Closes #52115

## Context

Removes the third party edit prediction integration for Sweep AI ahead
of their servers shutting down.

This PR will not affect those who are already use or plan to use their
open weighted models. The code removed by this PR was required for
integrating with their proprietary API.


## Self-Review Checklist

<!-- Check before requesting review: -->
- [x] I've reviewed my own diff for quality, security, and reliability
- [x] Unsafe blocks (if any) have justifying comments
- [x] The content is consistent with the [UI/UX
checklist](https://github.com/zed-industries/zed/blob/main/CONTRIBUTING.md#uiux-checklist)
- [x] Tests cover the new/changed behavior
- [x] Performance impact has been considered and is acceptable

Release Notes:

- Removed support for using the Sweep AI edit prediction provider
through their proprietary API, as the servers are shutting down
https://discord.com/channels/1100625416022138902/1100625417272045639/1480644297903575142,
https://x.com/wwzeng1/status/2033302698360180949

- zed-extensions/elixir#61 renames `HEEX` to `HEEx`; to prevent any
disruptions from that change, this adds the correct name alongside a
settings migration
- zed-extensions/elixir#101 adds support for `EEx` templates; this sets
the default language server to be the same as that of `Elixir` and
`HEEx`

This also adds a few more extensions that should be recognized as `HEEx`
files

Release Notes:

- N/A

---------

Co-authored-by: Finn Evers <finn@zed.dev>

After GPU device recovery clears the atlas, the next frame could
re-present a stale scene via the needs_present path (which skips scene
rebuilding). The stale scene references texture IDs that no longer exist
in the cleared atlas, causing an index-out-of-bounds panic.

Fix this by setting a force_render_after_recovery flag when device
recovery completes. The platform refresh loop reads this flag and passes
force_render: true in RequestFrameOptions, ensuring GPUI rebuilds the
scene before presenting.

Fixes ZED-5QT

Release Notes:

- N/A or Added/Fixed/Improved ...

This PR adds Git status badges next to file names in the Project Panel,
following my older PR #49802
These are enabled by having "git_status" true.

Screenshot
<img width="343" height="320" alt="image"
src="https://github.com/user-attachments/assets/b2c208bf-5027-4947-a5ee-eeb74fadb02b"
/>

I'd love to hear feedback about any of this :)
Especially feedback on these:

- File name colour is determined only by Git status, the diagnostic
badges remain separate. Should diagnostics also affect the filename
colour?
- (Unstaged) Modified files and staged modifications share the same
colour, in vscode staged modifications use a brownish colour by default
which I could not find the colours. I think differentiating them is
definetely something to add.

Release Notes

- Added git status indicators in Project Panel. It can be enabled by
setting `git_status_indicator` to `true` in `project_panel` settings.

---------

Co-authored-by: Smit Barmase <heysmitbarmase@gmail.com>

## Context

Follow up to #52388 
## How to Review

<!-- Help reviewers focus their attention:
- For small PRs: note what to focus on (e.g., "error handling in
foo.rs")
- For large PRs (>400 LOC): provide a guided tour — numbered list of
files/commits to read in order. (The `large-pr` label is applied
automatically.)
     - See the review process guidelines for comment conventions -->

## Self-Review Checklist

<!-- Check before requesting review: -->
- [x] I've reviewed my own diff for quality, security, and reliability
- [x] Unsafe blocks (if any) have justifying comments
- [x] The content is consistent with the [UI/UX
checklist](https://github.com/zed-industries/zed/blob/main/CONTRIBUTING.md#uiux-checklist)
- [x] Tests cover the new/changed behavior
- [x] Performance impact has been considered and is acceptable

Release Notes:

- N/A

## Context

Closes #52391

gopls sends semantic tokens for format args with [format] modifier.
Without a simple support for there is no special highlighting for format
args like `%s`. Its a simple change inside
`go/semantic_token_rules.json` that makes this a default behavior.

Currently (on 0.229 preview):
<img width="1136" height="675" alt="Screenshot 2026-03-25 at 11 34 57"
src="https://github.com/user-attachments/assets/cd3d9a52-6740-47b5-8f05-02514cb4f9df"
/>

After fix:

<img width="1136" height="675" alt="Screenshot 2026-03-25 at 11 58 37"
src="https://github.com/user-attachments/assets/5ec4277b-39c1-4fe1-bed5-8a6fa63c18c9"
/>


## Self-Review Checklist

<!-- Check before requesting review: -->
- [x] I've reviewed my own diff for quality, security, and reliability
- [x] Unsafe blocks (if any) have justifying comments
- [x] The content is consistent with the [UI/UX
checklist](https://github.com/zed-industries/zed/blob/main/CONTRIBUTING.md#uiux-checklist)
- [ ] Tests cover the new/changed behavior
- [x] Performance impact has been considered and is acceptable

Release Notes:

- gopls: Add semantic token rule for format string interpolations

Release Notes:

- N/A or Added/Fixed/Improved ...

