On Jan 18, 2006, at 12:28 PM, Chris Douty wrote:
On Jan 18, 2006, at 6:23 AM, Andrew Pinski wrote:
On Jan 18, 2006, at 6:16 AM, Eric Botcazou wrote:
OK, I can reproduce the failure.
Darwin specialists, what are we missing here?
Well it would be helpful if the reporter reran the link line with -v
so
people can see what is being linked. All I know it is not picking up
libgcc_s.1.dylib or picking up the wrong one.
I don't have access to 10.4.x to test with but I can test later
tonight on
10.3.9 which might (but I doubt it) expose the issue.
I think what is happening, although I don't know why, is that the
static libgcc.a is being linked so that the eh functions are not
present.
Yes the workaround is to add -fexceptions or -shared-libgcc to the
command line
when linking libgnat but I don't know if that is the correct fix or
some hacking
to config/darwin.h is needed. The LIBGCC SPECs assumes that you don't
use exceptions
when linking if have not supplied -mmacosx-version-min=10.4,
-fexceptions or
-shared-libgcc.
Thanks,
Andrew Pinski
