Implements worktree support for the agent panel sidebar
Before you mark this PR as ready for review, make sure that you have:
- [x] Added a solid test coverage and/or screenshots from doing manual
testing
- [x] Done a self-review taking into account security and performance
aspects
- [x] Aligned any UI changes with the [UI
checklist](https://github.com/zed-industries/zed/blob/main/CONTRIBUTING.md#uiux-checklist)
Release Notes:
- N/A
Closes#46661
This PR changes `fs.rename` to use the platform’s atomic no-overwrite
rename on all platforms when `overwrite` is `false`. This fixes a case
where concurrent renames to the same target could race past a separate
metadata check and end up overwriting each other.
In Project Panel, we can still rename entries in parallel without
worrying about OS internals not handling it correctly or making these
renames sequential.
Release Notes:
- Fixed an issue in the Project Panel where conflicting file moves could
overwrite each other instead of leaving the losing file in place.
Follow up on: https://github.com/zed-industries/zed/pull/49519
This PR reworks how Zed calculates diff num stats by moving the
calculation to the `RepositorySnapshot` layer, instead of the
`GitPanel`. This has a couple of benefits:
1. Snapshot recalculations are already set up to recompute on file
system changes and only update the affected files. This means that diff
stats don't need to manage their own subscription or states anymore like
they did in the original PR.
2. We're able to further separate the data layer from the UI. Before,
the git panel owned all the subscriptions and tasks that refreshed the
diff stat, now the repository does, which is more inline with the code
base.
3. Integration tests are cleaner because `FakeRepository` can handle all
the data and calculations of diff stat and make it accessible to more
tests in the codebase. Because a lot of tests wouldn't initialize the
git panel when they used the git repository.
4. This made implementing remote/collab support for this feature
streamline. Remote clients wouldn't get the same buffer events as local
clients, so they wouldn't know that the diff stat state has been updated
and invalidate their data.
5. File system changes that happened outside of Zed now trigger the diff
stat refresh because we're using the `RepositorySnapshot`.
I added some integration tests as well to make sure collab support is
working this time. Finally, adding the initial diff calculation to
`compute_snapshot` didn't affect performance for me when checking
against chromium's diff with HEAD~1000. So this should be a safe change
to make.
I decided to add diff stats on the status entry struct because it made
updating changed paths and the collab database much simpler than having
two separate SumTrees. Also whenever the UI got a file's status it would
check its diff stat as well, so this change makes that code more
streamlined as well.
Before you mark this PR as ready for review, make sure that you have:
- [x] Added a solid test coverage and/or screenshots from doing manual
testing.
- [x] Done a self-review taking into account security and performance
aspects.
Release Notes:
- N/A
This PR cleans up the git command spawning by wrapping everything in
GitBinary instead to follow a builder/factory pattern. It also extends
trusted workspace support to git commands.
I also added a `clippy.toml` configuration to our git crate that warns
against using `Command` struct to spawn git commands instead of going
through `GitBinary`. This should help us maintain the factory pattern in
the future
Before you mark this PR as ready for review, make sure that you have:
- [x] Added solid test coverage and/or screenshots from doing manual
testing
- [x] Done a self-review taking into account security and performance
aspects
Release Notes:
- git: Add trusted workspace support for Zed's git integration
The main issue is that we weren't forwarding the proto messages through
the collab server to the host. After fixing that I added integration
tests to cover local worktrees, remote worktrees, and ssh worktrees.
I also fixed a bug with FakeRepository where it wouldn't name its
current branch as a worktree when calling git worktree, which doesn't
match the behavior of the git binary.
Before you mark this PR as ready for review, make sure that you have:
- [x] Added a solid test coverage and/or screenshots from doing manual
testing
- [x] Done a self-review taking into account security and performance
aspects
Release Notes:
- git: Fix bug that caused the git worktree picker from displaying and
creating worktrees over collab
Add `agent_worktree_directory` to `GitSettings` for configuring where
agent worktrees are stored (default: Zed data dir). Remove `Copy` derive
from `GitSettings`/`GitContentSettings` (incompatible with String field)
and fix downstream `.as_ref().unwrap()` call sites.
Define `AgentGitWorktreeInfo` (branch, worktree_path, base_ref) and add
it to `DbThread` + `DbThreadMetadata` for persistence and session list
display.
