bug#17246: grep-2.19 planning

[Paolo Bonzini]

Il 12/04/2014 21:35, Paul Eggert ha scritto:
> Paolo Bonzini wrote:
> > I'm not sure that the complexity of C programs will be smaller than the
> > complexity of grep <=2.18's machinery.
>
> Unfortunately that machinery wasn't clearly separated from everything
> else, and this was confusing.  Having it be a separate script will make
> it clearer.  A nice plus is that the scripts are tiny.

You are comparing apples-to-oranges.

I've said multiple times that I'm willing to put effort into simplifying 
the current grep/egrep/fgrep implementation.

And the matcher machinery, no matter how confusing it is, is not 
breaking anything, while the scripts might have unanticipated problems 
with relocatability, copying binaries around, --program-prefix and 
friends, and perhaps others I have not thought about.

> > One remaining problem with C wrappers is that they make the grep
> > installation non-relocatable and "non-separable".
>
> They are still relocatable in the sense that if you move grep + egrep +
> fgrep together to some other directory, they'll continue to work
> together in the typical case.  That's good enough for a hack to support
> an obsoleted interface.

Stop referring to it as obsoleted!  The GNU coding standards say that 
GNU "considers" standards, it doesn't "obey" them.

It does not matter if it is obsolete*d*, it matters if it is obsolete, 
i.e. if nobody is using it---and that's obviously not the case, 
especially for fgrep.

Paolo