Now that we depend on the `git` executable being available for `jj git
fetch/push` tests, we might as well use it for `jj util gc` tests
too.
I also switched to using the Git backend in
`cli/tests/test_file_track_untrack_commands.rs`, which seemed to be
using the local backend for no good reason.
This will be used to compute the location to place commits from the
`--destination`, `--insert-after`, and `--insert-before` options. This
is used across the `new`, `duplicate`, and `rebase` commands, and can be
used for further commands as well.
The previous implementation of `new --insert-after` excluded children
that were ancestors of the new commit. This was done to avoid creating
cycles in the commit graph. However, this does not align with the
behavior of the `--insert-after` in the other commands, which explicitly
error if inserting a new commit after the given `--insert-after` commits
would create a cycle.
For consistency, and given that the desired behavior can be achieved
using by using `--insert-after` along with `--insert-before`, this
behavior is removed.
After system crash, file contents are often truncated or filled with zeros. If
a file was truncated to empty, it can be decoded successfully and we'll get
cryptic "is a directory" error because of an empty view ID. We should instead
report data corruption with the ID of the corrupted file.
#4423
The consensus about this change, if one ever existed, seems to have dissolved
into two cohorts, one which is happy with the change, and one which thinks we
should move both the bookmark and change id to the child commit instead of
leaving them on the parent.
We may decide to add flags to allow users to choose between the two behaviors,
but there are also other concerns such as where @ should go (parent or child).
Until we agree on a path forward it seems reasonable to delay the breaking
change by disabling it via the config option we added. I don't think it's
necessary to fully revert the code and new tests since we aren't announcing the
option.
#3419
This has bothered me for a while... It's useful to describe the target of the
split command as the "target commit", and it's a little confusing to have a
variable named "commit" in the code when we are dealing with several different
commits in the command.
I think this improves readability by being more precise.
When signing commits with `jj sign`, one might want to use a workflow
like:
```bash
jj fix && jj sign .. && jj git push
```
Making the default value for `-r`/`--revisions` configurable, will allow
such a workflow.
Co-Authored-By: Yuya Nishihara <yuya@tcha.org>
The output of `jj unsign` is based on that of `jj abandon`.
We output warnings when unsigning commits, which are not authored by the
user. This is encouraging to use `jj undo`, in case one unintentionally
drops signatures of others.
---
Co-authored-by: julienvincent <m@julienvincent.io>
Co-authored-by: necauqua <him@necauq.ua>
We always sign commits. This means commits, which are already signed,
will be resigned. While this is cumbersome for people using hardware
devices for signatures, we cannot reliably check if a commit is already
signed at the moment (see https://github.com/jj-vcs/jj/issues/5786).
We output warnings when signing commits, which are not authored by the
user. This is encouraging to use `jj undo`, in case one unintentionally
signs commits of others.
The output of `jj sign` is based on that of `jj abandon`.
---
Co-authored-by: julienvincent <m@julienvincent.io>
Co-authored-by: necauqua <him@necauq.ua>
This means that watchman is partially disabled until the user cleans up
untracked files. This isn't nice, but should be better than hiding untracked
files forever.
Fixes#5728Fixes#5602
There have been a number of users confused about why
their commits are immutable, or what to do about it, ex.
[https://github.com/jj-vcs/jj/discussions/5659].
Separately, I feel that the cli is too quick to suggest
`--ignore-immutable`, without context of the consequences. A new user
could see that the command is failing, see a helpful hint to make it not
fail, apply it and move on. This has wildly different consequences, from
`jj squash --into someone_elses_branch@origin` rewriting a single commit,
to `jj edit 'root()+'` rewriting your entire history.
This commit changes the immutable hint by doing the following:
* Adds a short description of what immutable commits are used for, and a
link to the relevant docs, to the hint message.
* Shows the number of immutable commits that would be rewritten if
the operation had succeeded.
* Removes the suggestion to use `--ignore-immutable`.
Previously, tools invoked by `jj fix` did not have their
`.current_dir()` set, and would just run from whatever directory the
user was in.
Now, the tools will always be invoked from the same directory that
`jj root` gives.
As a motivating example, consider a configuration path that's always at
the workspace root:
```toml
[fix.tools.something]
# previous:
command = ["cmd", "--config", "$root/tool.toml"]
# ^^^^^^ not possible
# now:
command = ["cmd", "--config", "./tool.toml"]
# ^^ now, just use a relative path
```
