Re: FYI: Stack traces

[Alexandre Duret-Lutz]

>>> "Akim" == Akim Demaille <address@hidden> writes:

>>> "Akim" == Akim Demaille <address@hidden> writes:
 Akim> There is one thing that remains not satisfying: an m4 script that
 Akim> issues warnings and then an m4_fatal will not see the warnings, since
 Akim> they are processed after a _successful_ run of m4.  

 >> Why can't they be processed after a failed run?  Is that just
 >> because the xsystem() is not the right interface for the job,
 >> or is it something else?

 Akim> Maybe I misunderstood your suggestion

Reading your comments about m4_exit/m4_fatal I think the problem
is that I simply have not clue what the problem really is :)

I thought your algorithm was

   xsystem("m4 [...lots of args...] > tmpfile");
   process_traces_from("tmpfile");

hence the comment about xsystem(), which abort on error
even though process_traces_from could have displayed a few
warning before eventually conking out.
-- 
Alexandre Duret-Lutz