On Sat, 17 Nov 2007, chromatic wrote:
> On Friday 16 November 2007 06:22:27 Andy Dougherty wrote:
>
> > For the first time ever, I was able to successfully complete these two
> > tests:
> >
> > t/compilers/pge/p5regex/p5rx.................ok
> > 355/960 skipped: various reasons
> > t/compilers/pge/p6regex/01-regex.............ok
> >
> > Somebody (most likely chromatic) has plugged a very big leak.
> > (Unfortunately, I haven't been testing parrot regularly, so all I know
> > is that it's been somewhere in the past two months (i.e. since my last
> > PLATFORMS update of 17 September 2007)).
>
> Great! I'd love to know which patch it was, but it's nice to know that our
> memory use is somewhat better now.
>
> Are you able to compile and test most of the rest of Parrot now? Are there
> any other tests or examples which run out of memory? My Valgrind-fu keeps
> improving....
Alas, binary searching has proven to be painfully slow, and just painful
to boot.
As for other tests, nothing appears to be gobbling up memory unreasonably.
Most of the test suite passes (maybe 30 test failures or so) but it seems
to hang reliably in pmc/threads.t test 7, so I haven't completed a clean
run yet (and hence don't have a PLATFORMS patch). I also haven't
successfully diagnosed that problem. I haven't even looked at 'make
languages' in ages.
--
Andy Dougherty [EMAIL PROTECTED]
