Re[4]: [Gc] Re: Partially scanning a VM's call stack

Ivan Maidanski ivmai at
Wed Aug 19 04:52:02 PDT 2009


Juan Jose Garcia-Ripoll <juanjose.garciaripoll at> wrote:
> 2009/8/19 Ivan Maidanski <ivmai at>:
> >> This looks like the easiest solution.  OTOH `GC_gcj_malloc ()' appears
> >> to be undocumented, so I'd rather go for GC_new_kind + GC_MAKE_PROC.
> >
> > It's public (exported from gc_gcj.h). It's used in GCC/GCJ (and Mono AFAIK) projects.
> The fact that something is exported in a header and commented in the
> mailing list does not make it look like a stable feature of the
> garbage collector. In particular, I can barely gather how those
> descriptors work from your emails.

AFAIK, this is supported since 5.0 and there's no plans to change its API at pesent.
Of course, this is just my opinion (including the fact it's public).
(And, of course, my free advices could lead someone/everyone to nowhere - I apologies.)

> If the GCJ methods for collection and allocation are useful, supported
> and with a stable interface, then it should be documented somewhere by
> the library. Otherwise it is hard for non-GCJ/GCC developers
> (including myself) to even figure out what the benefits and proper
> ways of using it are, or whether they will survive a future change of
> mood of Mono/GCJ developers :-)

Absolutely true. Are where any volunteers? It would be good to place it to SourceForge "Feature Request" tracker (just not to be lost in the ML).

> Indeed it would be nice for the boehm-gc project to have a wiki where
> uses of the garbage collector and techniques are documented, if not by
> the developers, at least by skilled users. Developers could then pass
> by and sanctify those documents.
> Just my 2c, talking as a long time user who had (and has) a hard time
> figuring out optimal ways to use this otherwise wonderful library.
> Juanjo
> Instituto de FМsica Fundamental, CSIC


More information about the Gc mailing list