On Tue, 24 Apr 2007, chromatic wrote:

> On Tuesday 24 April 2007 13:30, Andy Dougherty wrote:
> 
> > > Andy, could you update to r18323, remove the -G's, and see if it
> > > now runs to completion on your Solaris box?
> 
> > Thanks for the heads-up.  I'm afraid testing will have to wait until
> > tomorrow.  The rsync server at svn.perl.org is still down, and the
> > snapshot on svn.perl.org hasn't picked up that patch yet.
> 
> Here's the patch (before I added comments); it should apply cleanly to 
> anything in the past week or two.

Thanks.  That indeed does help.  Without this patch, 

         ./parrot  t/compilers/pge/p5regex/p5rx.t

would only get to test 185.  Today, I made it to test 227.  Granted, it 
took 22 minutes to get there, but it did get there!

So, good detective work.  I think you've plugged an important leak.  
Alas, there are still (at least) two problems:

1.  It's still leaking memory badly for me.  It improved from test 185
to 227, but that's a long way from completing all 960 tests.

2.  Garbage collection really slows the program down (I observed factors
of 10 difference in speed with and without -G), and I have a vague
unsupported suspicion that the slowdown grows faster than linearly with
the allocated memory.

Still, keep up the good work!

-- 
    Andy Dougherty              [EMAIL PROTECTED]

Reply via email to