Closes#46574
The `default_working_directory()` function in
`crates/terminal_view/src/terminal_view.rs:1585` returned `None` when no
project directory was available. The code comment (now removed)
incorrectly claimed "None implies `~` on whichever machine we end up
on". However, when `None` is passed to `alacritty_terminal`, it uses the
process CWD, not the home directory. On macOS when Zed is launched from
a .`app` bundle, the CWD is `/`.
Added a fallback at the end of` default_working_directory()` that
explicitly returns the home directory when no project directory is
found: `directory.or_else(dirs::home_dir)`
This ensures:
1. `CurrentProjectDirectory` with no project open → home directory
2. `FirstProjectDirectory `with no project open → home directory
3. `AlwaysHome `→ home directory (explicitly, not relying on shell
behavior)
4. Always `{ directory }` with invalid directory → home directory
Release Notes:
- Fixed terminal opening in `/` instead of home directory when no
project is open.
Closes#46365
The behavior I'm trying to handle here is documented in
https://man7.org/linux/man-pages/man5/proc_pid_exe.5.html.
[@merlinz01
](https://github.com/zed-industries/zed/issues/46365#issuecomment-3723996684)
points to some other alternatives but they don't seem necessary or
justified here.
# Testing
I've manually replaced the binary of a process running this build, to
simulate an update, and then opened a new tree and I don't see the error
message.
I don't know if there's a reasonable way to unit-test this? Let me know.
It seems like it would require using something like rusty-fork to spawn
a specific binary for the test, which I don't think I see already in the
Zed tree and perhaps it would be overkill to add just for this?
Release Notes:
- Fixed "Failed to load environment variables" after applying an update
This should move almost all access of the threadstore behind the trait,
which unlocks external agents supplying the list
Release Notes:
- N/A
---------
Co-authored-by: Zed Zippy <234243425+zed-zippy[bot]@users.noreply.github.com>
Some language servers return local symbols in `workspace/symbol`
responses, and languages like Julia can have multiple overloads for the
same generic function name. In these cases, symbols may appear identical
in the symbol picker, making it hard to distinguish between them. Adding
line numbers helps users identify the correct symbol.
In the future, we may also want to consider using `containerName` from
the LSP response for additional context.
> Before
<img width="671" height="573" alt="Screenshot 2026-01-10 at 19 21 08"
src="https://github.com/user-attachments/assets/838c3eb7-0e1d-46ba-9076-286e971afc2c"
/>
> After
<img width="671" height="573" alt="Screenshot 2026-01-10 at 19 43 19"
src="https://github.com/user-attachments/assets/31979d79-0169-4202-afc2-24fc7492dd47"
/>
Closes #ISSUE
Release Notes:
- Added line numbers to the project symbols picker to help distinguish
between symbols with the same name
---------
Co-authored-by: Matt Miller <mattrx@gmail.com>
Adds a very simple diff review button behind a feature flag, which
appears in the gutter of people who have the flag (currently just me;
it's not even enabled for staff, since it doesn't do anything yet).
Mostly this PR just adds the feature flag, the button, and tests.
Release Notes:
- N/A
The scheduler integration (#44810) removed the special handling for
realtime audio tasks that spawn them on a dedicated thread. This caused
a panic in the Mac dispatcher when RealtimeAudio priority was passed to
the regular dispatch path.
This restores the original behavior: RealtimeAudio tasks are spawned on
a dedicated thread via dispatcher.spawn_realtime(), using a bounded
channel to send runnables to it.
Release Notes:
- N/A
Add permission checking to the remaining 7 tools that require granular
permissions:
- `edit_file`: Checks path against permission rules
- `delete_path`: Checks path against permission rules
- `move_path`: Checks both source and destination paths
- `create_directory`: Checks path against permission rules
- `save_file`: Checks all paths, denies if any are blocked
- `fetch`: Checks URL against permission rules
- `web_search`: Checks query against permission rules
Each tool follows the pattern established in PR #46155 (terminal tool):
- `Allow` = proceed without prompting
- `Deny` = return error immediately
- `Confirm` = prompt user for confirmation
The deny > confirm > allow precedence is enforced by the
`decide_permission_from_settings()` function.
Release Notes:
- N/A
---------
Co-authored-by: Amp <amp@ampcode.com>
This PR creates a unified picker for all the Git-related pickers:
branch, worktree and stash. The main motivation for this is to bring
awareness and ease of access for the worktree and stash pickers that
were previously (and arguably) hidden amidst the sea of commands in the
command palette. As worktrees in particular become more relevant for
AI-related flows, having an easier way to reach for its picker will be
beneficial.
