[Gc] GC_malloc bug?

Boehm, Hans hans.boehm at hp.com
Mon Nov 29 11:40:36 PST 2004


It would be interesting to know whether you are trying to allocate
a particularly large object at the time.  GC_PRINT_STATS output
while it's in this loop might also be enlightening.

Hans

> -----Original Message-----
> From: gc-bounces at napali.hpl.hp.com
> [mailto:gc-bounces at napali.hpl.hp.com]On Behalf Of MenTaLguY
> Sent: Sunday, November 28, 2004 10:34 PM
> To: gc at napali.hpl.hp.com
> Subject: [Gc] GC_malloc bug?
> 
> 
> With Inkscape, we've been having trouble with hangs resulting from
> GC_malloc() going into apparently infinite loops, calling
> GC_collect_or_expand() repeatedly.  It seems to be more common on
> Windows than Linux, but happens on both platforms.
> 
> The heap isn't normally particularly large at the time (maybe 
> 15MB), and
> in our case we also have GC_no_dls turned on, so there really 
> shouldn't
> be that much to trace.
> 
> The one slightly unusual thing we do is that we currently allocate a
> very large number of small objects with GC_malloc_uncollectable(); I
> don't know whether that is likely to create a pathological 
> case for the
> collector.
> 
> We're using libgc 6.3; can anyone give me advice for further 
> diagnosing
> the problem?
> 
> -mental
> 


More information about the Gc mailing list