1 |
Paul de Vrieze wrote: |
2 |
> On Sunday 25 January 2004 04:02, Robin H. Johnson wrote: |
3 |
> |
4 |
>>A normal CVS checkout of the gentoo-x86 from the tree (locally on the |
5 |
>>CVS server) takes 8 minutes. Your SVN co took 3.8 times as long as the |
6 |
>>CVS checkout so thats nearly 30 minutes to do a checkout :-(. |
7 |
> |
8 |
> |
9 |
> Both take too long to my sense. Which is basically what I experienced before. |
10 |
|
11 |
Well thats 30 minutes for the whole tree initially, as i understand thats |
12 |
because svn is doing alot more in the background than CVS is, anyway i might |
13 |
just have not known how to operate it correctly so speeding it up might be a |
14 |
possibility. |
15 |
|
16 |
However note that thats 30 minutes for the whole tree, just grabbing the part |
17 |
you're going to be working on or even updating your working copy will never |
18 |
take that long once you have cheked it out initially. |
19 |
|
20 |
$ find /usr/portage -maxdepth 1 -iname '*-*'|sort|wc -l |
21 |
104 |
22 |
|
23 |
So grabbing a whole subbranch is still just 16 seconds on average (of course |
24 |
they differ in size) |
25 |
|
26 |
It could be better but its not that bad, especially considering all the |
27 |
atvantages you get. |
28 |
|
29 |
>>I am really worried about lock contention however. We already keep all |
30 |
>>of the cvs locks on a tmpfs filesystem for speed and disk fragmentation |
31 |
>>reasons. |
32 |
> |
33 |
> |
34 |
> I don't know how subversion does this, but you are right that it is possible |
35 |
> that subversion does global locking. In that case subversion is out of the |
36 |
> question. |
37 |
> |
38 |
> Paul |
39 |
> |
40 |
|
41 |
-- |
42 |
gentoo-dev@g.o mailing list |