[pcre-dev] [Bug 1368] Easier to use function for multi-segme…

Top Page
Delete this message
Author: Philip Hazel
Date:  
To: pcre-dev
Subject: [pcre-dev] [Bug 1368] Easier to use function for multi-segment matching
------- You are receiving this mail because: -------
You are on the CC list for the bug.

http://bugs.exim.org/show_bug.cgi?id=1368




--- Comment #1 from Philip Hazel <ph10@???> 2013-07-14 19:08:46 ---
There is currently a discussion starting about a new API for PCRE. I hope to
make the current proposal available for general discussion fairly soon, when
the maintainers have thrashed it about enough. One of the features is a cleaner
way of passing back additional information after a match or a partial match, in
particular the first and last characters inspected and the start of match
position. That should make it easier to understand how to implement
multi-segment matching using pcre_exec(). However, implementing a "resume where
stopped" feature for pcre_exec() (and indeed also for the JIT) is so
complicated that I am sure it will not be done. A re-run of the match, having
extended the data, is the only possible approach.

For pcre_dfa_exec(), using the RESTART flag, it already does exactly resume the
algorithm, except that if the match fails, it cannot re-start at the next
character because the previous data is not available.


--
Configure bugmail: http://bugs.exim.org/userprefs.cgi?tab=email