## Context

We now batch multiple db operations together if they happen very
frequently (similar to how we handle it in `thread_history.rs`)

## Self-Review Checklist

<!-- Check before requesting review: -->
- [x] I've reviewed my own diff for quality, security, and reliability
- [x] Unsafe blocks (if any) have justifying comments
- [x] The content is consistent with the [UI/UX
checklist](https://github.com/zed-industries/zed/blob/main/CONTRIBUTING.md#uiux-checklist)
- [x] Tests cover the new/changed behavior
- [x] Performance impact has been considered and is acceptable

Release Notes:

- N/A

---------

Co-authored-by: Gaauwe Rombouts <mail@grombouts.nl>

Follow up to bzip support to clean up some of this code


## Self-Review Checklist

<!-- Check before requesting review: -->
- [x] I've reviewed my own diff for quality, security, and reliability
- [x] Unsafe blocks (if any) have justifying comments
- [x] The content is consistent with the [UI/UX
checklist](https://github.com/zed-industries/zed/blob/main/CONTRIBUTING.md#uiux-checklist)
- [x] Tests cover the new/changed behavior
- [x] Performance impact has been considered and is acceptable

Release Notes:

- N/A

## Context

Follow up to #52193

We changed this code in #52193, but moving the thread title read outside
of the spawn, causes a loop of title updated events

## Self-Review Checklist

<!-- Check before requesting review: -->
- [x] I've reviewed my own diff for quality, security, and reliability
- [x] Unsafe blocks (if any) have justifying comments
- [x] The content is consistent with the [UI/UX
checklist](https://github.com/zed-industries/zed/blob/main/CONTRIBUTING.md#uiux-checklist)
- [ ] Tests cover the new/changed behavior
- [x] Performance impact has been considered and is acceptable

Release Notes:

- N/A

## Summary

- Added `TarBz2` variant to `AssetKind` enum for `.tar.bz2` / `.tbz2`
archives
- Implemented `extract_tar_bz2` using the `bzip2` feature of
`async-compression` (already a workspace dependency, just enabled the
feature flag)
- Wired up both streaming and file-based extraction paths in
`github_download.rs`
- Added `.tar.bz2` / `.tbz2` URL detection in both
`LocalExtensionArchiveAgent` and `LocalRegistryArchiveAgent`

This unblocks ACP registry entries (like Goose) that only ship
`.tar.bz2` archives.

Reference: https://github.com/block/goose/issues/8047

## Test plan

- [ ] Verify `cargo check` and `clippy` pass (confirmed locally)
- [ ] Test downloading an ACP agent that ships a `.tar.bz2` archive
(e.g., Goose)
- [ ] Verify existing `.tar.gz` and `.zip` agent downloads still work

Release Notes:

- Added support for `.tar.bz2` archives in ACP agent server downloads,
unblocking registry entries like Goose that only ship bzip2-compressed
tarballs.

---------

Co-authored-by: Ben Brandt <benjamin.j.brandt@gmail.com>

Needs some more work.

Release Notes:

- N/A

This PR moves the `CompletionIntent` enum from the `cloud_llm_client`
crate to the `language_model` crate, as it is no longer part of the
Cloud interface.

Release Notes:

- N/A

Fixes a double borrow panic when resizing the bottom dock, introduced in
#52276

Release Notes:

- N/A

Release Notes:

- Fix copy for some Wayland users

This PR simplifies the flexible-width agent panel's rendering such that
for now, it will not grow to fill the center area when there are no
files open. The problem with doing that is that it makes the bottom dock
invisible.

We could potentially revisit this, and just shrink the width of the
agent panel when the bottom dock was toggled, but that might feel
unstable, so for now I'm doing a simpler thing and just keeping the
center area visible even when it is empty.

Release Notes:

- N/A

Release Notes:

- Fixed Git Graph button flicker in Git panel

Release Notes:

- Fixed window title displaying garbled characters (e.g. "Zed â██
Settings") on Linux X11 due to missing UTF-8 encoding for the
`_NET_WM_NAME` property.

Release Notes:

- Fixed 1-2 seconds of audio silence when unmuting for the first time
during a call with Bluetooth headphones when `mute_on_join` is enabled.

## Context

Since https://github.com/zed-industries/zed/pull/52232/, there have been
a lot of problems related to panels randomly hiding and showing
themselves. However, the code in that PR is pretty innocuous. It
shouldn't be having that kind of effect. However, it turns out there's a
bug in how GPUI subscriptions work that caused there to be orphaned
observe callbacks firing spuriously. The core sequence is captured in
`test_unsubscribe_during_callback_with_insert`, essentially you have to
have two observers on the same item (`SettingsStore`, in this case),
that both drop themselves. If the first observer _adds_ a callback onto
the _same item_, then the second observer will never successfully drop
itself.

