Re: Re: What is the future of gcc-2.95.3-prerelease ?

[tprinceusa at mindspring dot com]

I'm in the middle of a testsuite run of gcc-2.97 on cygwin; I'll go later and see if the old -march=pentiumpro -ffast-math bugs are still there, as they were last week.  I've never had an answer as to whether that combination of options is supported; it does work for boot-strapping now, even though that has been stated as being supported only for default options, but obviously can't be depended upon in general.  The last time I tried to use the gnats reporting system, I got kicked out for lack of credentials, so I've never re-submitted these bugs myself since sending a test case to gcc-bugs.  Still, if 2.97 can't handle relatively simple options such as those and -Os, I'll continue to prefer the 2.95 series.  Do bugs submitted on gcc-bugs expire after a few months if no one succeeds in getting the case into the test suite, and they don't crop up in multiple applications?  I do remember a statement just a few days ago that bugs aren't important if they only show up in a few peo!
!
ple's applications.

Zack Weinberg <zackw@stanford.edu> wrote:
> On Thu, Feb 15, 2001 at 04:55:39PM -0500, tprinceusa@mindspring.com wrote:
> 
> I would prefer to use 2.95.3 over any other version; it runs my g77
> cases faster on linux that either 2.95.2 or 2.97, and doesn't share
> the long-standing bugs introduced early in 2.96 and perpetuated in 2.97.

Which long-standing bugs would those be?  We can't fix them if no one
tells us about them.

zw