This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: PATCH: PR middle-end/24901: Alignment bug with --checking=all
> | > Steve, seems like this is relevant to all active branches
> | > including 3.4, right? Are you going to check it in those also?
> | > --Kaveh
> |
> | This patch is relevent to the other branches but I don't know if it
> | is worth putting it in. My thinking was that since it is the main
> | line where most work happens, it is on the main line where people
> | (including me) are most interested in building with --checking=all.
> | I don't object to putting it on the branches (especially 4.1) if
> | people think it is worthwhile.
>
> I believe it qualifies for 4.1 at least.
> -- Gaby
IMHO "where most work happens" is not necessarily the criteria to
decide branches where we should put this fix. First, it has zero
chance of breaking the released compiler because they ship with "fold"
checking disabled and the relevant code is #ifdef'ed out. Instead we
should see if it would help testing GCC and discovering bugs, and I
think the answer is yes.
If my box is sufficiently fast, I try to add checking options even for
3.4 and 4.0, e.g.:
http://gcc.gnu.org/ml/gcc-testresults/2006-01/msg01305.html
http://gcc.gnu.org/ml/gcc-testresults/2006-01/msg01304.html
Many moons ago, I tried "fold" checking and it didn't work. I'd
consider using "fold" checking on those branches as well assuming your
patch fixes it. (Right now my "fast" box is unavailable so it'll take
a long while for me to check.) But I do think this patch is a good
candidate for backporting to all active branches.
--Kaveh
--
Kaveh R. Ghazi ghazi@caip.rutgers.edu