Re: make -j # bootstrap fails

[Matt Schalit <mschalit at pacbell dot net>]

DJ Delorie wrote:
> 
> > Building gcc on multiprocessor machines is a fine thing - when it works.
> >
> > Sometimes it does, sometimes it dies and invoking
> > make bootstrap
> > afterwards finishes the bootstrap w/o error.
> >
> > Any explanation for this behaviour?
> >
> > # make -v
> > GNU Make version 3.79.1, by Richard Stallman and Roland McGrath.
> > Built for hppa2.0w-hp-hpux11.00
> 
> Could you be more specific, please?  How many processors does your
> machine have?  What were the exact error messages you saw?  Which gcc
> (2.x, 3.0, cvs) were you trying to build?  Which -j value did you
> give?




I had the same problem where gmake -j 10 bootstrap
failed sometimes on my dual PII, i586-sco-sysvuw7.1.1,
compiling gcc-3.0, where it succeeded on 2.95.3.

It's not repeatable on my system.
I didn't save the gmake output when it did happen.
I can't make it happen anymore :-/
I just tried 5 times.

The only difference is that I installed autogen,
but the make output doesn't contain the word autogen,
so I don't know.

What I can tell you is that autogen _was_ accessed
during the time that I was doing these tests,
according to    ls -lu /usr/local/bin/auto*


Regards,
Matthew