This PR adds two more cases to existing LSP completion cases.
- When text_edit exists: (New test)
1. we use text_edit, over insert_text and label
- When edit range exists (and text_edit is None): (New test)
1. insert_text is used over label if exists
2. label is used otherwise
- When not edit range exists (and text_edit is None): (Existing test)
1. insert_text is used over label if exists
2. label is used otherwise
Release Notes:
- N/A
Closes#18153
When calling `uninstall_extension`, the `work_dir` associated with this
`extension` doesn't have its corresponding `FileHandle` properly closed,
preventing the deletion of the `work_dir`. As seen in the image below,
after installing the `toml` extension, `zed.exe` holds two `Handle`s for
the folder `C:\Users\36477\AppData\Local\Zed\extensions\work\toml`.
Therefore, after deleting `extension_dir` and then calling
`this.update(...)`, `zed.exe` releases these two `Handles`, and only
then can the folder
`C:\Users\36477\AppData\Local\Zed\extensions\work\toml` be deleted. See
the corresponding file handles are closed after calling
`this.update(...)`:
However, if there is a running server of the extension, the error will
persist. At this point, I haven’t found a direct way to terminate all
running servers of the extension. Since this feature might affect the
`LspStore` structure, I paused my work here.
See when `toml` extension is running, we can not delete
`C:\Users\36477\AppData\Local\Zed\extensions\work\toml` since
`C:\Users\36477\AppData\Local\Zed\extensions\work\toml\taplo.exe` is
still running:
cc @ConradIrwin You're the expert in this area—what are your thoughts?
Release Notes:
- N/A
This PR removes the `ZedAssistant2` icon.
I went to use it as a placeholder icon, but noticed that the icon wasn't
loaded properly due to a name mismatch.
However, since we aren't using it anywhere I'm opting to remove it.
Release Notes:
- N/A
This PR fixes an issue where we were looking for context server project
settings in _all_ worktrees, not just visible ones.
This meant that if you had a single file worktree open (e.g.,
`settings.json`) this could impact whether context servers defined in
project settings would load.
Release Notes:
- Fixed an issue where context servers defined in project settings would
not be respected in some scenarios.
Co-authored-by: Wilhelm Klopp <wil.klopp@gmail.com>
This is a clean-up PR in anticipation of introduction of Debugger
Registry. I wanna get rid of DebugAdapterKind (or rather, it being an
enum).
Release Notes:
- N/A
---------
Co-authored-by: Anthony Eid <hello@anthonyeid.me>
Co-authored-by: Anthony <anthony@zed.dev>
<img width="620" alt="Screenshot 2025-03-27 at 2 29 13 PM"
src="https://github.com/user-attachments/assets/dd023507-61bc-4722-a095-f65f4b6c746a"
/>
We'll iterate on the UI, but first the goal is to just get it to work at
all so we can see if it's useful in terms of getting correct output
faster.
Release Notes:
- N/A
---------
Co-authored-by: Agus Zubiaga <hi@aguz.me>
Now if you refocus Zed manually (e.g. cmd-tab), we hide the "View Panel"
notification automatically.
This also fixes a related subscription leak.
Release Notes:
- N/A
Longer write-up, sorry if this got a bit too long.
This PR removes a small gap between the editor gutter and the horizontal
scrollbar, if present, by stretching the scrollbar track the entire
witdth of the editor.
https://github.com/user-attachments/assets/d5c18b03-d1ff-4d48-a3da-5d0fb80ee967
This gap which can be seen in the bottom left of the video can cause
bugs when interacting with it using the cursor, as accidentally clicking
on it would trigger a vertical scroll instead of dragging the horizontal
scroll. Also for cases where themes provide a non-transparent scrollbar
track background, which can be seen in the video, the small gap is
visible whilst scrolling horizontally.
This gap is present because the horizontal editor scrollbar is layouted
based upon the `content_origin`, which offsets the whole layout by the
horizontal gutter margin to the right. However, the scrollbar should be
layouted based upon the editor text bounds to be properly painted over
the entire editor text hitbox.
