| Package home | Report new bug | New search | Development Roadmap | Status: Open | Feedback | All | Closed Since Version 0.4.0 |
| Request #11917 | Faster regexp matching with \G instead of substr() | ||
|---|---|---|---|
| Submitted: | 2007-08-28 11:06 UTC | ||
| From: | joror | Assigned: | instance |
| Status: | Closed | Package: | PHP_LexerGenerator (version 0.3.4) |
| PHP Version: | 5.2.3 | OS: | |
| Roadmaps: | (Not assigned) | ||
[2007-08-28 11:06 UTC] joror
(Daan Broekhof)
Description:
------------
I noticed that the regular expressions used in the lexer were constructed in a manner of preg_match() on a substr() from a certain offset of the text.
A better performing solution is to use the 'offset' parameter of preg_match(), in combination with the \G offset place identifier in the regular expression.
The performance improvement is quite significant (not orders of magnitude), and it probably also reduces alot of memory swapping/usage due to the string not being chopped up at every match.
Comments
[2007-12-02 04:40 UTC] instance
(Alan Langford)
Changes integrated with other work; update is in CVS.