[pcre-dev] [Bug 2698] Performance regression caused by r1236

Top Page
Delete this message
Author: admin
Date:  
To: pcre-dev
Subject: [pcre-dev] [Bug 2698] Performance regression caused by r1236
https://bugs.exim.org/show_bug.cgi?id=2698

Zoltan Herczeg <hzmester@???> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |hzmester@???


--- Comment #2 from Zoltan Herczeg <hzmester@???> ---
I cut back EARLY_FAIL_ENHANCE_MAX too much.

If you are interested about this optimization, you can read section "Single
character repetition optimizations" in

https://zherczeg.github.io/sljit/pcre2_jit.html

There is no perfect value for this constant, the "enhanced" iterators can slow
down ceartain patterns, but speed up others. But dropping it from 3 to 1 was
too much, I probably wanted to drop it to 2 (that means "enhance" the first two
iterators).

Could you try to change this constant to the following and see what happens:

#define EARLY_FAIL_ENHANCE_MAX (1 + 2)

--
You are receiving this mail because:
You are on the CC list for the bug.