Replies: 5 comments 2 replies
|
i'd say that the short answer is "yes". if a submodule is using git-bug, that isn't really pertinent to the parent repository. it's a "third party dependency", and we shouldn't care about our dependency's issues. a worktree isn't a separate repository, it's just a separate "namespaced" working directory for a branch and is in fact the same repository as what you're calling the parent. |
There's more than one question in my description - which has the short answer "yes". If it's related to the submodule, then I think I agree (and We agree on what a worktree is ... but my intuition is that running |
|
apologies, i meant to quote this with the opening line of my comment:
|
i don't agree with this idea. a branch is often created to perform a unit of work (in a typical workflow, after an issue is created); issues (bugs) are part of the repository as a whole, and not part of a branch. if i'm in |
|
I need to do some tests on the current behavior ... from that description, I'm wondering if my understanding is wrong. |
|
My intuition was completely wrong ... sorry for the chatter. |
|
not to worry! discussion is always welcome :) |
Uh oh!
There was an error while loading. Please reload this page.
The following code works for Git submodules, where we presumably want
git-bugto manipulate bugs and users in the parent Git repository.git-bug/repository/gogit.go
Lines 179 to 202 in 7468b37
The code shown above also works for Git worktrees as the syntax of the
.gitfile also contains a path that will resolve to the parent Git repository with contents that look something like:Note that
mainabove is arbitrary and simply a folder that allows worktrees to be sibling directories to the actual repository.The difference between submodules and worktrees is that, for a worktree, there's an associated branch, One of
git-bug's features is that you can create a branch, add CRDT operations to existing (or new) bugs and users, then merge those operations back into the originating branch. This works fine for the parent repository as there's only one active branch at a time. But with the current code, if we manipulate bugs or users in a worktree, we alter the parent repository's operations now instead of when the associated branch is merged back into the parent repository.For work that's accepted (merged), this doesn't make much of a difference, but for work that's abandoned in an unmerged worktree, do we want the operations created for that branch immediately added to the parent or do we want those operations to also be abandoned?
Revisiting the submodule behavior above, if a submodule repository is managing bugs and users using
git-bugs, do we want to always be using the parent repository path?All reactions