[Webkit-unassigned] [Bug 38117] Differences between subpattern matching in use of pcre and Yarr Intrepreter

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=38117





--- Comment #2 from Peter Varga <pvarga at inf.u-szeged.hu>  2010-05-20 02:22:37 PST ---
(In reply to comment #1)
Thanks for your detailed answer.

I'm working on the implementation of the cases which are missing from the YARR
JIT and cause fallback to pcre. I'm planning to apply some optimizations on
YARR Interpreter too, which I noticed during the work on JIT.
Most of my previous bug reports (like this) were raised during this work.
Now I'm working on the JIT implementation of the quantified (iterative) group matching.

The current conception is to avoid to allocate stack frames for terms in
advance and store information about terms in every iteration on stack. This
information is usually the match count of each term and the start and
end position of the previous matching in case of capturing groups. A part of this
solution can be found at this bug report:
https://bugs.webkit.org/show_bug.cgi?id=38988
This patch is motivated by the "do I really want this on the stack" TODO.

Thanks,
Peter

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.