This fixes real dependency cycle that would expose slightly different versions
of jj_lib types to unit tests, one from super::* and another from testutils::*.
The testutils dependency can be removed by splitting lib/tests to a separate
crate.
These helpers are going to be needed to port the git2 code in the lib
tests to gitoxide. Since the cli tests already depend on testutils, this
helps with avoiding duplicating the code
Although this behaviour is accepted by git, it's a degenerate case.
Especially because we implicitely rely on being able to parse out the
remote from the refname (i.e., `refs/remotes/<remote>/<branch>`).
Branches can have forward slashes, but if remotes can also have them,
parsing the refname becomes ambiguous, and a pain. Especially because it
would be totally legal to have a branch "c" on remote "a/b" and a branch
"b" on remote "a".
Fixes#5731
When updating the working copy results in no changes, we currently
return `None` from `cli_util::update_working_copy()`. However, a no-op
update is still an update, so I think it makes more sense to always
return a `CheckoutStats`. We already skip printing stats if they're
zero.
This was discussed in the Discord a while ago, and this is the logical and consistent
conclusion. Implementing it as such makes it consistent with both `jj edit` and `jj new`
which make hidden commits such as predecessors visible.
This actually was Martins work, I just added the tests.
Co-Authored-by: martinvonz <martinvonz@google.com>
`gix::Repository` instances are larger than 1KB on 64-bit machines. This
makes Clippy warn about an unbalance between variants of `GitFetchImpl`:
the `Git2` variant requires 8 bytes, while the `Subprocess` variant
requires 1128 bytes.
Boxing the `gix::Repository` makes `GitFetchImpl` instances smaller, as
well as the other structs embedding them such as `GitFetch`.
Since git CLI appends " " to sideband messages to overwrite the previous
line, we need to undo that. Our GitSidebandProgressMessageWriter expects
original sideband messages.
Given the previously‐stated intention of making this default
for the 0.27 release, prepare for that decision ahead of time by
enabling subprocessing by default on trunk. This will help surface
any regressions and workflow incompatibilities and therefore give
us more information to decide whether to keep or revert this commit,
without inconveniencing any users who haven’t already opted in to
the bleeding edge.
Please feel free to revert without hesitation if any major issues
arise; this is not intended as a strong commitment to enable this
option for the next stable release if it turns out to not be ready. In
that case, it’s better that we learn that early on in the cycle,
rather than having to revert at the last minute or, worse, cutting
a stable release that we later find contains a serious regression.
See the discussion in #5623. This prevents future accidents if member variables
are added to RewriteRootCommit which need to be propagated to EditCommitError
during conversion.
The previous name violated the "snake case" convention for variable names. I
suspect the intention was to use `RewriteRootCommit` as a type declaration, not
the argument name. Since the argument isn't used, we can use an underscore as
the name and leave the type declaration.
#cleanup
there are many ways to implement `heads` for your custom backend:
one involves calling `self.evaluate_revset` with a proper revset.
However, it can return an error, but `heads` does not allow it.
In our implementation of the index backend we do exactly the above
and we ended up with several unwraps which we are trying to avoid
as it could potentially crash our prod jobs.
P.S. The same logic applies to many methods in this trait, but I am doing one at a time.
Since we need to scan directory entries recursively in order to detect new
untracked paths, it doesn't make sense to reuse the .gitignore code path.
This change means all untracked file paths are listed in "jj status" even if
the whole directory is untracked. It might be a bit verbose, but should be
okay. Directories like node_modules should be excluded by .gitignore, not by
auto-track pattern.
Fixes#5389
With git.subprocess = true, it's more important to skip unneeded remote
operations. Each remote command may involve user intervention if authentication
requires manual step.
This change also means that the remote connection is no longer reused in git2
impl. I think the added cost is acceptable. The git2 impl will hopefully be
removed soon, and the remote branch name is needed only when cloning new repo.
The current comment uses `[noeol]`, which can be difficult to
understand. In this commit, the brackets are changed to parentheses to
make it clear that there is no semantic meaning to the comment, and the
wording is changed to be more clear to the user.
Git doesn't show this information in their "diff3" style either, and it
looks a bit strange having two different bracketed sections of text next
to each other, so I think it would be better to remove it.
