Re: [Nmh-workers] Proposed solution to Debian Bug#143485

[Harald Geyer]

> > > >             execlp (mhlproc, mhl, "-form", filter, "-noclear", NULL);
> > > > -           fprintf (stderr, "unable to exec ");
> > > > -           perror (mhlproc);
> > > > +           write(2, "unable to exec ", 15);
> > > 
> > > Reminds me of Hollerith constants, i.e. having to manually count the
> > > length of a string literal.  Can't we just stick with the fprintf(),
> > > perror(), etc., and add judicious use of fflush()?
> > 
> > No. fflush() won't help since stderr should be unbuffered anyway. But
> > the vfork()ed child might still overwrite the parents data. Of course
> > we could stick with fprintf() if we dropped vfork()...
> 
> Apologies, I obviously haven't been following closely enough;  I thought
> we were getting rid of vfork() calls in the patch.

Of course we could do that too, but so far nobody has pushed that forward
and I didn't want to change only this call of vfork(). The important 
thing, which actually fixes the initial bug, is that we call fflush()
*before* replfilter().

Harald