Here are some comparison images with `scrollbar.track.background` and
`gutter.background` set to red for visibility.
| | Current `main` | With this change |
| - | - | - |
| Default position / Fully scrolled to the left | <img width="842"
alt="left_main"
src="https://github.com/user-attachments/assets/8b053fc8-5271-4b58-8404-dcabf49bf702"
/> | <img width="842" alt="left_fix"
src="https://github.com/user-attachments/assets/459df723-05d5-4813-a6a4-038f7d662495"
/> |
| Scrolled to the right | <img width="216" alt="scroll_main"
src="https://github.com/user-attachments/assets/9c1fcc0d-fbb4-49af-9645-f258f5a7217b"
/> | <img width="216" alt="scroll_fix"
src="https://github.com/user-attachments/assets/8dd2e585-7802-415b-a05a-fb40a882323e"
/> |
---
#### Small downsight of this approach
Currently, the scrollbar thumb aligns with the indent guides if the
editor is fully scrolled to the left and the track background is
transparent. This is because the indent guides are layouted according to
the content margin.
With this change, however, the scrollbar thumb will shift a few pixels
to the left and will overlap the indent guides if present.
| Current `main` | With this change |
| - | - |
| <img width="295" alt="cur_indent"
src="https://github.com/user-attachments/assets/92753951-6f35-4c39-94eb-21c445f8d2f5"
/> | <img width="381" alt="fix_indent"
src="https://github.com/user-attachments/assets/899d945c-49f8-4117-bc48-52501d55cc33"
/> |
To circumvent this, the scrollbar thumb could be layouted with a small
offset so that the thumb aligns properly with the indent guides whilst
the scrollbar track spans the whole editor width. This would lead to
some questions on how to account for the gap during layouting and
dragging of the thumb though, but might work for a gap that small. Happy
to implement this fix, should that be preferred 😄
(VSCode does not have the indent guide issue, as they do not layout the
text in the editor with any offset unlike Zed does)
Release Notes:
- Removed a small gap between the editor gutter and horizontal
scrollbar.
Closes#25761#21603
When `text_edit` is not available we directly fallback to to `label`.
That means, when we have range to replace, we never use `insertText` and
only use it when we haven't found any range.
This PR fixes, this and uses `insertText` as fallback first, and then
`label`.
Release Notes:
- Fixed an issue where accepting LSP snippet completion would insert the
label instead of expanding the snippet.
Part of #27171
Follows-up the change in
https://github.com/zed-industries/zed/pull/22346 to consider the case
where the assistant-panel is disabled via settings (as also noted in
[this
comment](https://github.com/zed-industries/zed/pull/22346#issuecomment-2558372412),
Notably, only the explicit case is considered here. Can extend this
change to also cover the implicit case where the button is disabled if
requested.).
Currently, if the user toggles the right dock, the assistant panel will
be shown even if it is disabled via settings, because it has the highest
priority (see
https://github.com/zed-industries/zed/pull/22346#issuecomment-2564890493).
With this change, the assistant panel is no longer activated when
disabled and the dock with the next highest activation order is
activated instead.
I did not opt in to make the priority configurabe, as I agree with
https://github.com/zed-industries/zed/pull/22346#issuecomment-2564890493
that this will most likely rarely be used (the active panel is only none
on the first toggle of the dock, afterwards it remains set for the
remainder of the session).
Release Notes:
- `workspace::ToggleRightDock` will no longer open the assistant panel
when it is disabled via settings.
This PR hardcodes the font size for the OS notification and adjusts the
copywriting on the `DoneStreaming` scenario.
1. Reason for the former change is because notifications always have a
fixed width and height, so any responsive design strategy here wouldn't
fully work.
2. Reason for the latter is because when the assistant response is done
streaming, that _can_ mean "changes have been applied" (previous label)
but it can also not mean that. So, I'm making it more generic now.
