On 5/11/05, Jeff Squyres <jsquyres_at_[hidden]> wrote:
> On May 11, 2005, at 9:11 AM, Travis Spencer wrote:
>
> > On 5/11/05, Jeff Squyres <jsquyres_at_[hidden]> wrote:
> >> Can you specify what didn't work with the solaris threads?
> >
> > I uploaded the errors to
> > http://travisspencer.com/stash/tmp/lam_err.txt and included the specs
> > about my compiler and OS in my original post. What other information
> > would be helpful?
>
> Good enough -- thanks!
OK. If you do want more info about this or something else, just say
so. This project is extremely cool, and, if I can say thanks by
providing a little beta testing, I'm happy to do so.
> Those linker errors are just downright odd, though. I can think of
> only one reason this might happen: did you have a clean source tree?
> I.e., is it possible that there was some other kruft in your tree left
> over from a previous build?
Ya this is possible but unlikely. Between build attempts, I usually
remove the whole source tree, unpack the tarball anew, and restart
with a clean slate to avoid this sort of problem. That said, however,
I did see this a few times with different version of the source
(including 7.2), so maybe I got a little tired and skipped a step at
some point. If I see it again, I check things really carefully.
> I have to admit that we probably never tried gcc with the Solaris
> threads -- we only ever tested Solaris threads with the Solaris
> compilers.
GCC is all we use at PSU; even the non-CS departments that work on
Solaris use GCC. I think we have access to Sun's Forte (or whatever)
compiler, but no one uses it.
--
Regards,
Travis Spencer
Portland, OR USA
|
