Front page | perl.perl5.porters |
Postings from July 2016
* Karl Williamson <public@khwilliamson.com> [2016-07-06 21:48]:
> I'm looking for some more ideas.
It seems to me you framed the problem as being about runs, and then got
stuck in that frame for its design, which would seem to imply that this
has something to do with quantifiers.
Really though, what you want is to match another character that is like
some other already-matched character. That doesn’t need to involve any
sort of quantifier.
The pattern matching language already has a way of saying “give me more
of the same as this”: you capture the thing you want to match, then you
use a backref to match more of it.
(?x) ( [=,.-] {2} ) \1+
So the most generic primitive for match script runs would be one which
takes a template string from either a literal or a backreference, looks
at the script of each character, then matches a sequence of characters
that has the same sequence of scripts. (Very vaguely this is akin to the
samecase/samemark methods in Perl 6.) Something like
(.)(*script:\1)*
(No idea what the spelling of the primitive ought to look like. I keep
thinking this belongs into the \p{} syntax, but it fits oddly there for
multiple reasons.)
That’s obviously too clunky in daily use, so it’s clearly not the end
point of the design. Some flag may turn out to be the right answer. But
I think this is the right place to start, or at least the right initial
direction.
Once we have a properly composable primitive we can look at what people
need/use it for most (or how it trips them up most often!), and iterate
in that direction.
Regards,
--
Aristotle Pagaltzis // <http://plasmasturm.org/>
Thread Previous
|
Thread Next