Re: More problems with `make -n' in automake-generated rules.

[Stefano Lattarini]

Hello Ralf.

On Saturday 06 November 2010, Stefano Lattarini wrote:
> On Saturday 06 November 2010, Ralf Wildenhues wrote:
> > Hi Stefano,
> > 
> > * Stefano Lattarini wrote on Sat, Nov 06, 2010 at 05:52:57PM CET:
> > > Hi Ralf, I've just spotted a bug in the patch ...
> > 
> > > > -             $output_rules .= "\$(srcdir)/$headerfile: 
> > > > \$(srcdir)/${derived}_vala.stamp\n".
> > > > -               "address@hidden test -f \$@; then :; else \\\n".
> > > > -               "\t  rm -f \$(srcdir)/${derived}_vala.stamp; \\\n".
> > > > -               "\t  \$(am__cd) \$(srcdir) && \$(MAKE) \$(AM_MAKEFLAGS) 
> > > > ${derived}_vala.stamp; \\\n".
> > > > -               "\tfi\n";
> > > > +             $output_rules .= "\$(srcdir)/$headerfile: 
> > > > \$(srcdir)/${derived}_vala.stamp\n"
> > > > +               . "address@hidden test -f \$@; then :; else rm -f 
> > > > \$(srcdir)/${derived}_vala.stamp; \n"
> > > ... here (missing "fi").  It causes a failure in `vala2.test'.
> > > 
> > > I'll install a fix later if you don't beat me (right now I'm doing other
> > > testing, and prefer not to be sidetracked by this issue).
> > 
> > That's what I get for forgetting one testsuite addition.  Thanks for
> > tracking this down, and please push the fix, ideally together with a
> > new test
> Is a new test really needed?  After all, I noticed the bug because it
> broke the pre-existing `vala2.test'...
I've gone ahead and comitted the fix without adding a new testcase.
Please let me know if you still think a testcase addition would be
valuable, in which case I'll add it in a follow-up patch.

Regards,
   Stefano