[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: "rx"



On Tue, Oct 22, 2013 at 2:59 AM, Evan Hanson <evhan@xxxxxxxxxxxxxxx> wrote:
On 2013-10-20 12:26, John Cowan wrote:
> Alex Shinn scripsit:
> > How about renaming it "submatch"?
> >
> >   submatch?
> >   submatch-string
> >   submatch-start
> >   submatch-end
>
> I think that's confusing if you aren't actually using submatches.  I
> would favor either switching to "match" (and no, I don't think
> name collision is so important in this case), or using the explicit
> form "regexp-result", which describes what the object actually is.

I agree that this is confusing, even in the presence of "real"
submatches; you'd then have a submatch, with its own submatches
("subsubmatches"?), and no match to speak of. Some of the procedure
names would be OK, but as the record type name I don't think it's as
clear as "match".

Well, it _is_ a submatch, 0.  This may or may not
contain nested submatches 1 or higher, which may
themselves contains nested submatches, etc., but
I've never heard anyone call a nested submatch a
subsubmatch.

Anyway, the match- prefix both conflicts with the
pattern matching library and obscures any relation
to regular expressions, so that's not an option.

I could live with:

  rx-match-*
  regexp-match-*
  submatch-*

Anyone with a preference is free to email me their
(ordered) vote.

-- 
Alex