GNU bug report logs - #15191
faster DFA.C state merge

Previous Next

Package: grep;

Reported by: Ivan Yanikov <dobrokot <at> gmail.com>

Date: Mon, 26 Aug 2013 06:58:03 UTC

Severity: wishlist

Tags: moreinfo, patch

Done: Paul Eggert <eggert <at> cs.ucla.edu>

Bug is archived. No further changes may be made.

Full log


View this message in rfc822 format

From: help-debbugs <at> gnu.org (GNU bug Tracking System)
To: Ivan Yanikov <dobrokot <at> gmail.com>
Subject: bug#15191: closed (Re: bug#15191: faster DFA.C state merge)
Date: Sat, 17 May 2014 03:26:02 +0000
[Message part 1 (text/plain, inline)]
Your bug report

#15191: faster DFA.C state merge

which was filed against the grep package, has been closed.

The explanation is attached below, along with your original report.
If you require more details, please reply to 15191 <at> debbugs.gnu.org.

-- 
15191: http://debbugs.gnu.org/cgi/bugreport.cgi?bug=15191
GNU Bug Tracking System
Contact help-debbugs <at> gnu.org with problems
[Message part 2 (message/rfc822, inline)]
From: Paul Eggert <eggert <at> cs.ucla.edu>
To: Norihiro Tanaka <noritnk <at> kcn.ne.jp>, 15191-done <at> debbugs.gnu.org
Cc: Ivan Yanikov <dobrokot <at> gmail.com>
Subject: Re: bug#15191: faster DFA.C state merge
Date: Fri, 16 May 2014 20:24:54 -0700
Norihiro Tanaka wrote:
> If my rewriting of the patch is right, it causes slowdown rather by
> building heap.

Thanks for checking it.  I'll close this bug report, since it seems to 
have been fixed in a different way.

[Message part 3 (message/rfc822, inline)]
From: Ivan Yanikov <dobrokot <at> gmail.com>
To: Paul Eggert <eggert <at> cs.ucla.edu>
Cc: bug-grep <at> gnu.org
Subject: Re: faster DFA.C state merge
Date: Mon, 26 Aug 2013 05:26:11 +0400
[Message part 4 (text/plain, inline)]
I have polished the patch (
    faster code for simple cases of merging 2 or 1 sets,
    fixes of code style, like fancy 2+2 indentation,
    important fix: memory leaks under valgrind - I improperly returned array from a function )

New patch available on URL https://gist.github.com/dobrokot/6337339 and included as attach

> One thing that'd help would be a test case that illustrates the need for the patch.

Not sure how to properly to send test-case, and how to reference grep compiled old/new binary. I had put it on URL http://dobrokot.ru/dump/slow_dfa_merge.2013-08-26.tar.gz
I can send it in a attach for a "history", if binary/large attaches are allowed in maillists.
Real regex contains sensitive private data, and it's huge. So I had little obfuscated it and reduce to kilobytes. The runtime difference is not so great as in real example (1 to 100), but still large 
( 2-3 times faster ).

Times for new/old version: 2.3sec / 8.7sec

> I assume you're willing to assign copyright to the FSF for the change?  (I can send you copies of the paperwork, if so.)

You mean, I should somewhere explicitly state, that I am agree with GPL and give FSF rights to distribute and use the code from patch?
If so, I am surely agree with the license, and (proudly) give the permission to use/distribute code from my patch :)

Or you mean some modification of authorship lines in the AUTHORS/THANKS and beginning of dfa.c, which should contain my name now ?

Which paperwork do you mean? Real paper which I should sign with pen and ink?

Thanks for attention, Ivan Yanikov.

On 25.08.2013 11:02, Paul Eggert wrote:
> Ivan Yanikov wrote:
>> How I can properly send/commit it for review?
> You've already done that, sort of, with the pointers
> to the patch.  One thing that'd help would be a test
> case that illustrates the need for the patch.  Also,
> I assume you're willing to assign copyright to the
> FSF for the change?  (I can send you copies of the
> paperwork, if so.)
>

[grep-2.14-faster-dfa-merger.diff (text/plain, attachment)]

This bug report was last modified 11 years and 12 days ago.

Previous Next


GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.