All of lore.kernel.org
 help / color / mirror / Atom feed
From: "Shawn O. Pearce" <spearce@spearce.org>
To: Adrian Bunk <bunk@kernel.org>
Cc: git@vger.kernel.org
Subject: Re: git-revert is a memory hog
Date: Sun, 27 Jan 2008 12:38:08 -0500	[thread overview]
Message-ID: <20080127173808.GX24004@spearce.org> (raw)
In-Reply-To: <20080127172748.GD2558@does.not.exist>

Adrian Bunk <bunk@kernel.org> wrote:
> I'm not sure whether this is already known, but when recently working 
> for some time from a computer with "only" 512 MB RAM I ran into the huge 
> memory usage of git-revert when it tries to revert old commits.
> 
> Example (in Linus' kernel tree with git 1.5.3.8):
> 
> $ git-revert d19fbe8a7
...
> In top you can see that this took > 800 MB of RAM !
> 
> I don't know how easy it would be to implement, but shouldn't git-revert 
> be able to be as fast and less memory consuming as
>   git-show d19fbe8a7 | patch -p1 -R

Its more like:

	git-diff-tree -M d19fbe8a7^ d19fbe8a7 | git-apply -R --index

In other words its doing rename detection.  Its possible that the
rename detector fired for added/removed paths and it took some
significant amount of memory to figure out what was renamed.
Maybe we hung onto stuff for too long, but it has to do rename
detection and that takes memory to store the matrix and file
contents.  Once memory is allocated by git we don't give it back
to the kernel, even if we free'd it internally.

If you really are tight on memory and have to do a revert you can
use the above, but minus the -M, to setup the change, but you may
run into trouble if renames were involved.

-- 
Shawn.

  reply	other threads:[~2008-01-27 17:38 UTC|newest]

Thread overview: 24+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2008-01-27 17:27 git-revert is a memory hog Adrian Bunk
2008-01-27 17:38 ` Shawn O. Pearce [this message]
2008-01-28  6:01   ` Jeff King
2008-01-28  5:59 ` Jeff King
2008-01-29 21:51   ` Linus Torvalds
2008-01-29 22:15     ` Junio C Hamano
2008-01-29 22:20     ` Jeff King
2008-01-29 22:30       ` Jeff King
2008-01-29 22:36       ` Junio C Hamano
2008-01-29 22:45         ` Jeff King
2008-01-29 22:51           ` Jeff King
2008-01-29 22:49         ` Linus Torvalds
2008-01-29 22:54           ` Jeff King
2008-01-29 22:53         ` Junio C Hamano
2008-01-29 22:57           ` Jeff King
2008-01-29 23:19         ` Junio C Hamano
2008-01-29 23:50           ` Junio C Hamano
2008-01-30  4:40           ` [PATCH] Optimize rename detection for a huge diff Junio C Hamano
2008-01-30  6:57             ` Luke Lu
2008-01-30  7:24               ` Luke Lu
2008-02-13  9:53             ` Junio C Hamano
2008-02-13 10:19               ` David Kastrup
2008-02-13 10:23                 ` Junio C Hamano
2008-02-14  3:00               ` Junio C Hamano

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20080127173808.GX24004@spearce.org \
    --to=spearce@spearce.org \
    --cc=bunk@kernel.org \
    --cc=git@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.