| 1 |
On Monday 19 June 2006 09:58, Trenton Adams wrote: |
| 2 |
> Hi guys, |
| 3 |
> |
| 4 |
> I'm just curious about something. I've noticed many people report |
| 5 |
> problems with their repository syncing simply because someone was in |
| 6 |
> the middle of committing to the repository. Couldn't this be resolved |
| 7 |
> by replacing the syncing mechanism with *svn* as opposed to rsync? |
| 8 |
> After all, it does have atomic transactions, and you cannot get a |
| 9 |
> partially updated repository when you use it. |
| 10 |
|
| 11 |
Perhaps it could. But I don't think that problem is important enough to make |
| 12 |
Portage usage depend on Subversion. |
| 13 |
|
| 14 |
[...] |
| 15 |
|
| 16 |
> One other benefit of this mechanism, that I can think of, would be |
| 17 |
> that I could easily roll back to my last sync date or release, if some |
| 18 |
> of the updated ebuilds caused me problems. |
| 19 |
|
| 20 |
That really should not be necessary. What you should of course do is file a |
| 21 |
bug so the problems can be fixed for everyone. |
| 22 |
|
| 23 |
> Any thoughts? |
| 24 |
|
| 25 |
Currently the developers are still using CVS for the tree. Migrating to |
| 26 |
another VCS does have a high priority but they have not even chosen which VCS |
| 27 |
to migrate to. There is a Google Summer of Code project which is supposed to |
| 28 |
test the different options to give a good basis for making a decision. This, |
| 29 |
however, affects the Gentoo developers only. |
| 30 |
|
| 31 |
-- |
| 32 |
Bo Andresen |
Archives