Re: [RFC] GC: external thread suspension

[Keith Seitz <keiths at redhat dot com>]

Boehm, Hans wrote:

> One other problem is that this is currently only implemented for
> linux-like settings.  We eventually need win32 and Darwin
> implementation, at least if this is going to be a public part of the
> interface.

Yeah, I haven't looked at anything but linux. Eventually, we'll want it 
all, I would guess. Maybe someone will step up to the proverbial plate. 
I only have hardware to do win32 (albeit *painfully* slowly) and x86 linux.

> I think that in the current gcj version, Solaris would also need a
> separate implementation.  But that's no longer needed for 7.0, and hence
> almost certainly not worth doing.
>
> Also note that the win32 threads code was partly rewritten for 7.0.  For
> this piece, a patch against 7.0 would be much more interesting than one
> against the current gcj tree.

Noted.

> I don't think any of this is essential for including it.  But if it's
> included in gc.h, it needs to be documented as platform-specific.

When I submit a final patch, I will do this.

Thanks,
Keith