This is the mail archive of the
libstdc++@gcc.gnu.org
mailing list for the libstdc++ project.
Re: Google Summer of Code
- From: Jonathan Wakely <cow at compsoc dot man dot ac dot uk>
- To: Jeyasankar Kottalam <gcc at jey dot kottalam dot net>
- Cc: chris jefferson <caj at cs dot york dot ac dot uk>, libstdc++ <libstdc++ at gcc dot gnu dot org>
- Date: Wed, 1 Jun 2005 12:05:19 +0100
- Subject: Re: Google Summer of Code
- References: <20050531223545.rxy6ta5zy00kso08@kottalam.net> <429D85C7.4090700@cs.york.ac.uk> <20050601035358.wcg1spf97rocggs4@kottalam.net>
On Wed, Jun 01, 2005 at 03:53:58AM -0700, Jeyasankar Kottalam wrote:
> >I don't know if anyone is already working on them, or how hard they
> >would be to write, or if they are going to be borrowed from somewhere
> >else, but if google would pay someone to produce a nice finished version
> >of either regular expressions or random number generators (the two big
> >bits of TR1 we still seem to be missing), and a whole bunch of tests to
> >go with them, that seems like it would only be a good thing :)
>
> Hm... interesting. Would the regex library involve hand-rolling a parser and
> state machine? Or interfacing with an existing regex library? I haven't
> looked
> at the actual specification in TR1.
It needs to work with arbitrary character types, in the same way as
basic_iostream and basic_string do. It also should support almost every
RE syntax (perl, ecmascript, extended, basic, posix etc.)
That requires a regex_traits type to describe whether a character is a
metacharacter in the current syntax and a parser that will build a state
machine based on those traits. I see it as a framework for defining RE
grammars, along with a handful of predefined grammars such as basic and
extended REs.
It's quite a big task, which I'd love to have time to tackle, but don't.
It might be possible to borrow the Boost regex code that inspired the TR1
interface.
jon
--
Always forgive your enemies; nothing annoys them so much.
- Oscar Wilde