Re: [committed] Clean up boehm-gc/configure.in a bit

[andreas tobler <toa at pop dot agri dot ch>]

Hi Nathanel,

Nathanael Nerode wrote:

> > > * acinclude.m4: Move certain code to configure.in.
> > >       * configure.in: Receive code from acinclude.m4.
> > >       * aclocal.m4: Rebuild.
>
>
> > And this one breaks my solaris 9 multilib boehm-gc. The exact way as it does with my pending libffi update.
>
> Please tell me what this way was, since I didn't see the problem with
> your libffi change.

Sorry, I wasn't very clear :(

I was in hurry and forgot simpy to explain what breaks.

> > I couldn't figure out why. Unfortunately no time atm. And I did not commit/send the libffi patch yet since this issue is not solved.
>
>
> You're absolutely sure it's *this* one, not one of the three commits I made
> fairly soon after this one?  I don't want to go barking up the wrong tree.
> Please, if possible, verify which commit caused the problem.
>
> If it really was this commit, it's a shell code ordering problem of the
> most bizarre sort.

Yes, I'm pretty sure since this is the same failure I get on libffi.

The one Kaveh explained. It has to do with the multilib machinery.

Unfortunately I can't help much at the moment. I can offer to run tests 
on my sol9 box, but only in the evening.

We're under high pressure at work and I fried my powerbook so everything 
is slowed down.


Regards,

Andreas