Closes AI-33
Release Notes:
- N/A
Add `remove_worktree()` and `rename_worktree()` to the `GitRepository`
trait with `RealGitRepository` implementations that shell out to `git
worktree remove/move`.
Implement all 4 worktree methods (`worktrees`, `create_worktree`,
`remove_worktree`, `rename_worktree`) on `FakeGitRepository` backed by
`FakeGitRepositoryState`, with `simulated_create_worktree_error` for
test-time fault injection.
Add `set_create_worktree_error()` helper on `FakeFs`.
Add `parse_worktrees_from_str` helper and 7 new tests covering real git
operations and fake worktree lifecycle.
Closes AI-31
Release Notes:
- N/A
Here's some backstory:
* on macOS, @cole-miller and I noticed that since roughly Oct 2025, due
to some changes to latest macOS Tahoe, for any spawned child process we
needed to reset Mach exception ports
(https://github.com/zed-industries/zed/issues/36754 +
https://github.com/RemiKalbe/zed/commit/6e8f2d2ebe5a93753625a3026aeb996de8cb436b)
* the changes in that PR achieve that via `pre_exec` hook on
`std::process::Command` which then abandons `posix_spawn` syscall for
`fork` + `execve` dance on macOS (we tracked it down in Rust's std
implementation)
* as it turns out, `fork` + `execve` is pretty expensive on macOS
(apparently way more so than on other OSes like Linux) and `fork` takes
a process-wide lock on the allocator which is bad
* however, since we wanna reset exception ports on the child, the only
official way supported by Rust's std is to use `pre_exec` hook
* posix_spawn on macOS exposes this tho via a macOS specific extension
to that syscall `posix_spawnattr_setexceptionports_np` but there is no
way to use that via any standard interfaces in `std::process::Command`
* thus, it seemed like a good idea to instead create our own custom
Command wrapper that on non-macOS hosts is a zero-cost wrapper of
`smol::process::Command`, while on macOS we reimplement the minimum to
achieve `smol::process::Command` with `posix_spawn` under-the-hood
Notably, this changeset improves git-blame in very large repos
significantly.
Release Notes:
- Fixed performance spawning child processes on macOS by always forcing
`posix_spawn` no matter what.
---------
Co-authored-by: Cole Miller <cole@zed.dev>
Reverts zed-industries/zed#47799
Now that we have a week to let this be tested in Nightly we can merge it
again
Release Notes:
- Fix to many file descriptors error in MacOS
Closes#47064Closes#20806
The previous implementation used Zed's custom fsevent crate, which
spawned a dedicated thread for each watched path. On large projects,
this could result in 100+ threads just for filesystem watching.
This PR removes the `fsevent` crate and switches to using `notify's`
FsEventWatcher directly. The notify implementation maintains a single
FSEvents stream that watches all paths on one thread. When paths are
added or removed, the stream is stopped, the path list is updated, and
the stream is restarted.
Ref: https://github.com/notify-rs/notify/blob/main/notify/src/fsevent.rs
As a result, Zed now uses one thread for filesystem watching regardless
of how many paths are watched.
Release Notes:
- On macOS, Zed now uses significantly fewer resources when watching
filesystem changes
---------
Signed-off-by: Marco Mihai Condrache <52580954+marcocondrache@users.noreply.github.com>
Co-authored-by: Anthony Eid <anthony@zed.dev>
Closes#26866
### Summary
Adds a git graph to Zed, accessible via the `git_graph::Open` action if
a project has an active repository. There's still more to do, but this
is a solid foundation to expand upon. The code structure is in line with
Zed's codebase and shouldn't require architectural changes to add
missing features.
The git graph can be opened via the command palette (`git graph: open`)
or by binding a key to `git_graph::Open`. It's available when the
project has an active git repository.
### Architecture
Similar to the Debugger, the git graph is split between a data layer and
a view/UI layer. When the view layer is rendering, it queries the data
layer for its active state. This setup allows the data layer to lazily
request graph data (only when needed for rendering), abstracts collab
from the view layer, allows most of the data loading to happen on a
background thread, and makes caching easy to implement.