Note that the actions/commands remain the same; you can still look for
`git: worktree` or `git: view stash`. The difference is that these
actions will take you to the unified picker with the correct
corresponding active view.
https://github.com/user-attachments/assets/99d1cd6f-a19d-47d3-9bca-d7257e7ed5b8
Release Notes:
- Git: Unify the branch, worktree, and stash pickers into one, making it
easier to find and access them from one another.
New workspace/window creates a different `WeakEntity<WorktreeStore>` for
the same path, which was not considered before.
Closes https://github.com/zed-industries/zed/issues/46318
Release Notes:
- Fixed worktree trust not applied in window reuse
Release Notes:
- Fixed zed no longer rendering on windows in a timely manner if it gets
overloaded with foreground work
Co-authored-by: John Tur <john-tur@outlook.com>
Closes https://github.com/zed-industries/zed/issues/46590https://github.com/user-attachments/assets/8464a3b6-9258-4d15-9168-56178bf46437
`use smol::channel::{Receiver, Sender};` do not propagate the value
change to all receivers, hence only one language server received the
"worktree trusted" message.
The fix uses `watch` kind of channels instead, and cleans up its state
more eager to rely on the sender drop as another form of notification.
Release Notes:
- Fixed groups of language servers not starting after worktree trust
approval
Stumbled across this upon a PR review - `unfold_ranges` takes a slice of
Ranges and already clones the ranges itself, so we do not need to clone
the selections prior to passing these.
Release Notes:
- N/A
Closes https://github.com/zed-industries/zed/issues/45553
## Summary
When using applications that hide the terminal cursor (e.g., via ANSI
escape sequence `\e[?25l`), the IME candidate window appeared at the
bottom-left of the screen, and the composition text was not displayed at
all. This affects tools like Claude Code that hide the cursor during
operation.
The root cause was that cursor position calculation was skipped when the
cursor shape was `Hidden`. This fix separates cursor visibility from IME
position calculation by always computing cursor bounds for IME
positioning.
## Movies
### Zed 0.218.6
https://github.com/user-attachments/assets/a66d569a-178d-4bca-a577-fcc65e9fd2c4
### With this fix
https://github.com/user-attachments/assets/456323d1-6c18-45ea-88e8-fb5375c00f98
Release Notes:
- Fixed IME candidate window and composition text position in terminal
when cursor is hidden
Release Notes:
- N/A
Description:
E.g. Instead of saying:
2026-01-07T09:08:42-06:00 ERROR [crates/snippet_provider/src/lib.rs:46]
failed to parse snippet
Caused by:
expected an integer
Now the logs give more valuable information like:
2026-01-07T10:47:39-06:00 ERROR [crates/snippet_provider/src/lib.rs:48]
invalid snippet in /Users/oscarvarto/Library/Application
Support/Zed/extensions/installed/django-snippets/./snippets/python.json
(re_path)
Closes#40110
Changes:
- `DeleteToPreviousSubwordStart` now deletes `\n` separately from
preceding subwords and whitespace.
- `DeleteToNextSubwordEnd` now deletes `\n` and any following whitespace
separately from subsequent subwords.
- Added an `ignore_newlines` flag to both actions to optionally retain
the old behavior.
These modifications align the subword commands with their word
counterparts and with other popular editors like VSCode and Sublime.
Related to: https://github.com/zed-industries/zed/pull/16848
Release Notes:
- Improved `DeleteToPreviousSubwordStart` and `DeleteToNextSubwordEnd`
interactions around newlines. You can opt-in into the previous behavior
by adding `{"ignore_newlines": true}` to either action's binds in your
keymap.
---
This is my first contribution to Zed! If anything should be done
differently, please let me know. Happy to learn :)
Previously, a new composition string was processed before a new
composition result. Thus, in the case that a keystroke creates a new
composition result and composition string, the processing of the new
result would overwrite the new string, even though they should both be
displayed. Processing them in the reverse order fixes this issue.
Fixes#42201
Additionally: while fixing this issue, I discovered that the Japanese
IME sometimes moves the cursor far away from the inserted text. WPF
works around this issue by only respecting the IME's requested cursor
position if it's adjacent to uncommitted text. So, we copy this
workaround.
Release Notes:
- N/A
PR #46306 changed cancellation to wait for tools to complete before
returning. This was correct behavior - it allows tools like terminal to
capture their output on cancellation. The real issue was that many tools
didn't check for cancellation, so they would continue running until they
finished.
