1 |
On 1/19/08, Robin H. Johnson <robbat2@g.o> wrote: |
2 |
> My core concern with the SVN http://, was the crappy performance it |
3 |
> provided compared to svn://. The main rsync tree has never been |
4 |
> available for iterative syncing via http://, just had tarball snapshots |
5 |
> and deltas instead. |
6 |
|
7 |
If I understand correctly, the performance of svn under apache is |
8 |
better than the svnserver, the same for git... Well... This is only |
9 |
for my experience. |
10 |
In git case, apache is used to transfer files, and it is much better |
11 |
in this than the most available alternatives. |
12 |
In svn case, apache provides the concurrency missing from svnserve. |
13 |
|
14 |
> > Also using none secured protocols, exposes users to man-in-the-middle attacks. |
15 |
> The existing http:// had this problem already, it's not a new one. |
16 |
> git:// and svn:// do both have patches around adding support for adding |
17 |
> TLS. This however just adds overhead, I really need to finish the |
18 |
> tree-signing work I was doing, as that protects the content better (MITM |
19 |
> is still possible on SSL without it, just a lot harder as an attacker |
20 |
> has to deal with the SSL stream first). |
21 |
|
22 |
Even if tree signing will be available, the developers should work in |
23 |
secured channel... ssh or https... The users will benefit from the |
24 |
signing and not require secured channel. |
25 |
|
26 |
Until signing will be available, I think it is very important for us |
27 |
to provide reliable source. |
28 |
|
29 |
Regards, |
30 |
Alon Bar-Lev. |
31 |
-- |
32 |
gentoo-dev@l.g.o mailing list |