#### Graph Loading
The graph data is loaded in two phases:
1. `Repository::graph_data()` streams commit structure (SHA, parents,
refs) in chunks of 1000 via `git log`
2. `CommitDataReader` lazily fetches full commit details (author,
timestamp, subject) on-demand using a persistent `git cat-file --batch`
process
This two-phase approach makes the initial loading of the graph as fast
as possible, because `git log` takes significantly longer when all the
needed graph data is queried through it. Zed then lazily loads commits
in the user's viewport through `cat-file --batch`. This makes scrolling
to any place in the graph extremely snappy and benefits the
collaborative architecture by only fetching data needed to render the
graph. It also allows Zed to share commit data between different graph
visualizations (e.g., date order vs. topological order).
#### Performance
Tested on both the Zed and LLVM repositories with good performance in
both cases. The two-phase loading approach and lazy fetching keep the UI
responsive even with large commit histories.
#### Testing
I added property testing that builds randomized commit graphs and
verifies that the graph is constructed correctly. This also works as an
integration test and will be expanded in the future to test collab graph
visualization, graph filtering, commit actions, etc.
### New Crate
- `git_graph` (GPL-licensed) — contains UI and graph computation logic
### Not Yet Implemented
- Remote repository support (collab)
- Filtering by branch
- Commit actions (checkout, cherry-pick, etc.)
- Search
- Open commit view for selected commit
- Resizable columns
- Column filtering
#### Reference
<img width="1624" height="976" alt="Screenshot 2025-01-22 at 8 15 39 PM"
src="https://github.com/user-attachments/assets/0f10924a-3964-462f-b320-42d84d02f7bf"
/>
Special thanks to [Alberto Slavica](https://github.com/pyundev) for
submitting #44405, which was a good base to work off of.
Release Notes:
- git: Add initial version of git graph
---------
Co-authored-by: pyundev <pyundev@users.noreply.github.com>
Co-authored-by: Cole Miller <cole@zed.dev>
Co-authored-by: Zed Zippy <234243425+zed-zippy[bot]@users.noreply.github.com>
Release Notes:
- Opening bundled files, keymap, and local release notes now opens in
remote windows instead of opening a new local zed window
- Opening the settings files, keymap files, task files, debug files and
logs will now open within wsl windows instead of opening a new local zed
window
## Motivation
This PR unifies the async execution infrastructure between GPUI and
other components that depend on the `scheduler` crate (such as our cloud
codebase). By having a scheduler that lives independently of GPUI, we
can enable deterministic testing across the entire stack - testing GPUI
applications alongside cloud services with a single, unified scheduler.
## Summary
This PR completes the integration of the `scheduler` crate into GPUI,
unifying async execution and enabling deterministic testing of GPUI
combined with other components that depend on the scheduler crate.
## Key Changes
### Scheduler Integration (Phases 1-5, previously completed)
- `TestDispatcher` now delegates to `TestScheduler` for timing, clock,
RNG, and task scheduling
- `PlatformScheduler` implements the `Scheduler` trait for production
use
- GPUI executors wrap scheduler executors, selecting `TestScheduler` or
`PlatformScheduler` based on environment
- Unified blocking logic via `Scheduler::block()`
### Dead Code Cleanup
- Deleted orphaned `crates/gpui/src/platform/platform_scheduler.rs`
(older incompatible version)
## Intentional Removals
### `spawn_labeled` and `deprioritize` removed
The `TaskLabel` system (`spawn_labeled`, `deprioritize`) was removed
during this integration. It was only used in a few places for test
ordering control.
cc @maxbrunsfeld @as-cii - The new priority-weighted scheduling in
`TestScheduler` provides similar functionality through
`Priority::High/Medium/Low`. If `deprioritize` is important for specific
test scenarios, we could add it back to the scheduler crate. Let me know
if this is blocking anything.
### `start_waiting` / `finish_waiting` debug methods removed
Replaced by `TracingWaker` in `TestScheduler` - run tests with
`PENDING_TRACES=1` to see backtraces of pending futures when parking is
forbidden.
### Realtime Priority removed
The realtime priority feature was unused in the codebase. I'd prefer to
reintroduce it when we have an actual use case, as the implementation
(bounded channel with capacity 1) could potentially block the main
thread. Having a real use case will help us validate the design.
