| 1 |
On 27/03/06, Ryan Phillips <rphillips@g.o> wrote: |
| 2 |
> Aron Griffis <agriffis@g.o> said: |
| 3 |
> > Have you followed the threads in the past regarding using other |
| 4 |
> > version control systems for portage? Some devs have done benchmarks |
| 5 |
> > and found that there are blocking issues with subversion, particularly |
| 6 |
> > because of its repo-wide revisions that prevent multiple commits from |
| 7 |
> > happening simultaneously. |
| 8 |
> |
| 9 |
> In actuality, Subversion does 98% of the commit in an initial |
| 10 |
> transaction, and the blocking only occurs in the last 2% with the FSFS |
| 11 |
> filesystem. It really isn't an issue and shouldn't prevent us from |
| 12 |
> adopting it. |
| 13 |
|
| 14 |
All svn commits are atomic, and that requires some kind of global |
| 15 |
lock. I'd say the (slight) performance penalty is worth it for that |
| 16 |
feature alone. I'd also point out that the KDE project have everything |
| 17 |
in a single svn repository and can manage >10,000 commits per month |
| 18 |
with no problems. There are various testimonials around from people |
| 19 |
claiming to be running svn on multiple GB repositories with >17,000 |
| 20 |
commits a month. |
| 21 |
|
| 22 |
-- |
| 23 |
gentoo-dev@g.o mailing list |
Archives