## The Problem
When the user pressed Escape to cancel during a tool operation, tools
would continue running because they never checked for the cancellation
signal. The thread correctly waited for tools to complete (so terminal
could capture output), but tools like edit_file, grep, fetch, etc. would
just keep going.
## The Fix
Add cancellation handling to all tools using the same pattern as
`terminal_tool`: use `select!` to race between the tool's main work and
`event_stream.cancelled_by_user()`. When cancelled, tools break out of
their loops or return early.
## All Tools Now Cancellation-Aware
| Tool | Change |
|------|--------|
| `edit_file_tool` | Checks cancellation in edit event processing loop |
| `terminal_tool` | Already handled cancellation |
| `grep_tool` | Checks cancellation in search result iteration loop |
| `fetch_tool` | Checks cancellation during HTTP fetch |
| `web_search_tool` | Checks cancellation during web search |
| `find_path_tool` | Checks cancellation during path search |
| `read_file_tool` | Checks cancellation during buffer open |
| `copy_path_tool` | Checks cancellation during file copy |
| `move_path_tool` | Checks cancellation during file move/rename |
| `delete_path_tool` | Checks cancellation during delete operation |
| `create_directory_tool` | Checks cancellation during directory
creation |
| `save_file_tool` | Checks cancellation during buffer open and save |
| `restore_file_from_disk_tool` | Checks cancellation during buffer open
and reload |
| `open_tool` | Checks cancellation during authorization |
| `diagnostics_tool` | Checks cancellation during buffer open |
| `ContextServerTool` (MCP) | Checks cancellation during external server
calls |
**Synchronous tools (no async work, return immediately):**
- `list_directory_tool` - Reads worktree snapshot synchronously
- `now_tool` - Returns current time immediately
- `thinking_tool` - Returns immediately
## MCP Tools Automatically Handled
MCP tools (user-defined tools via context servers) are now automatically
cancellation-aware without any user action. The `ContextServerTool`
wrapper races the external server request against
`event_stream.cancelled_by_user()`.
## Testing
- Added `CancellationAwareTool` test helper that mirrors the
cancellation pattern
- Updated `test_cancellation_aware_tool_responds_to_cancellation` to
properly await the cancel task and verify the tool detected cancellation
Release Notes:
- Fixed a regression where pressing Escape wouldn't immediately cancel
in-progress tool operations
This PR makes it so we use a proper type for the Responses API `input`
rather than a `serde_json::Value`.
It should have never used `serde_json::Value` to begin with.
Release Notes:
- N/A
## 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 PR adds a right-click action to the breadcrumb button to copy the
path, when there is some (i.e., untitled buffers and others won't have
this feature).
https://github.com/user-attachments/assets/94ece12c-3071-4a07-a04d-8d89e10fd59e
Release Notes:
- Added the ability to right-click the breadcrumb button when it's a
file path and copy the path.
## Summary
Fixes#33633 - folds corrupting after external file modifications.
The existing fold persistence stored raw byte offsets, which become
stale when files are modified externally (git operations, other editors,
sync tools). This caused folds to capture wrong content on restore -
users reported "wrong lines getting folded" with systematic offsets.
**Solution: Content fingerprinting**
- Store 32-byte content samples at fold boundaries
- On restore, validate fingerprints match; if not, search buffer for new
positions
- Handles edge cases: short folds (< 32 bytes), duplicate content,
boundary positioning
**Bonus fix: Ungraceful exit survival**
- Entity IDs change between sessions, but workspace cleanup's CASCADE
DELETE was wiping folds before new editors could save
- Now migrates folds to new entity_id immediately after restore
## Test plan
- [x] Unit test for fingerprint storage/retrieval
(`test_save_and_get_editor_folds_with_fingerprints`)
- [x] Manual test: fold sections, add content at file start externally,
reopen → folds restore at correct positions
- [x] Manual test: fold sections, Ctrl+C, reopen → folds survive
- [x] Existing fold tests pass (`cargo test -p editor`)
## Release Notes
- N/A
---------
Co-authored-by: Hector <hector@cyberneticwilderness.com>
This PR adds support for using the OpenAI Responses API through the Zed
provider.
This is gated behind the `open-ai-responses-api` feature flag.
Part of CLO-34.
Release Notes:
- N/A
This makes it a smoother workflow to capture an edit prediction example
from Zed and then write what you'd expect the patch to be.
Release Notes:
- N/A
Danilo Leal