crash in config.guess on Power Mac G5 under Mac OS X 10.4.6
Torbjorn Granlund
tege at swox.com
Fri May 19 15:46:58 CEST 2006
Vincent Lefevre <vincent at vinc17.org> writes:
On 2006-05-19 14:00:09 +0200, Torbjorn Granlund wrote:
> > You asked to have popup window for every Unix signal, and you get
> > what you ask for.
>
> Not for every signal (SIGQUIT isn't even concerned), only those that
> are generally caused by a crash. But a crash is a bug.
>
> Perhaps so, but a signal is not a crash.
Agreed, but an abrupt termination due to segmentation fault,
a bus error or a bad asm instruction is a crash... unless
the signal is trapped to handle the problem in a clean way.
Wouldn't this be a nice fix?
Please look at the code. Setting up signal handling seems risky,
and will likely cause more breakage than it fixes.
If you have a specific suggestion, please don't hesitate to write
a patch for me to look at, though.
All the configure stuff is, as you surely are more than aware of,
a tiptoe dance act.
--
Torbjörn
More information about the gmp-bugs
mailing list