Cocoa mail archive
Re: [ANN] RegexKit - An Objective-C Framework for Regular Expressions Using the PCRE Library
FROM : Alastair Houghton
DATE : Mon Sep 03 01:12:51 2007
On 1 Sep 2007, at 20:49, John Engelhart wrote:
> Announcing RegexKit - A framework for regular
> expressions using the PCRE library.
Hi John,
Any chance of changing the name to PCREKit or something more
specific? It's just that there are a number of regexp libraries, all
with subtly different implementations (e.g. POSIX, PCRE, Oniguruma,
ICU...). It can be a bit puzzling at times when confronted with apps
using the various different libraries (and/or with various different
options enabled), so I think it'd be good to make it *really* obvious
to people that your library is using PCRE.
Incidentally, does PCRE have good (i.e. native) support for UTF-16?
Oniguruma and ICU both do (and ICU includes a powerful implementation
of the regex character class feature that lets you query Unicode
attributes), which makes them a good choice for integration with
Cocoa, but if you have to e.g. transcode to UTF-8 in order to use
regex matching, it's going to be somewhat more expensive.
BTW, nice documentation. I was going to ask what tools you used to
do it, but it looks like you included them in the source
distribution. You should consider packaging up the doc. building
tools separately, as it looks like they're an improvement on headerdoc.
On 1 Sep 2007, at 22:20, John C. Randolph wrote:
> Sweet!
>
> Of course, I still want Apple to get this into the foundation kit,
> but this is quite welcome.
Indeed. It seems strange that they still haven't done this, given
that they've got ICU working under the covers... ICU includes a
perfectly reasonable (and Unicode-enabled) regular expression engine.
(BTW, it seems likely that an Apple-supplied implementation would be
based on ICU, whose regex engine uses an enhanced Perl-compatible
syntax.)
For those who are interested, there's also
<http://aarone.org/cocoaicu/>
though I don't know how it compares (it probably won't be as
efficient as an Apple implementation could be, as Apple have more
control over the internal representation of strings).
Kind regards,
Alastair.
--
http://alastairs-place.net
DATE : Mon Sep 03 01:12:51 2007
On 1 Sep 2007, at 20:49, John Engelhart wrote:
> Announcing RegexKit - A framework for regular
> expressions using the PCRE library.
Hi John,
Any chance of changing the name to PCREKit or something more
specific? It's just that there are a number of regexp libraries, all
with subtly different implementations (e.g. POSIX, PCRE, Oniguruma,
ICU...). It can be a bit puzzling at times when confronted with apps
using the various different libraries (and/or with various different
options enabled), so I think it'd be good to make it *really* obvious
to people that your library is using PCRE.
Incidentally, does PCRE have good (i.e. native) support for UTF-16?
Oniguruma and ICU both do (and ICU includes a powerful implementation
of the regex character class feature that lets you query Unicode
attributes), which makes them a good choice for integration with
Cocoa, but if you have to e.g. transcode to UTF-8 in order to use
regex matching, it's going to be somewhat more expensive.
BTW, nice documentation. I was going to ask what tools you used to
do it, but it looks like you included them in the source
distribution. You should consider packaging up the doc. building
tools separately, as it looks like they're an improvement on headerdoc.
On 1 Sep 2007, at 22:20, John C. Randolph wrote:
> Sweet!
>
> Of course, I still want Apple to get this into the foundation kit,
> but this is quite welcome.
Indeed. It seems strange that they still haven't done this, given
that they've got ICU working under the covers... ICU includes a
perfectly reasonable (and Unicode-enabled) regular expression engine.
(BTW, it seems likely that an Apple-supplied implementation would be
based on ICU, whose regex engine uses an enhanced Perl-compatible
syntax.)
For those who are interested, there's also
<http://aarone.org/cocoaicu/>
though I don't know how it compares (it probably won't be as
efficient as an Apple implementation could be, as Apple have more
control over the internal representation of strings).
Kind regards,
Alastair.
--
http://alastairs-place.net
