Pull request: Raise SIGFPE instead of abort in __gmp_exception`

Yichao Yu yyc1992 at gmail.com
Thu Jan 28 18:44:54 UTC 2016


On Thu, Jan 28, 2016 at 12:45 PM, Torbjörn Granlund <tg at gmplib.org> wrote:
> nisse at lysator.liu.se (Niels Möller) writes:
>
>   > I guess I'm not familiar with hg/gmp work flow. Is checked in
>   > different from commited to the repo? I don't see the change appearing
>   > in the repo[1] and my local fresh clone doesn't have it either.
>   >
>   > [1] https://gmplib.org/repo/gmp/
>
>   Maybe there's something wrong with the mirroring to the public repo?
>
> Indeed.  The freebsd-update utility is truly impressive.  It deleted
> lots of stuff, including /usr/bin/ssh.
>
> Now mirroring works again.

Thanks. I see the change now.

However, if I understand correctly, this may cause a regression on OSX
x64 for multi-threaded programs since the signal may be delivered to
any thread instead of the current thread. Unconditionally trying the
old approach first should solve this issue.

>
> --
> Torbjörn
> Please encrypt, key id 0xC8601622


More information about the gmp-devel mailing list