux/gcc/testsuite/accessing_ivars.x    (timeout = 300)
> /usr/bin/ld: warning: type and size of dynamic symbol `__objc_class_name_Object' are not defined^M
> output is: /usr/bin/ld: warning: type and size of dynamic symbol `__objc_class_name_Object' are not defined^M
> 
> PASS: objc/execute/accessing_ivars.m compilation,  -O
> PASS: objc/execute/accessing_ivars.m execution,  -O 
> 
> The new one does. As I said, it is an objc compiler bug.

Yeah, it is an objc compiler bug, which I've fixed in
http://gcc.gnu.org/ml/gcc-patches/2001-12/msg02645.html
(bootstrapped/regression tested on {i386,ia64,alpha}-redhat-linux).
Ok to commit?

	Jakub