Re: [PATCH] Replace strcmp_icase with strequal_icase

From: Duy Nguyen <pclouds@gmail.com>
Date: Mar 9, 2013, 12:40 UTC
Message-ID: <CACsJy8DfsgzEiMELy4UHJ2fvExZjcmvRop4m7u1H3OibrJiSPg@mail.gmail.com>
In-Reply-To: <20130309122200.GA7755@paksenarrion.iveqy.com>

On Sat, Mar 9, 2013 at 7:22 PM, Fredrik Gustafsson <iveqy@iveqy.com> wrote:

Show 11 quoted lines

> On Sat, Mar 09, 2013 at 07:05:37PM +0700, Duy Nguyen wrote:
>> On Sat, Mar 9, 2013 at 6:08 PM, Fredrik Gustafsson <iveqy@iveqy.com> wrote:
>> > Actually when implemented a str[n]equal_icase that actually should work.
>> > I break the test suite when trying to replace
>> > strncmp_icase(pathname, base, baselen)) on line 711 in dir.c and I don't
>> > get any significant improvements.
>>
>> Hmm.. mine passed the test suite.
>
> Using my patch or your own code? Maybe I just did something wrong. Could
> you see any improvements in speed?

It's the one I posted in [1] and yes it improves speed, numbers in [1].

[1] http://thread.gmane.org/gmane.comp.version-control.git/217712/focus=217724

Show 12 quoted lines

>> > I like work in this area though, slow commit's are my worst git problem.
>> > I often have to wait 10s. for a commit to be calculated.
>>
>> Personally I don't accept any often used git commands taking more than
>> 1 second (in hot cache case). What commands do you use? What's the
>> size of the repository in terms of tracked/untracked files?
>
> It's a small repository, 100 MB. However I have a slow hdd which is
> almost full. I often add one file and make an one-line change to an
> other file and then do a git commit -a. That will make git to look
> through the whole repo, which isn't in the kernel RAM cache but needs to
> be reed from the hdd.

"commit -a" does not run exclude (what I'm improving here). It's probably stat problem. If you already know what files you have changed, "git add path..." then commit without -a might help. Or turn on core.ignorestat (read doc about it first).

-- 
Duy