Re: gcc/Makefile.in targets 'compare' and 'gnucompare' behave differently

[Phil Edwards <phil at jaj dot com>]

On Fri, Oct 25, 2002 at 05:17:42PM -0700, Zack Weinberg wrote:
> On Fri, Oct 25, 2002 at 06:57:07PM -0400, DJ Delorie wrote:
> > 
> > >   for file in *.o; do \
> > >     test -f stage$stage/$file && (cmp --ignor...bad_compare) || true; \
> > >   done
> > 
> > That adds back in some of the overhead, unless we assume that the
> > shell is running the test command internally.
> 
> I think that's a safe assumption with any modern shell

I agree.

> > Otherwise, I can't think of a better way, unless we can check for
> > cmp's exit code being exactly 1 (it returns 2 if one of the files is
> > missing).
> 
> for file in *.o; do \
>   cmp ...
>   if [ $? -eq 1 ] ; then
>     # bad
>   fi
> done

I've just finished testing

  for file in *.o; do \
    cmp --ignore-initial=16 $$file stage$$stage/$$file > /dev/null 2>&1; \
    test $$? -eq 1 && echo $$file differs >> .bad_compare || true; \
  done

and this performs correctly when a file is missing from the stage*
subdirectory.  It also maintains the low overhead; 'make gnucompare'
continues to take about one-third to one-fourth the time of 'make compare'.

I'll post a patch tomorrow after some sleep.


Phil

-- 
I would therefore like to posit that computing's central challenge, viz. "How
not to make a mess of it," has /not/ been met.
                                                 - Edsger Dijkstra, 1930-2002