------- Comment #4 from aoliva at gcc dot gnu dot org 2008-06-26 18:10 -------
I have not tried 4.4. Anyhow, I have looked further into this issue, and I
came to the conclusion that a memory clobber (that should have always been
there anyway) will get things right. I cannot tell for sure that it guarantees
GCC will not ever get smart and delete the initializer because it cannot see
that it is used (the ASM needs not be a VUSE if it is marked as using only the
pointer, after all), but I find it unlikely that this will ever break. So,
nevermind... Just another case of inlining exposing latent problems.
--
aoliva at gcc dot gnu dot org changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|UNCONFIRMED |RESOLVED
Resolution| |INVALID
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36639
