When cloning with the branch option:
- Only the specified branch will be fetched
- The trunk alias is only set if the specified branch happens to be the default branch
- The clone fails if the branch does not exist in the remote
Since we can now update tags internally, local and remote tags may be out of
sync in non-colocated Git repositories. "tags" template should indicate that.
This is a stripped-down version of cmd_bookmark_set(). Since tags shouldn't
usually be rewritten, I've made it fail on updating existing tags by default.
This will help "jj bookmark track" know whether absent remote ref can be created
for the specified remote. "jj bookmark" subcommands shouldn't depend on
gix::Repository API.
Although we don't have "jj git tag set"/"delete" commands, this fixes weird undo
behavior #6325. The discussion in #6325 is derailed, but there would be another
issue for the "abandon unreachable" behavior.
Fixes#6325
Previously untracking a file with `jj file untrack` would not remove the
file from the git index (intent-to-add), making the output of
`jj status` and `git status` inconsistent.
`jj bisect run --command` splits the argument on spaces, so
e.g. `--command 'cargo test'` will run `cargo` with `test` as
argument. It is hard to pass more complex commands to because we don't
parse e.g. quotes as the shell would (and we wouldn't match the user's
preferred shell even if we did).
This patch deprecates the `--command` argument in favor of positional
arguments, so the user can do things like this:
```
jj bisect run --range ..main -- bash -c 'jj duplicate -r xyz -B @ && cargo test'
```
That will run the tests with the `xyz` change applied.
- Added support for workspace-specific configuration using `jj config` commands.
- Introduced `--workspace` option to `jj config edit`, `jj config set`, and `jj config unset`.
- Workspace configuration file is stored in `.jj/workspace-config.toml` per workspace.
- Simplified workspace path handling logic.
- Updated documentation and test snapshots to reflect new behavior.
This improves flexibility for users working across multiple workspaces.
Co-authored-by: Eidolon <furyhunter600@gmail.com>
Co-authored-by: ase <adam@sandbergericsson.se>
Remote names are potentially sensitive? Unsure, so better to redact them.
For getting the "*": This could have been done without modifying the
Rust side, with e.g. `stringify(bookmark).ends_with("*")`. However, this
feels brittle.
With adding `.synced()` to the template language, bookmark formatting
_could_ be done entirely in templates now. I don't know if that's
desired, so leaving for a potential future commmit.
I'm trying to go through the code and make sure that the working copy
is always updated to whatever the working-copy commit in the operation
says. The operation we pass to `LockedWorkingCopy::finish()`, that
is. That wasn't the case after importing the Git HEAD commit. There,
we passed the parent commit instead. It doesn't matter in practice
since the commit is empty but I think it is clearer this way.
When debugging why a workspace has a particular state, I often run `jj
op log -T id` to find an operation id. Then I pass that to `jj debug
object operation`. Then I get the view id from there and pass it to
`jj debug object view`. To simplify that process, this commit adds a
`jj debug object view --op` which accepts the usual operation
expression (not just a full operation id). Since I often use the
command is for debugging workspaces for other users and/or broken
workspaces, I made it not snapshot.
