Re: GC brokenness in `master'

[Ludovic Courtès]

Hi Neil,

Neil Jerram <address@hidden> writes:

> address@hidden (Ludovic Courtès) writes:

>> After some manual bisecting, I found the offending commit (dated
>> 2008-08-16):
>>
>>   
>> http://git.savannah.gnu.org/cgit/guile.git/commit/?id=82ae1b8eb3413e6be6bd2aa032986fc7782e85ac
>
> But presumably this doesn't matter much, if we are going to switch to
> BDW-GC.  What's your latest thinking on that?

Yes, I'm still considering switching GCs anyway, but I wanted to know
what happened exactly (and whether it was really a GC issue, which
wasn't 100% obvious).

While doing this, I remembered that Valgrind can't be used with BDW-GC,
which can be annoying.  I'll email the GC list about that.

>> Strangely enough, the "simple subr" case, which is not GC-intensive, is
>> slower with BDW-GC.  This needs further investigation...
>
> It's not very much slower - and IMO, not serious enough to be a
> significant factor against BDW-GC.

Agreed.

> But it's nicer to understand things fully, so I agree that
> investigation would be good too.

Yes.

Thanks,
Ludo'.