Release Notes:
- N/A
This PR removes the `predict-edits-non-eager-mode` feature flag.
The feature is shipped, and we aren't referencing the flag anywhere
anymore.
Release Notes:
- N/A
This PR contains the following updates:
| Package | Type | Update | Change |
|---|---|---|---|
| [log](https://redirect.github.com/rust-lang/log) |
workspace.dependencies | patch | `0.4.26` -> `0.4.27` |
---
### Release Notes
<details>
<summary>rust-lang/log (log)</summary>
###
[`v0.4.27`](https://redirect.github.com/rust-lang/log/blob/HEAD/CHANGELOG.md#0427---2025-03-24)
[Compare
Source](https://redirect.github.com/rust-lang/log/compare/0.4.26...0.4.27)
</details>
---
### Configuration
📅 **Schedule**: Branch creation - "after 3pm on Wednesday" in timezone
America/New_York, Automerge - At any time (no schedule defined).
🚦 **Automerge**: Disabled by config. Please merge this manually once you
are satisfied.
♻ **Rebasing**: Whenever PR becomes conflicted, or you tick the
rebase/retry checkbox.
🔕 **Ignore**: Close this PR and you won't be reminded about this update
again.
---
- [ ] <!-- rebase-check -->If you want to rebase/retry this PR, check
this box
---
Release Notes:
- N/A
<!--renovate-debug:eyJjcmVhdGVkSW5WZXIiOiIzOS4yMDcuMSIsInVwZGF0ZWRJblZlciI6IjM5LjIwNy4xIiwidGFyZ2V0QnJhbmNoIjoibWFpbiIsImxhYmVscyI6W119-->
---------
Co-authored-by: renovate[bot] <29139614+renovate[bot]@users.noreply.github.com>
Co-authored-by: Kirill Bulatov <kirill@zed.dev>
Release Notes:
- Git: Fix crash when staging a hunk that overlaps multiple unstaged
hunks.
---------
Co-authored-by: Max Brunsfeld <maxbrunsfeld@gmail.com>
This is another in the series of PRs to make the GitStore own all
repository state and enable better concurrency control for git
repository scans.
After this PR, the `RepositoryEntry`s stored in worktree snapshots are
used only as a staging ground for local GitStores to pull from after
git-related events; non-local worktrees don't store them at all,
although this is not reflected in the types. GitTraversal and other
places that need information about repositories get it from the
GitStore. The GitStore also takes over handling of the new
UpdateRepository and RemoveRepository messages. However, repositories
are still discovered and scanned on a per-worktree basis, and we're
still identifying them by the (worktree-specific) project entry ID of
their working directory.
- [x] Remove WorkDirectory from RepositoryEntry
- [x] Remove worktree IDs from repository-related RPC messages
- [x] Handle UpdateRepository and RemoveRepository RPCs from the
GitStore
Release Notes:
- N/A
---------
Co-authored-by: Max <max@zed.dev>
Co-authored-by: Max Brunsfeld <maxbrunsfeld@gmail.com>
This PR fixes a bug where breakpoints would be rendered on incorrect
lines when openings a git hunk that contained breakpoints. This also
disables breakpoints from being shown in deleted git hunks as well.
Note: There's some unexpected behavior when using an anchor to get a
display point that is in an open git hunk, where the
`anchor.to_point().col == 0`.
```rust
let position = multi_buffer_anchor
.to_point(&multi_buffer_snapshot)
.to_display_point(&snapshot);
```
The above code will return a display point that is one line below where
the anchor actually represents when it's in an opened hunk diff. Which
causes the bug shown below
https://github.com/user-attachments/assets/bd15d02a-3cdc-4c8e-841f-bef238583351
@ConradIrwin Is this expected behavior when calling
`.to_display_point(&snapshot)`?
Release Notes:
- N/A
This change makes the git panel and project panel behave the same, on
Linux and macOS, and adds prompts.
Release Notes:
- Changed the git panel to prompt before restoring a file.
