| stellarpower | Hi, I have just set up a local 2a "repository" on my machine using the git-import command; this is to track a personal gitlab repository. I made some commits, and pushed this up to the master branch on gitlab okay. However, if I then run `brz missing` against the branch, it lists my local commits as extra to those on the server, and likewise | 13:34 |
|---|---|---|
| stellarpower | complains if I try to run `brz pull` (which I would expect to come back saying there is nothing new to pull as the branches should be in-sync. | 13:34 |
| stellarpower | IDK if that message sent, I ca't see it on my screen, I may try again with a different browser.. | 13:35 |
| stellarpower1 | Hi, I have just set up a 2a-format local "repository" using the git-import command - this is so that I can pull and push to/from a gitlab project. I checked out the master branch, added some files, committed, and then pushed those changes to gitlab upstream. All okay. | 13:44 |
| stellarpower1 | Then I come back this morning and find that if I run `brz missing git+ssh://git@gitlab.com/user/project,branch=master` in the master branch, breezy tells me that I have four extra commits in my local branch, and four missing from the remote branch (that it doesn't list). I do also get a pre-amble before saying `brz ERROR: not supported by the git | 13:44 |
| stellarpower1 | smart server`. Equally if I run `brz pull git+ssh://git@gitlab.com/user/project,branch=master` it tells me that the branches have diverged and I need to merge, when I would expect it simply to spit out that there is nothing new to pull. | 13:44 |
| stellarpower1 | I'm not very well-versed in version control and I know I have pushed and pulled to/from remote git repos using breezy before, but this is also a problem I encounter frequently at work, and have never understood what it is that I have been doing wrong, so I tend to reclone the branch and move my changes over, push that, and then delete. Or simply | 13:44 |
| stellarpower1 | only ever push, don't make any changes anywhere else other than my local branch on disc. Now that I've encountered this with an almost-empty repo it seems to be inherent in the workflow I was following. | 13:44 |
| stellarpower1 | Is there somehting I'm missing here, or is this perhaps an inherent limitation with the repo format or how I have set things up? I like the idea of keeping each branch in a separate directory (I think this is called non co-located branches) and the merge being between files so I can see what I'm doing with meld, but, I wouldn't be opposed to making | 13:44 |
| stellarpower1 | each branch use git storage underneath if this made thngs simpler to work with in this respect, as I need to be able to pull changes reliably from the remote repo, without ending up having to merge (in effect nothing) each time, as has happened in the past. | 13:44 |
| stellarpower1 | Can anyone give me some pointers please? Or is this a bug I should file on launchpad? Thanks!! | 13:44 |
| gegoxaren[m] | Do you have a common repo for your banches? You can try to brach the remote branch into a new branch on your system, then merge them locally. Or you could do an `--overwrite`. | 13:52 |
| gegoxaren[m] | 🤷 | 13:54 |
| gegoxaren[m] | `brz merge :parent --overwrite` | 13:55 |
| gegoxaren[m] | Or something. | 13:55 |
| stellarpower1 | I guess mainly I want to know why it's saying they've diverged when it's a one-way operation. Nothing else has touched the gitlab repo since pushing to it from that branch | 14:06 |
| stellarpower1 | So it presumably mustn't be identifying that the commits up there are in fact the same as the ones sitting locally | 14:07 |
| stellarpower1 | But yeah, it's that format I think. Rather than e.g. cloning with git, there's a central dir containing the .bzr hidden directory, and then each branch is in a separate folder with itsn ame under there, containing the working tree | 14:09 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!