This is the mail archive of the gcc-bugs@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: Bug 14562 : copyrename & PRE


On Mon, 2004-03-15 at 15:45, Daniel Berlin wrote:
> >
> > Yeah, that does seem to fix it. We aren't going to run into similar
> > problems anywhere else are we? I see other places where names_match_p()
> > is used that its not clear to me won't result in similar things. but
> > then, i dont know PRE very well.
> 
> There's one function named reaching_def that probably has the problem, but
> it isn't used and i'm going to remove it.  The rest of the uses of
> names_match_p are lexical equivalence uses or unused (part of strength
> reduction that probably isn't worth it to implement since i'm working on
> changing algorithms), so they should be correct.
> 
> GVN-PRE won't use any of these functions (It's conceptually much
> simpler), or need to match names at all, since it's value based.
> I just have some speed and memory concerns about the algorithm, but i
> seriously doubt it could be *slower* than PRE is now.
> :)
> 
> The local PRE tree on my laptop has some speedup changes, so if you could
> commit that fix, i'd much appreciate it, so that i didn't have to copy a
> new tree.
> 
> I can also do it later tonight when i get home if you aren't in a rush.
> 

no rush, I dont have a tree in great condition at the moment either :-)
I wouldnt want to check it in without a bootstrap run :-)

Andrew


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