This is the mail archive of the gcc-patches@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: cfg merge part 17 - loop datastructure updates


> > On Mon, 13 May 2002, Jan Hubicka wrote:
> > 
> > > 	* Makefile.in (OBJS): Add loop-new.o.
> > > 	(function.o, reg-stack.o, bb-reorder.o, tracer.o):
> > > 	Add dependency on basic-block.h.
> > > 	(loop-new.o): New.
> > 
> > Those changes don't seem to be in the patch sent.
> Oops, I will dig out the proper entry.
> > 
> > > 	* loop-new.c: New file.
> > 
> > Nor does this file.  And I don't think "loop-new.c" is a good name - if
> > the loop optimiser gets replaced again later, will you call the next one
> > "loop-new-new.c" or "loop-verynew.c"?  In time "-new" will be a bad
> > description.  Better would be another generic name, e.g. "loop-opt.c", or
> > a name with a version number rather than "-new", e.g. "loop-v2.c".
> > 
> > (This is presuming that the existing loop.c will go away in due course.
> > If not, then the name needs to be one effectively distinguishing the
> > function from loop.c.)
> My plan has been to replace loop by loop-new once loop is gone, but we can
> call it loop-cfg (but cfgloop is already taken) or loopopt...

Or prague-loop after haifa-sched :))

Honza
> 
> Honza
> > 
> > -- 
> > Joseph S. Myers
> > jsm28@cam.ac.uk


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