Re: Give people with crashes more of a clue
Kyle Jones <kyle_jones(a)wonderworks.com> writes:
William M. Perry writes:
> Yoshiki Hayashi <t90553(a)m.ecc.u-tokyo.ac.jp> writes:
>
> > Could someone please test this patch before I send it to
> > xemacs-patches? You need to run autoconf after applying this
> > patch. Thanks.
>
> There should probably be some way to turn off these checks, just in case
> they get a false positive. Or a developer _wants_ to compile with known
> bad configurations to try and fix the
> problem. --i-know-what-im-doing-dammit=yes ?
To someone able to track down a compiler bug and provide workaround,
hacking the configure script is no hurdle at all. If they can't
manage that simple task, then maybe they don't know what they're
doing, dammit.
As for false positives, we want these to be reported, so adding a
switch to defeat the checks is counterproductive.
Another approach would be to boil the mis-compiled code into a short
bug-report class of program, and then gcc it during the configure
script. If it crashes, or whatever, then gcc has the bug. Then,
allow for '--ignore-failed-skip-syntax-test'.
--
-Justin
vallon(a)mindspring.com