* Stefano Lattarini wrote on Sun, Jan 02, 2011 at 04:31:18PM CET:
> On Sunday 02 January 2011, Ralf Wildenhues wrote:
> > Feel free to push the patch series
> > (as far as OKed) on a new branch based off of maint,
> >
> Hmmm.. currently my private branch is based off of master, since I
> thought it would be wrong to apply such slightly backward-incompatible
> changes to maint. Rebasing on maint would require some modifications
> to the patches in this series, and/or preliminary changes to maint
> (which still lacks definition of e.g. `$original_AUTOMAKE' in the
> testsuite).
>
> Would be OK with you if I'd just push to a new public branch off of
> master?
Yes, sure. Sorry, I didn't pay close enough attention.
> > > +{
> > > + set +x
> >
> > Curious: why turn off tracing here?
> >
> Because, while I was testing and debugging the test cases, those traces
> turned out to be useless and a little confusing; it was much more useful
> and clear to just display the contents of the edited file. And once I
> was done with debugging, I saw no reason to revert this temporary
> disabling of shell traces.
Whichever you prefer.
> > > + sed <$2 >$2-t -e "s|^\\(AUTOMAKE_OPTIONS\\) *=.*|\\1 = $1|" \
> > > + -e "s|^\\(AM_INIT_AUTOMAKE\\).*|\\1([$1])|"
> > > + mv -f $2-t $2
> > > + set -x
> > > + cat $2
> >
> > To avoid caching please 'rm -rf autom4te.cache' here.
> >
> OK, done here and in warnings-overide.test too.
>
> Well, with two differences:
> - I've used 'rm -rf autom4te*.cache', for consistency with what is
> used and suggested in other tests (e.g., asm.test).
> - I've added various 'rm -rf autom4te*.cache' calls in the main code,
> not a single one in set_strictness() and/or set_wanings().
OK.
Thanks,
Ralf
