Re: [PATCH] Do not recurse walking the CFG in phiprop

[Richard Guenther <rguenther at suse dot de>]

On Tue, 21 Apr 2009, Gerald Pfeifer wrote:

> On Tue, 21 Apr 2009, Dave Korn wrote:
> > Richard Guenther wrote:
> >> As suggested by Paolo this fixes/generalizes get_all_dominated_blocks
> >> and uses it to avoid recursion in the CFG walk in phiprop.
> > Thanks, this totally solves my stack problem.
> 
> Excellent.  Thanks for pushing, Dave, and fixing Richi!
> 
> libjava/classpath/gnu/javax/swing/text/html/parser/HTML_401F.java
> exploding is something I have been reporting in Feburary and then
> October 2007 already, not because of stack overflows but since it
> was the reason why GCC could no longer build within 512MB of main
> memory on i386-unknown-freebsd (needing some 700MB).
> 
> Is this something that could be retrofitted into GCC 4.4 as well,
> Richi?  (The regression would be failure to bootstrap within half
> a GB of main memory.)

I don't think the patch fixes that.  Or does it?

Thanks,
Richard.