[PATCH]: Improved handling of COND_EXPR in middle-end passes

Roberto COSTA roberto.costa@st.com
Wed Nov 22 11:23:00 GMT 2006


Hello,
I ping about a patch I posted last week, which got no reaction at all:
http://gcc.gnu.org/ml/gcc-patches/2006-11/msg00922.html

The patch improves handling COND_EXPR nodes used as expressions (I mean, 
not as statements) in a few middle-end passes.
Basically, it consists in enhanced optimization opportunities and a bug fix.

I tested it, I bootstrapped the compiler with no regressions on 
i686-pc-linux-gnu.
I don't see disruptive changes in the patch, most of them are just added 
lines that handle COND_EXPR nodes in passes that weren't doing anything 
special about them.
Tell me if you can approve it or if I need reviewers to scan each line 
of code... and I'll do it, what I wish is to be able to commit these lines.
As a matter of fact, the patch is a collateral effect of a study I did 
about COND_EXPRs. I'm no longer working on that and I'm concerned that 
the patch will be lost if too much time passes. (and I guess it's not 
very exciting for anybody to see the result of her own work lost because 
of lack of reviewers :-( ).

Cheers,
Roberto
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: gcc_cond_patch3
URL: <http://gcc.gnu.org/pipermail/gcc-patches/attachments/20061122/38a306d1/attachment.ksh>


More information about the Gcc-patches mailing list