bug#16893: [PATCH] Avoid matching line-by-line for case-insensitive with grep

[Glenn Morris]

This message was sent to the address@hidden address with no 
"Package:" specified in the body. So it ended up on the help-debbugs
mailing list rather than bug-grep. I have assigned it to grep and am
sending this mail, which will appear on the bug-grep list.

For new reports, either use the bug-grep address, or remember to use
Package: grep at the start of the body. They both have identical results.

Norihiro Tanaka wrote:

> Now grep and awk matchers doesn't waste buffer in case-sensisitive matching.
> So I think that we can avoid line-by-line matching for them.
>
> It enable to speed up case-sensitive matching with grep or awk matcher
> without trivial_case_ignore as fast as when with it.
>
> In bug#16232:
>> The following times 2.16, 2.17 and 2.17+patch two ways:
>> 
>> $ yes jjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjjj | head -10000000 > k
>> $ for i in 16 17 18; do echo $i; env LC_ALL=en_US.UTF-8 time
>> /p/p/grep-2.$i/bin/grep -i foobar k; done
>> 16
>>        15.96 real        14.57 user         0.12 sys
>> 17
>>         1.13 real         1.07 user         0.06 sys
>> 18
>>         1.96 real         1.89 user         0.06 sys
>> 
>> The above search takes more than 70% longer with the proposed patch.
>
> Therefore, I think 30% slow-down is caused by the line-by-line matching
> for them.

[See attachment at http://debbugs.gnu.org/16893]