This is the mail archive of the gcc@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Optimization patches


> On Fri, 2003-03-21 at 11:49, Jan Hubicka wrote:
> > > 
> > > Jan --
> > > 
> > > Janis and her magic regresion-finding scripts (thanks Janis!), seem to
> > > have marked you as the culprit for several of the outstanding
> > > high-priority PRs, including PR 8634, PR 7916, and others.  
> > 
> > Zdenek has sent patch for 8634 - it is nasty problem in addressof
> > ellimination versus constant memory initializers.
> 
> Do you have a URL?  I would like to review outstanding patches if I can;
> that's an easy way to get the bug count down. :-)
> 
> > Does "nice" error message from elliminate_virtual_regs looks sane?
> 
> If the input is invalid, I think it's perfectly reasonable to fix it as
> you suggest.  People who play with inline asm/register asms are

It is invalid.
I will dig out the patch I made for that in december then.

> hard-core; if they get a hard-core error message that's OK.  But it
> should be an error message, not an internal error.  I'm agreeing with
> you. :-)
> 
> Also, would you change any such PRs to be marked as ice-on-illegal, even
> before we do this fix?

I really should force myself to updated the PRs of bugs I analyze/make
patch for as the patch often jsut gets missed.  Thanks for reminding me
:)

Honza
> 
> Thanks!
> 
> -- 
> Mark Mitchell <mark at codesourcery dot com>
> CodeSourcery, LLC
> 


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]