GNU bug report logs - #16232
[PATCH] grep: make --ignore-case (-i) faster (sometimes 10x) in multibyte locales

Previous Next

Full log

Message #29 received at 16232 <at> debbugs.gnu.org (full text, mbox):

From: Jim Meyering <jim <at> meyering.net>
To: Pádraig Brady <P <at> draigbrady.com>
Cc: 16232 <16232 <at> debbugs.gnu.org>
Subject: Re: bug#16232: [PATCH] grep: make --ignore-case (-i) faster
 (sometimes 10x) in multibyte locales
Date: Fri, 10 Jan 2014 20:52:00 -0800

[Message part 1 (text/plain, inline)]

On Fri, Jan 10, 2014 at 5:49 PM, Pádraig Brady <P <at> draigbrady.com> wrote:
> Cool so it does this transformation:
>
>   sed 's/./[\L&\U&]/g'
>
> Though multi byte case handling has all sorts of edge cases (pardon the pun),
> and it may not be always valid to treat each character independently?
> For example see some of the tests in:
> http://git.sv.gnu.org/gitweb/?p=gnulib.git;a=blob;f=tests/unicase/test-ulc-casecmp.c;hb=HEAD

It seems you're right.  Since it's a many-to-one mapping in some
cases, simply using one lower case character and one upper case
version won't cover all possibilities.

> I wonder might this faster path be restricted to a safer but very common input subset of:
>
> (MB_CUR_MAX == 1 || (in_utf8 && *c < 0x80))

That sounds like a good approach.
Now I need another test case, to demonstrate that the current code can
cause trouble.

> Also are the following printfs in the test redundant?
>
>> +data=$(      printf "I:$I $i:i")
>> +search_str=$(printf "$i:i I:$I")

Good catch.  Those were vestiges of pre-factoring code, where they
were needed.  Here's the patch to fix that part, in your name:

[k.txt (text/plain, attachment)]

Previous Next

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