Boost logo

Boost :

From: Beman Dawes (bdawes_at_[hidden])
Date: 2001-10-31 09:42:22


Various Boost libraries were proposed to the C++ standards committee's
Library Working Group (LWG) at the Redmond meeting for inclusion in the
upcoming Library Technical Report (TR). What follows is my recollection of
the reaction from the LWG.

There doesn't seem to be any question of regular expressions being useful,
and that they should be included in the TR in some form or another.

One question: How does the Boost regex proposal compare to other regex
libraries? Answer given: Whenever possible, compatibility has been
maintained with the Henry Spenser, Perl, GNU, and POSIX regex libraries,
and the proposal contains POSIX compatibility functions. That reply seemed
to reassure several LWG members.

Matt Austern mentioned a couple of concerns (which will also apply so some
other libraries):

* There may well be other regex proposals. How does the LWG decide
between them?

* There are other regex libraries available; does the LWG look at them
too?

[My personal reaction to the first question is to not worry about it.
During the initial development of the standard library, many people told
the LWG that they were about to propose something, often a "better" string
class. Almost invariably, the proposals never materialized. The second
question can be dealt with in proposals by mentioning other well known
existing libraries, if any, which cover the same ground, and perhaps
pointing out major differences. Don't get into a deep analysis,
however. Leave that to others.]

Based on reaction to a very brief presentation I made, it looks like there
is also LWG interest in Darin Adler continuing development of his string
algorithms because they enhance the usefulness of both strings and regular
expressions.

It looks to me that because regular expression are viewed as important, the
Boost Regular Expression library will receive a lot of scrutiny before
being accepted. But the initial reaction seemed favorable.

What might be helpful at this point is to identify some other regex
libraries (commercial as well as free), and describe (a couple of sentences
only) how they differ from Boost regex.

--Beman


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