## Testing
- All GPUI tests pass
- All scheduler tests pass
- Clippy clean
## Architecture
```
┌─────────────────────────────────────────────────────────────┐
│ GPUI │
│ ┌──────────────────────┐ ┌────────────────────────────┐ │
│ │ gpui::Background- │ │ gpui::ForegroundExecutor │ │
│ │ Executor │ │ - wraps scheduler:: │ │
│ │ - scheduler: Arc< │ │ ForegroundExecutor │ │
│ │ dyn Scheduler> │ └────────────┬───────────────┘ │
│ └──────────┬───────────┘ │ │
│ │ │ │
│ └──────────┬──────────────────┘ │
│ ▼ │
│ ┌───────────────────────┐ │
│ │ Arc<dyn Scheduler> │ │
│ └───────────┬───────────┘ │
│ ┌──────────────┴──────────────┐ │
│ ▼ ▼ │
│ ┌──────────────────┐ ┌────────────────────┐ │
│ │ PlatformScheduler│ │ TestScheduler │ │
│ │ (production) │ │ (deterministic) │ │
│ └──────────────────┘ └────────────────────┘ │
└─────────────────────────────────────────────────────────────┘
```
Release Notes:
- N/A
---------
Co-authored-by: Antonio Scandurra <me@as-cii.com>
Co-authored-by: Claude Opus 4.5 <noreply@anthropic.com>
Co-authored-by: Yara <git@yara.blue>
Co-authored-by: Zed Zippy <234243425+zed-zippy[bot]@users.noreply.github.com>
This happens when a user's default branch has different remote and local
versions on their system. Because we would get a diff using just a
branch name without including it's remote name as well
This is a better default because it's far less confusing from a user's
perspective to debug what's going on.
Release Notes:
- git: Fix BranchDiff showing incorrect diffs when default local branch
is out of sync with the remote default branch
Closes https://github.com/zed-industries/zed/issues/42845
Repro steps:
https://github.com/zed-industries/zed/issues/42845#issuecomment-3687413958
Initial investigation and Zed memory trace:
https://github.com/zed-industries/zed/issues/42845#issuecomment-3687877977
The PR consists of 2 commits:
*
[first](https://github.com/zed-industries/zed/pull/45669/changes/732d308c8d7e9af3649ac71ea65a9c029af820fc)
adds cosmetic fixes to remove backtraces from logs yet again and print
paths in quotes, as file descriptors may return empty paths.
It also stubs the cause if OOM in project panel: that one traversed all
worktrees in `for worktree_snapshot in visible_worktrees` and "accepted"
the one with empty paths + never called `entry_iter.advance();` in "no
file name found for the worktree" case, thus looping endlessly and
bloating the memory quite fast.
*
[second](https://github.com/zed-industries/zed/pull/45669/changes/7ebfe5da2fc6d32f3fa2d71c761f8b2ec26d945b)
adds something that resembles a fix: `fn current_path` on macOS used the
file handler to re-fetch the worktree root file path on worktree root
canonicalization failure.
What's odd, is that `libc::fcntl` returns `0` in the case when external
volume is not mounted, thus resulting in the `""` path string that is
propagated all the way up.
*
[third](https://github.com/zed-industries/zed/pull/45669/changes/1a7560cef3e9fac604124c19f46b1f9c7b91815f)
moves the fix down to the platform-related FS implementations
The "fix" now checks the only usage of this method inside `async fn
process_events` for an empty path and bails if that is the case.
I am not sure what is a better fix, but this stops any memory leaks and
given how bad the situation now, seems ok to merge for now with the
`TODO` comment for more clever people to fix properly later.
----------------
Now, when I disconnect the SMB share and reconnect it again, Zed stops
displaying any files in the project tree but the ones opened as editors.
As before, at first, when the share is unmounted, Zed fails to save any
changes because of the timeouts.
Later, when the share is re-connected, macOS Finder hangs still but Zed
starts to react on saves yet still only shows the files that are open as
editors.
The files can be edited and saved from now on.
Later, when Finder finally stops hanging and indicates that the share is
mounted fully, the rest of the file structure reappear in the project
panel, and all file saves are propagated, hence can be observed in the
share in Finder.
