[Date Prev][Date Next][Thread Prev][Thread Next]
- Subject: Re: [ANN] matchext 0.1.0 - fork and extension of Lua 5.3 pattern matching
- From: Jonathan Goble <jcgoble3@...>
- Date: Sun, 24 Jan 2016 15:26:06 -0500
On Sun, Jan 24, 2016 at 9:17 AM, Jay Carlson <firstname.lastname@example.org> wrote:
> Thank you for following up on this. It builds cleanly on OS X 10.11 Xcode
> 7.2, although I haven't really tested it. (Add a test suite to the wish list.
Yes, I plan to add a test suite before I go too far with this. Probably, it
will be set up for automatic testing via Travis CI (and possibly Appveyor for
Windows testing as well).
> IMO, the most important feature for a Unicode pattern matcher is for "." and
> inverted ranges to match either whole codepoints or nothing. I think that
> would mean: given valid utf8 patterns and haystacks, no invalid utf8 can be
> output. "First, do no harm..."
The main obstacles to UTF-8 pattern matching are for ".", sets, and quantifiers
to match on codepoints instead of bytes. Everything else works "as-is". I have
no intention of making classes like "%a" match on the Unicode meaning; that
would be far too complicated and bloated.
> I think the slnunicode package has a Unicode-aware pattern matcher, if you
> want other ideas.
I'll look into that. Thanks for the tip.