Re: git and larger trees, not so fast?

From: Linus Torvalds <torvalds@linux-foundation.org>
Date: Aug 9, 2007, 21:41 UTC
Message-ID: <alpine.LFD.0.999.0708091426050.25146@woody.linux-foundation.org>
In-Reply-To: <20070809165218.9b76ebf7.seanlkml@sympatico.ca>

On Thu, 9 Aug 2007, Sean wrote:

Show 13 quoted lines

> 
> This makes things _much_ better, however the final commit in the 
> test script still shows a lot of user time:
> 
> ## time git commit -q -m 'buurrrrn' -a
> real    0m2.299s
> user    0m1.065s
> sys     0m1.317s
>
> ## time git commit -q -m 'hurry' 50/500
> real    0m16.944s
> user    0m15.466s
> sys     0m1.133s

In the case where we do a partial commit (never mind that it's all the changes: when you give a path limiter, it's "partial"), "git commit" one goes through another path, and triggers the same issue with

	git read-tree --index-output=tmp-index -i -m HEAD

which has the same O(n**2) issue (except it uses "unpack_trees()" rather than "read_tree()", so Junio's patch does nothing for it).

So "builtin-read-tree.c" (or rather unpack-trees.c) would need the same kind of logic.

		Linus