Re: Embedding Perl but redefining its malloc()

On Tue, Sep 29, 2009 at 08:56:37AM -0500, David Nicol wrote:
> On Sun, Sep 27, 2009 at 4:24 AM, Raphael Manfredi
> <Raphael_Manfredi@pobox.com> wrote:
> >  It's not necessarily a bad idea to make the two orthogonal
> > though, as memory allocation is something belonging to a lower layer
> > than dynamically binding language routines to OS routines.
> 
> One of the things that would become possible after organizing memory
> allocation to consistently use an API would be building perl with the
> Boehm garbage collector and ifdeffing out all the reference counting.
> For benchmarking purposes.

You're welcome to do that right now, but you'll find that no lexical
filehandles are closed when they go out of scope, and no DESTROY methods
will be called.

At which point Advice from Klortho #11912 clearly applies.

Possibly also #11961, but I don't know the performance characteristics of
the Boehm GC. :-)

Nicholas Clark

• Embedding Perl but redefining its malloc() by Raphael Manfredi
• Re: Embedding Perl but redefining its malloc() by Joshua ben Jore
• Re: Embedding Perl but redefining its malloc() by Raphael_Manfredi
• Re: Embedding Perl but redefining its malloc() by Joshua ben Jore
• Re: Embedding Perl but redefining its malloc() by Raphael_Manfredi
• RE: Embedding Perl but redefining its malloc() by Jan Dubois
• Re: Embedding Perl but redefining its malloc() by Raphael_Manfredi
• Re: Embedding Perl but redefining its malloc() by David Nicol
• Re: Embedding Perl but redefining its malloc() by Nicholas Clark
• Re: Embedding Perl but redefining its malloc() by Steffen Mueller
• Re: Embedding Perl but redefining its malloc() by Raphael_Manfredi
• Re: Embedding Perl but redefining its malloc() by Steffen Mueller
• Re: Embedding Perl but redefining its malloc() by Aristotle Pagaltzis