regex(3)/grot

[Kyle Evans]

On Fri, Apr 7, 2017 at 9:19 AM, Kyle Evans <kevans91 at ksu.edu> wrote:

> Hello!
>
> Over the next week or so, I plan to address inconsistencies in regex(3)
> handling of invalid bounds/atom constructs [1] between BREs and EREs, and
> found grot [2]. This appears to have once upon a time been the vehicle for
> regression testing of regex(3), but seems to have been replaced by the
> netbsd test suite (see: [3]).
>
> Does grot still serve a purpose, or can it go away? All of its tests seem
> to have been split out into individual tests grouped by functionality in
> the netbsd-tests suite [4], which is a model that seems ideal. On top of
> that, I'm not smart enough to actually build any of the targets in its
> Makefile and I don't feel a compelling urge to make it work on its own.
>
> Thanks,
>
> Kyle Evans
>
> [1] https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=166861
> [2] https://svnweb.freebsd.org/base/head/lib/libc/regex/grot/
> [3] https://svnweb.freebsd.org/base/head/lib/libc/tests/
> regex/Makefile?view=markup
> [4] https://svnweb.freebsd.org/base/head/contrib/netbsd-
> tests/lib/libc/regex/data/
>

Further inspection revealed that the netbsd-tests bits are actually just an
obviously derived/improved version of grot/, leading me to further believe
that grot/ can/should go away.