It feels that one good improvement to add on top is some "disconnected"
indicator that clearly shows that the file is not properly handles in
the OS.
This requires much more changes and thinking as nothing like that exists
in Zed yet, hence not done.
Release Notes:
- Fixed Zed OOM-ing when macOS file descriptors become invalid
This PR improves the `edit prediction: Capture Example` in several ways:
* fixed bugs in how the uncommitted diff was calculated
* added a `edit_predictions.examples_dir` setting that can be set in
order to have the action automatically save examples into the given
folder
* moved the action into the `edit_predictions` crate, in preparation for
collecting this data passively from end users, when they have opted in
to data sharing, similar to what we did for Zeta 1
Release Notes:
- N/A
Update the behavior of `git::Amend` to ensure that the latest head
commit message, if available, is always loaded into the commit message
editor, regardless of its state. The previous text, if any, is now also
restored after the amend is finished.
- Update `FakeGitRepository.show` to include a message in the returned
`CommitDetails` so we can assert that this specific commit message is
set in the commit message editor.
- Add default implementation for `FakeGitRepository.commit` and
`FakeGitRepository.run_hook` to ensure that tests are able to run and
don't panic on `unimplemented!()`
- Refactor `GitPanel.load_last_commit_message_if_empty` to
`GitPanel.load_last_commit_message`, ensuring that the head commit
message is always loaded, regardless of whether the commit message
editor is empty.
- Update `GitPanel.commit_changes` to ensure that the pending amend
state is only updated if the editor managed to actually commit the
changes. This also ensures that we don't restore the commit message
editor's contents when amending a commit, before the amend is actually
processed.
- Update `CommitModal.amend`, removing the call to
`GitPanel.set_amend_pending` as that is now handled by the background
task created in `GitPanel.commit_changes`.
- Split the `commit` and `amend` methods from the event handlers so that
the methods can be called directly, as is now being done by
`CommitModal.on_commit` and `CommitModal.on_amend`.
Release Notes:
- Updated the `git: amend` command to always load the latest head
commit message, and to restore any previously entered text in the commit
message editor after the amend completes
The bug was introduced in this recent PR:
https://github.com/zed-industries/zed/pull/42819. Since it's still in
nightly, there is no need for release notes.
I also polished the feature a bit by:
- Ensuring branch names are always a single line so the branch picker's
uniform list uses the correct element height.
- Adding tooltip text for the filter remotes button.
- Fixing the create branch from default icon showing up for non-new
branch entries.
Release Notes:
- N/A
Closes#26649
Release Notes:
- Improved git support for remotes handling. Git will now check pushRemote and pushDefault configurations before
falling back to branch remote.
---------
Co-authored-by: Jakub Konka <kubkon@jakubkonka.com>
Update the `fs::RenameOptions` used by
`project::lsp_store::LocalLspStore.deserialize_workspace_edit` in order
to always set `create_parents` to `true`. Doing this ensures that we'll
always create the folders for the new file path provided by the language
server instead of failing to handle the request in case the parent
- Introduce `create_parents` field to `fs::RenameOptions`
- Update `fs::RealFs.rename` to ensure that the `create_parents` option
is respected
Closes#41820
Release Notes:
- Fixed a bug where using language server's file renaming actions could
fail if the parent directory of the new file did not exist
Closes#28667
Release Notes:
- Fixed git not preserving file mode when committing. Now if an input file is executable it will be preserved when committed with Zed.
---------
Signed-off-by: 11happy <soni5happy@gmail.com>
Signed-off-by: 11happy <bhuminjaysoni@gmail.com>
Co-authored-by: Jakub Konka <kubkon@jakubkonka.com>
Path canonicalization on windows will now favor keeping the drive letter
intact when canonicalizing paths. This helps some lsps with mapped
network drive compatibility.
Closes#41336
Release Notes:
- N/A
We now run git pre-commit hooks before we commit. This ensures we don't
run into timeout issues with askpass delegate and report invalid error
to the user.
Closes#43157
Release Notes:
- Fixed long running pre-commit hooks causing committing from Zed to
fail.
Co-authored-by: Cole Miller <cole@zed.dev>
This reverts commit b0a7defd09.
It looks like this doesn't interact correctly with the project diff or
with staging, let's revert and reland with bugs fixed.
