1 |
On Wed, Nov 26, 2014 at 9:17 PM, Paige Thompson <erratic@××××××××××.sx> wrote: |
2 |
> A good example of why I don't think they will be using git for portage: |
3 |
> ``git clone https://git.kernel.org' |
4 |
> |
5 |
> This takes a very long time and a lot of bandwidth and a lot of space. |
6 |
> Why move data that isn't going to be used? With rsync I believe you can |
7 |
> exclude categories: |
8 |
> http://www.gentoo-wiki.info/TIP_Exclude_categories_from_emerge_sync |
9 |
|
10 |
Anybody who wants to use rsync can still do so after the migration. |
11 |
|
12 |
Git is a LOT faster at syncing than cvs, which is what we're all stuck |
13 |
with right now. |
14 |
|
15 |
The whole gentoo repository with history takes about 2GB, but we will |
16 |
be separating history from the current tree so initially the |
17 |
repository will be quite small. |
18 |
|
19 |
For anybody who does want the history git is very efficient at |
20 |
syncing. You would need to store the whole repository, but the |
21 |
bandwidth needed for syncs is very low. |
22 |
|
23 |
> |
24 |
> If they were gonna do that with git every category would have to be a |
25 |
> submodule. I could be wrong but I just want to believe it hasn't been |
26 |
> made available despite being supported for a good reason. |
27 |
|
28 |
There are a bunch of reasons that it hasn't happened. You can read |
29 |
the tracker bug to get a sense of the past roadblocks. At this point |
30 |
the biggest item left is getting the new infra set up, which is |
31 |
underway. |
32 |
|
33 |
-- |
34 |
Rich |