RE: building pam and x11 in phase2

> > (a) it makes it easier to use the
> > intermediate as a testbed, for example
> > to go build devchk against the lsbsi
> > we can't use the phase3 (final), since
> > it has no build tools; currently we
> > also can't use phase2 because it doesn't
> > have everything required by the LSB.
> 
> unless things have changed, phase2 was not lsb complient 
> enough for testing when I last played with it.  

> Having a tarball of the build tools you could plop on a 
> phase3/final would not be a bad thing.  Stuart and I
> had talked about it.

No, it's not, really.  I didn't mean full-blown
testing, just a way to build things that might
need a build environment - specifically the
aforementioned devchk (or hdrchk, if you prefer,
seems to go by both names).  Adding a build drop-in
is probably a better approach. 

> > (b) there's a chance that the support
> > for pam will be better in phase3 if
> > it's present when various packages are
> > configured - and configuration is done
> > against what's found in phase2.
> 
> This makes sense although it would be worthwhile to find out 
> if there is truth to the hypothesis.

I'm trying a test run against an lsbsi built
using this sequence.  Don't know if I have
the patience to go examine individual binaries
to see if they've configured themselves to use
X11.  We've finally got some pam tests (I should
really see if I can get those built up and run
inside the lsbsi), but in the meantime we were
using Samba built for pam as a test, and I saw
some problems with this on the lsbsi, at least
on Itanium, to the point where I was concerned
that "just building pam" wasn't enough. Although
since I didn't see the same problems on ia32,
I'm less convinced it's a system build/integration
issue...