Release Notes:
- N/A
Adds a simple notification when cloning a repo using the integrated git
clone on Zed. Before this, the user had no feedback after starting the
cloning action.
Demo:
https://github.com/user-attachments/assets/72fcdf1b-fc99-4fe5-8db2-7c30b170f12f
Not sure about that icon I'm using for the animation, but that can be
easily changed.
Release Notes:
- Added notification when cloning a repo from zed
New error variants were stabilized in 1.83, and this might've led to us
mis-handling not-a-directory errors.
Co-authored-by: Dino <dino@zed.dev>
Release Notes:
- N/A
Co-authored-by: Dino <dino@zed.dev>
In my local setup, I always enforce git-commit signing with GPG/SSH
which automatically enforces `git commit -S` when committing. This
changeset will now show a modal to the user for them to specify the
passphrase (if any) so that they can unlock their private key for
signing when committing in Zed.
<img width="1086" height="948" alt="Screenshot 2025-11-07 at 11 09
09 PM"
src="https://github.com/user-attachments/assets/ac34b427-c833-41c7-b634-8781493f8a5e"
/>
Release Notes:
- Handle automatic git-commit signing by presenting the user with an
askpass modal
Closes [#41674](https://github.com/zed-industries/zed/issues/41674)
Description:
Creating a branch from a base requires switching to the base branch
first, then creating the new branch and checking out to it, which
requires multiple operations.
Add base_branch parameter to create_branch to allow a new branch from a
base branch in one operation which is synonymous to the command `git
switch -c <new-branch> <base-branch>`.
Below is the video after solving the issue:
(`master` branch is the default branch here, and I create a branch
`new-branch-2` based off the `master` branch. I also show the error
which used to appear before the fix.)
[Screencast from 2025-11-07
05-14-32.webm](https://github.com/user-attachments/assets/d37d1b58-af5f-44e8-b867-2aa5d4ef3d90)
Release Notes:
- Fixed the branch-picking error by replacing multiple sequential switch
operations with just one switch operation.
Signed-off-by: ayu-ch <ayu.chandekar@gmail.com>
- Adds a new action `git::PullRebase` which adds `--rebase` in the final
command invoked by existing Git-Pull implementation.
- Includes the new action in "Fetch/Push" button in the Git Panel
(screenshot below)
- Adds key-binding for `git::PullRebase` in all three platforms,
following the existing key-binding patterns (`ctrl-g shift-down`)
- Update git docs to include the new action.
Sidenote: This is my first ever OSS contribution
Screenshot:
<img width="234" height="215" alt="image"
src="https://github.com/user-attachments/assets/713d068f-5ea5-444f-8d66-444ca65affc8"
/>
---
Release Notes:
- Git: Added `git: pull rebase` for running `git pull --rebase`.
Related discussions #26084
Worktree creations are implemented similar to how branch creations are
handled on the branch picker (the user types a new name that's not on
the list and a new entry option appears to create a new branch with that
name).
https://github.com/user-attachments/assets/39e58983-740c-4a91-be88-57ef95aed85b
With this picker you have a few workflows:
- Open the picker and type the name of a branch that's checked out on an
existing worktree:
- Press enter to open the worktree on a new window
- Press ctrl-enter to open the worktree and replace the current window
- Open the picker and type the name of a new branch or an existing one
that's not checked out in another worktree:
- Press enter to create the worktree and open in a new window. If the
branch doesn't exists, we will create a new one based on the branch you
have currently checked out. If the branch does exists then we create a
worktree with that branch checked out.
- Press ctrl-enter to do everything on the previous point but instead,
replace the current window with the new worktre.
- Open the picker and type the name of a new branch or an existing one
that's not checked out in another worktree:
- If a default branch is detected on the repo, you can create a new
worktree based on that branch by pressing ctrl-enter or
ctrl-shift-enter. The first one will open a new window and the last one
will replace the current one.
Note: If you preffer to not use the system prompt for choosing a
directory, you can set `"use_system_path_prompts": false` in zed
settings.
Release Notes:
- Added git worktree picker to open a git worktree on a new window or
replace the current one
- Added git worktree creation action
---------
Co-authored-by: Cole Miller <cole@zed.dev>
Mikayla Maki