From: Brian Hurt <brian.hurt@ql...> - 2003-03-20 17:17:21
Opps. Too late :-)
---------- Forwarded message ----------
Date: 21 Mar 2003 01:50:36 +0900
From: Yutaka OIWA <oiwa@...>
Subject: [Caml-list] Announce: Regexp/OCaml syntax extension
I release a camlp4-macro package called Regexp/OCaml.
I believe Objective Caml is a powerful tool not only for writing an
interpreter and compiler but also for writing casual "script"
applications which were in a territory of Perl, Python and
Ruby. However, string decomposition is one of OCaml's weak points when
compared to those scripting languages: the interfaces of both Str
module and Pcre module are very primitive and cumbersome if they are
heavily used. Regexp/OCaml solves this point by providing syntax
support for regular expression matching.
Regexp/OCaml provides convenient syntax sugar for regular expression
match against strings using PCRE/OCaml library. The features of this
macro package are the following:
* Convenient syntax: similar to standard match-with expressions
* Binding matching substrings to variables: no more $1, $2, ...
* Automagical easy-to-use type-coercion: no flood of int_of_string etc.
* Support for optional-patterns: gives string option type etc.
* Default values for optional-patterns
For example, parsing an entry for some log file becomes as easy as follows:
while true do
let line = input_line ic in
Regexp.match line with
as hour : int, min: int, sec : int, name, line ->
let time = hour * 3600 + min * 60 + sec in
| "^# (.*)$" as meta_info ->
| _ -> ()
with End_of_file -> ()
This short code parses both line in format like
"(00:34:32) [foobar] something" and "# some meta info"
and binds appropriate data into variables which can be used inside "...".
Compare the code above with an equivalent without using syntax extension.
Regexp/OCaml is downloadable from the web location
In addition to the main macro called pa_regexp_match, the package also
contains two tiny macros:
1) pa_pragma changes an option for loaded camlp4 macros inside source code.
2) pa_once provides "once" construct which evaluates any expression
only once per execution (pa_regexp_match uses this internally).
Any comments will be greatly appreciated.
Yutaka Oiwa Yonezawa Lab., Dept. of Computer Science,
Graduate School of Information Sci. & Tech., Univ. of Tokyo.
PGP fingerprint = C9 8D 5C B8 86 ED D8 07 EA 59 34 D8 F4 65 53 61
To unsubscribe, mail caml-list-request@... Archives: http://caml.inria.fr
Bug reports: http://caml.inria.fr/bin/caml-bugs FAQ: http://caml.inria.fr/FAQ/
Beginner's list: http://groups.yahoo.com/group/ocaml_beginners