2007/11/3, Thomas Bächler :

James Rayner schrieb:

...

mlocate has replaced slocate as the default locate implementation, this is currently in [testing]. This was requested on the bug tracker a while back[1].

mlocate is used by Fedora at least and unlike slocate, no longer has the well known midnight updatedb crawl. In all other aspects it's compatible with and identical to slocate.

James

[1] http://bugs.archlinux.org/task/4490

As described by the README: mlocate is a locate/updatedb implementation. The 'm' stands for "merging": updatedb reuses the existing database to avoid rereading most of the file system, which makes updatedb faster and does not trash the system caches as much. The locate(1) utility is intended to be completely compatible to slocate. It also attempts to be compatible to GNU locate, when it does not conflict with slocate compatibility.

While I think this is a good thing, this list should be informed _before_ such a change is made and before the [arch] list is informed, and you should wait at least a day for any objections (don't wait for confirmation though, because then nothing will happen). Aaron was always pushing for more communication, so let's practice it.

Despite of all that, I like your initiative.

As is already visible on that bugreport - I'm fine with this change (didn't test the new package yet though). I hope Aaron, Travis, Dan and Eliott didn't change their thought too. ;-) -- Roman Kyrylych (Роман Кирилич)

Am Sat, 3 Nov 2007 17:48:51 +0100 schrieb Andreas Radke :

I'm not sure how mlocate follows the kernel file syscalls and if it will slow down any file operations. I'm also concerned how it works in chroots. I might try it later.

Ok, have to say sorry here. Mixed mlocate up with rlocate :-/ Here are my experiences: old slocate [root@workstation64 andyrtr]# time updatedb real 1m7.929s user 0m1.683s sys 0m2.566s [root@workstation64 andyrtr]# time updatedb real 1m19.958s user 0m1.690s sys 0m2.696s new mlocate: [root@workstation64 andyrtr]# time updatedb real 10m33.805s user 0m1.463s sys 0m33.304s [root@workstation64 andyrtr]# time updatedb real 0m21.013s user 0m0.617s sys 0m1.333s [root@workstation64 andyrtr]# time updatedb real 0m1.183s user 0m0.463s sys 0m0.417s [root@workstation64 andyrtr]# time updatedb real 0m0.868s user 0m0.413s sys 0m0.443s That'really faster after the first updatedb call! So I'm fine with it. Andy

On 11/3/07, Jason Chu wrote:

On Sat, Nov 03, 2007 at 09:26:25AM +0100, Thomas Bächler wrote:

...

James Rayner schrieb:

...

mlocate has replaced slocate as the default locate implementation, this is currently in [testing]. This was requested on the bug tracker a while back[1].

mlocate is used by Fedora at least and unlike slocate, no longer has the well known midnight updatedb crawl. In all other aspects it's compatible with and identical to slocate.

James

[1] http://bugs.archlinux.org/task/4490

As described by the README: mlocate is a locate/updatedb implementation. The 'm' stands for "merging": updatedb reuses the existing database to avoid rereading most of the file system, which makes updatedb faster and does not trash the system caches as much. The locate(1) utility is intended to be completely compatible to slocate. It also attempts to be compatible to GNU locate, when it does not conflict with slocate compatibility.

While I think this is a good thing, this list should be informed _before_ such a change is made and before the [arch] list is informed, and you should wait at least a day for any objections (don't wait for confirmation though, because then nothing will happen). Aaron was always pushing for more communication, so let's practice it.

Despite of all that, I like your initiative.

This was mentioned back on October 9th...

http://archlinux.org/pipermail/arch-dev-public/2007-October/002119.html

This is just James acting on the already existing confirmation.

Agreed. I poked the list a while back.

On Sun, Nov 04, 2007 at 07:06:59PM +0100, Tobias Powalowski wrote:

can we start signoff? for me it works on x86_64

works great on i686. Jürgen

On Mon, November 5, 2007 19:36, Roman Kyrylych wrote:

2007/11/4, Jürgen Hötzel :

...

On Sun, Nov 04, 2007 at 07:06:59PM +0100, Tobias Powalowski wrote:

...

can we start signoff? for me it works on x86_64

works great on i686.

Oh, I forget to say "signoff". :-) It works great here on i686.

Alright, done and dusted. Just got to remove slocate now.