While in there, I also fixed an unrelated bug that @maxbrunsfeld noticed
where if you have two callbacks on the same item, and an earlier
callback drops a later one, then the second callback will spuriously
fire.

I also added a few extra smoke tests to the subscription code, and a
test capturing the observed bug at the workspace level.

## Self-Review Checklist

<!-- Check before requesting review: -->
- [x] I've reviewed my own diff for quality, security, and reliability
- [x] Unsafe blocks (if any) have justifying comments
- [x] The content is consistent with the [UI/UX
checklist](https://github.com/zed-industries/zed/blob/main/CONTRIBUTING.md#uiux-checklist)
- [x] Tests cover the new/changed behavior
- [x] Performance impact has been considered and is acceptable

Release Notes:

- N/A

---------

Co-authored-by: max <max@zed.dev>

Many editors such as vim and emacs support "modelines", a comment at the
beginning of the file that allows the file type to be explicitly
specified along with per-file specific settings

- The amount of configurations, style and settings mapping cannot be
handled in one go, so this opens up a lot of potential improvements.
- I left out the possiblity to have "zed" specific modelines for now,
but this could be potentially interesting.
- Mapping the mode or filetype to zed language names isn't obvious
either. We may want to make it configurable.

This is my first contribution to zed, be kind. I struggled a bit to find
the right place to add those settings. I use a similar approach as done
with editorconfig (merge_with_editorconfig). There might be better ways.

Closes #4762

Release Notes:

- Add basic emacs/vim modeline support.

Supersedes #41899, changes:
- limit reading to the first and last 1kb
- add documentation
- more variables handled
- add Arc around ModelineSettings to avoid extra cloning
- changed the way mode -> language mapping is done, thanks to
`modeline_aliases` language config
- drop vim ex: support
- made "Local Variables:" handling a separate commit, so we can drop it
easily
- various code style improvements

---------

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Co-authored-by: Claude <noreply@anthropic.com>
Co-authored-by: Conrad Irwin <conrad.irwin@gmail.com>

MetalAtlas::remove() used tiles_by_key.get(key) to look up the texture
ID but only called tiles_by_key.remove(key) when the texture became
fully unreferenced. When a tile was removed while other tiles remained
on the same texture, the key stayed in tiles_by_key as a stale entry.

Once subsequent removals deleted the texture, get_or_insert_with could
return the stale tile referencing a now-deleted texture slot, causing an
unwrap panic in MetalAtlasState::texture().

Fix: change .get(key) to .remove(key) unconditionally, matching the WGPU
and DirectX atlas implementations which already do this correctly.

Closes ZED-5KV

Release Notes:

- N/A or Added/Fixed/Improved ...

Closes #43545

Remove existing entries for an item before adding new ones to the
navigation history. This prevents Go Back (Ctrl-O) from bouncing between
the same items repeatedly.

Navigation pattern A->B->A->B->C now creates history [A,B,C] instead of
[A,B,A,B,C], making backward navigation more efficient.

Release Notes:

- N/A

## Fixes
Closes #28174 
Closes  #30393
Closes #36610
Closes #38616

Updates #31819
Updates #21136

## Context

There was an issue where CJK input could not be entered correctly when
certain keys used during IME composition conflicted with custom
keybindings while the CJK IME was enabled.

(In particular, in Vim mode, the `jj` keybinding in INSERT MODE could
not be entered correctly.)

This issue had already been fixed on Windows, but not on macOS (and
judging from the code, it is likely also present on Linux, although this
has not been verified).

In this change, we aligned the behavior with Windows.

Currently, on Windows, when the CJK ( or other ) IME is enabled, IME
input is prioritized.
Following this behavior, we updated macOS to also prioritize IME input.

| Input source | Source type | ASCII-capable | Result | Behavior |
|---|---|---|---|---|
| Japanese Hiragana | `KeyboardInputMode` | `false` | `true` | IME
handles key first |
| Japanese Romaji | `KeyboardInputMode` | `true` | `false` | Keybinding
matching first |
| Korean | `KeyboardInputMode` | `false` | `true` | IME handles key
first |
| Chinese Pinyin | `KeyboardInputMode` | `false` | `true` | IME handles
key first |
| Armenian | `KeyboardLayout` | `false` | `false` | Keybinding matching
first |
| Ukrainian | `KeyboardLayout` | `false` | `false` | Keybinding matching
first |
| English (ABC) | `KeyboardLayout` | `true` | `false` | Keybinding
matching first |


(Since similar approaches are used in tools like Vim and Neovim, we
consider this policy to be appropriate.)



