[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <AANLkTinYk-xAk==1wtYy-XjVe-SGbZn-wiAK19Yj0jyr@mail.gmail.com>
Date: Wed, 19 Jan 2011 10:26:15 -0800
From: Linus Torvalds <torvalds@...ux-foundation.org>
To: Steven Rostedt <rostedt@...dmis.org>
Cc: nobody <darwinskernel@...il.com>,
Linux Kernel Mailing List <linux-kernel@...r.kernel.org>
Subject: Re: Linux 2.6.38-rc1
On Wed, Jan 19, 2011 at 10:07 AM, Steven Rostedt <rostedt@...dmis.org> wrote:
>
> When I first started using git (2006?) I was recommended to use "-k"
> when doing a fetch or pull. I don't even remember why it was
> recommended. I still do it, but is it necessary?
>
> All the help says is:
>
> -k, --keep
> Keep downloaded pack.
That's a different issue. That marks the resulting pack for "keeping",
which means that it won't participate in garbage collection.
IOW, it's a bad idea.
There _is_ a reason for it, namely that if your machine is really
slow, you may want to mark big packs for keeping, so that when you do
a gc it will be cheaper to do. In particular, the initial clone/fetch
may well be worth marking "keep". But even that is really only worth
it if you have such a slow machine (or the machine has too little
memory) that repacking is really a big pain (it's not like you need to
do gc very often - once a week would be overkill for most projects).
(There can be other reasons to mark packs as "keep" - namely to help
with http/rsync of the resulting archive if you have a huge amount of
history. In that case, it might be worthwhile keeping historical data
as a "keep" pack, so that rsync/http don't need to re-fetch it after a
GC or repack event.)
Linus
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists