GC problem (finalisers) & comments

[Erik Schnetter]

On Wed, 9 Apr 1997, Bernie Solomon wrote:

[snip-a-lot]

> Actually I do wonder whether it wouldn't be simpler just to plug in the
> Boehm garbage collector instead (e.g. as Sather does). Has anyone done
> this? (I can't recall if it can run incrementally though).

That would seem reasonable to me.  A gargabe collector is a complicated
thing.  If you want to have it thread safe, include finalization, want to
behave it reasonable if a part of the main memory is swapped out, want to
have weak pointers or tell the gc that some memory region will not contain
pointers, etc... then why not use a generic malloc replacement package
like Boehm's?  It's available for a lot of platforms, well tested, *does*
run incrementally if you want it to and has all sorts of pipes and
whistles.  And it has a new mode called "JAVA"something to ensure Java
finalization semantics. 

-erik

-----
Erik Schnetter, erik.schnetter@student.uni-tuebingen.de