This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

[PING] Re: [PATCH] ARM half-precision floating point, 5/8 (detect constant overflow)


Richard Guenther wrote:

If Joseph thinks this is ok the patch is ok if it passes bootstrap / testing on another target with lacking NaN/Inf support.

2009-04-15 Sandra Loosemore <sandra@codesourcery.com>


       gcc/
       * fold-const.c (fold_convert_const_real_from_real): Check for
       overflow.

I've still been unable to set up a test machine for this locally, and I did not get a response from the spu port maintainers when I asked for testing help a few weeks ago.


Is it reasonable for me to go ahead and commit the patch anyway, with the understanding that it can be reverted and/or re-worked if it causes problems for spu? Personally, I'm confident that Joseph has a good handle on the correctness issues of this change. I'll also note that you have compile with -pedantic to get the overflow warning at all, and the patch has no effect on the value stored in the overflowing constant. So it seems very low-risk, to me.

Here's a pointer to the original patch post:

http://gcc.gnu.org/ml/gcc-patches/2009-04/msg01126.html

-Sandra


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]