Boost logo

Boost :

From: Hartmut Kaiser (hartmutkaiser_at_[hidden])
Date: 2003-11-17 01:15:49


Eric Niebler wrote:

> I have begun work on a new library. It is beginning life as a
> regular expression pattern matcher, though I plan to morph it
> into something more powerful and general. I am hoping spark
> the interest of some developers, Spirit developers in
> particular, to make sure that the new library plays well with
> the rest of Boost. The working title for the project is "xpressive".
>
> xpressive will have the following features:

[snip]

> The two rules above are equivalent; they both match "hello "
> repeatedly, storing the result in capture number 1, and then
> match "world".

Hmmm... Looks interesting. This would be usable as a separate lexing
component, well integrated into Spirit. How this does relate to the Spirit
regex parser? Do you see any possibilities to combine these two?

> xpressive will differ from Spirit in that it will have
> backtracking semantics. (Spirit has greedy semantics, and
> does not do backtracking.)

But Spirit does backtracking!

> The project is still quite small. I have checked the source
> code into the boost sandbox at boost-sandbox/boost/xpressive
> for anybody who is interested.

I'll have a look at it later this week.

> Is anybody else interested?

Please add me to your list of interested people, who wish to collaborate.
I'm especially interested in a lexing component for Wave, which does not
rely on Re2C, but is equally fast, but well integrated into the C++/STL
idioms.

Regards Hartmut


Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk