Re: Bison 2.3

[H.Merijn Brand]

On Wed, 7 Jun 2006 20:38:38 +0200, Ralf Wildenhues <address@hidden>
wrote:

> * Paul Eggert wrote on Wed, Jun 07, 2006 at 08:27:17PM CEST:
> > "H.Merijn Brand" <address@hidden> writes:
> > 
> > > First two bloks of tests passed (bison 2.2 did not pass), after which I
> > > killed the tests, as AIX is soooooooo slow with the tests that I do not
> > > care to wait for it to crash, because of the fact that the test suite
> > > generates so many 'sh\d+.\d+' files on /tmp, that 'make check' cannot
> > > finish, as /tmp is flooded.
> > 
> > Thanks for letting us know.  The test suite problem is more an
> > Autoconf issue so I'm CC'ing this to bug-autoconf.
> 
> It's a long-known fact that the AIX /bin/sh is painfully slow due to
> this.  The GCC install manual has recommended setting CONFIG_SHELL to
> bash for years.

And if one does not have bash?

> We could consider special-casing AIX, or preferring bash in any case.
> Neither of which I'd like to do before 2.60, nor any other changes to
> the shell selection.  Adding a recommendation for setting CONFIG_SHELL
> to bash would be rather safe, though.

Sure.

> BTW, does the flooding lead to /tmp being full, /bin/sh running out of
> file descriptors, running out of file names, or did you just kill it at
> one point while it was still running?

My personal notes tell me both out of file descriptors, amd file system full
happened. Since then, I just kill the test suite once I see enough passes and
no fails

As a side note: I don't care about AIX at all. I would be the happiest man on
earth if all our IBM customers would switch to HP-UX or Linux. But in the
end, they still pay my salary, so I just have to do it :(

-- 
H.Merijn Brand        Amsterdam Perl Mongers (http://amsterdam.pm.org/)
using & porting perl 5.6.2, 5.8.x, 5.9.x  on HP-UX 10.20, 11.00, 11.11,
& 11.23, SuSE 10.0, AIX 4.3 & 5.2, and Cygwin.       http://qa.perl.org
http://mirrors.develooper.com/hpux/           http://www.test-smoke.org
                       http://www.goldmark.org/jeff/stupid-disclaimers/