Hi, Ian,
Sorry for this delayed reply, just because bug chasing is time-cosumping to
a beginner. I think I found the root cause of the problem and post my ideas
to it for your reference.
As mentioned before in this thread, A new floating point type of 32 bit was
introduced into GCC, and this floating point type will cause mistakes when
it is variable argument list, just like single precision floating type. I
add some checks in c_type_promotes_to() in gcc/c-typeck.c and now there
would be a warning when this 32 bit floating type varialbes appeared in
varialbe argument list and abort when executed.
It seems that this problem itself has nothing to do with rs6000 backend and
I was confused by the symptom at first.
Thank you very much for all your previous reply.
Best Regards
----------------
Yao Qi
Bejing Institute of Technology
_________________________________________________________________
Express yourself instantly with MSN Messenger! Download today - it's FREE!
http://messenger.msn.click-url.com/go/onm00200471ave/direct/01/
