1 |
On Mon, 2006-03-27 at 09:51 +0100, Chris Bainbridge wrote: |
2 |
> On 27/03/06, Ryan Phillips <rphillips@g.o> wrote: |
3 |
> > Aron Griffis <agriffis@g.o> said: |
4 |
> > > Have you followed the threads in the past regarding using other |
5 |
> > > version control systems for portage? Some devs have done benchmarks |
6 |
> > > and found that there are blocking issues with subversion, particularly |
7 |
> > > because of its repo-wide revisions that prevent multiple commits from |
8 |
> > > happening simultaneously. |
9 |
> > |
10 |
> > In actuality, Subversion does 98% of the commit in an initial |
11 |
> > transaction, and the blocking only occurs in the last 2% with the FSFS |
12 |
> > filesystem. It really isn't an issue and shouldn't prevent us from |
13 |
> > adopting it. |
14 |
> |
15 |
> All svn commits are atomic, and that requires some kind of global |
16 |
> lock. I'd say the (slight) performance penalty is worth it for that |
17 |
> feature alone. I'd also point out that the KDE project have everything |
18 |
> in a single svn repository and can manage >10,000 commits per month |
19 |
> with no problems. There are various testimonials around from people |
20 |
> claiming to be running svn on multiple GB repositories with >17,000 |
21 |
> commits a month. |
22 |
|
23 |
Well, CIA seems to differ from your stats. |
24 |
|
25 |
Gentoo: 8495 last month http://cia.navi.cx/stats/project/gentoo |
26 |
KDE: 7523 last month http://cia.navi.cx/stats/project/KDE |
27 |
|
28 |
Now, according to the stats, I would say that we have a similar commit |
29 |
rate as KDE, though not all of ours are on the same repository and some |
30 |
are on CVS and some are on SVN, so the stats aren't 100% accurate for |
31 |
getting just the main portage tree. |
32 |
|
33 |
-- |
34 |
Chris Gianelloni |
35 |
Release Engineering - Strategic Lead |
36 |
x86 Architecture Team |
37 |
Games - Developer |
38 |
Gentoo Linux |