blob: 6ba17969393b9f33d6ee7a690043ddd3a20ccf06 (
plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
|
[[!comment format=mdwn
username="http://adamspiers.myopenid.com/"
nickname="Adam"
subject="I think Matt is right."
date="2011-12-23T14:04:44Z"
content="""
I got bitten by this too. It seems that the user is expected to fetch
remote git-annex branches themselves, but this is not documented
anywhere.
The man page says of \"git annex merge\":
Automatically merges any changes from remotes into the git-annex
branch.
I am not a git newbie, but even so I had incorrectly assumed that git
annex merge would take care of pulling the git-annex branch from the
remote prior to merging, thereby ensuring all versions of the
git-annex branch would be merged, and that the location tracking data
would be synced across all peer repositories.
My master branches do not track any specific upstream branch, because
I am operating in a decentralized fashion. Therefore the error
message caused by `git pull $remote` succeeded in encouraging me to
instead use `git pull $remote master`, and this excludes the git-annex
branch from the fetch. Even worse, a git newbie might realise this
and be tempted to do `git pull $remote git-annex`.
Therefore I think it needs to be explicitly documented that
git fetch $remote
git merge $remote/master
is required when the local branch doesn't track an upstream branch.
Or maybe a `--fetch` option could be added to `git annex merge` to
perform the fetch from all remotes before running the merge(s).
"""]]
|