## Videos
### before fix

https://github.com/user-attachments/assets/9b204220-d0de-4819-8dc4-00ba85169ef6

### after fix


https://github.com/user-attachments/assets/045cc808-8d29-42d2-9de5-f903fcb602db


## How to Review

This change modifies the gpui layer rather than Zed itself, so please
verify the behavior at the gpui level, not just within Zed.

## Self-Review Checklist

<!-- Check before requesting review: -->
- [x] I've reviewed my own diff for quality, security, and reliability
- [x] Unsafe blocks (if any) have justifying comments
- [x] The content is consistent with the [UI/UX
checklist](https://github.com/zed-industries/zed/blob/main/CONTRIBUTING.md#uiux-checklist)
- [x] Tests cover the new/changed behavior
- [x] Performance impact has been considered and is acceptable

Release Notes:

- Fixed IME composition on macOS in Vim insert mode with multi-stroke keybindings like `jj`

Fixes a crash introduced by removing some defensive code in
https://github.com/zed-industries/zed/pull/52276

Release Notes:

- N/A

Co-authored-by: Mikayla Maki <mikayla.c.maki@gmail.com>

## Context

We currently mark the current prediction as "shown" when the subtle mode
indicator appears. This doesn't really make sense. We make an effort
elsewhere to differentiate between a user rejecting a prediction that
they viewed vs. never saw in order to compute accurate metrics.

## Self-Review Checklist

<!-- Check before requesting review: -->
- [x] I've reviewed my own diff for quality, security, and reliability
- [x] Unsafe blocks (if any) have justifying comments
- [x] The content is consistent with the [UI/UX
checklist](https://github.com/zed-industries/zed/blob/main/CONTRIBUTING.md#uiux-checklist)
- [x] Tests cover the new/changed behavior
- [x] Performance impact has been considered and is acceptable

Release Notes:

- N/A or Added/Fixed/Improved ...

Closes #21822

## Context

Exposes `preferred_line_length` from `AllLanguageSettings` to the
Extension
API's `LanguageSettings` struct. Currently only `tab_size` is available
to
extensions, which prevents language extensions (e.g. Dart) from reading
the
user's preferred line length and forwarding it to their language server
(e.g. as `dart.lineLength`).

Related: https://github.com/zed-extensions/dart/issues/2

## How to Review

Small change — follow how `tab_size` is plumbed through:
1. WIT definition (`language-settings` record)
2. `extension_api` Rust struct
3. Host-side bridge conversion

The new field follows the exact same pattern.

## Self-Review Checklist

<!-- Check before requesting review: -->
- [x] I've reviewed my own diff for quality, security, and reliability
- [ ] Unsafe blocks (if any) have justifying comments
- [x] The content is consistent with the [UI/UX
checklist](https://github.com/zed-industries/zed/blob/main/CONTRIBUTING.md#uiux-checklist)
- [x] Tests cover the new/changed behavior
  - Compile-time verified via WIT bindings; no runtime behavior change
- [x] Performance impact has been considered and is acceptable

Release Notes:

- N/A

This PR adds support to paste external files and directories in Agent
Panel, along with the existing image paste path.

  Release Notes:

  - Added support for pasting files and folders into the Agent Panel.

Follow up to https://github.com/zed-industries/zed/pull/52343

We'll be experimenting showing the org name in the avatar trigger button
when that's not the personal one.

Release Notes:

- N/A

---------

Co-authored-by: Marshall Bowers <git@maxdeviant.com>

`PathList` stores both sorted paths and the original insertion order
(for display in the project panel). Previously, `PartialEq`, `Eq`, and
`Hash` were derived, which meant two `PathList` values with the same
paths but different display orderings were considered unequal.

This change replaces the derived impls with manual ones that only
compare
the sorted `paths` field, matching the semantic intent: a `PathList`
identifies a set of directories, and the display order is not part of
that identity.

Release Notes:

- N/A

Follow-up to the discussion in #51670.

@ConradIrwin, thanks for the review!

You're right about the other editors. VS Code and JetBrains use
dedicated shortcuts for bottom panel tabs, which exactly mirrors how
`ctrl-`` already behaves in Zed, so we don't need a new keybinding.

The reason I originally proposed `cmd-shift-j` was just to match the UI
pattern of the Debugger, which clearly shows `cmd-shift-d` on hover.
Because of a minor UI discoverability bug, the Terminal button's tooltip
was completely blank, which made it look like it just lacked a shortcut
entirely.

It turns out the Terminal button's internal `toggle_action()` method was
just telling the UI to look up `ToggleFocus` instead of
`terminal_panel::Toggle` (which is what `ctrl-`` is actually bound to).

I've updated this branch with a simple one-word fix that just points the
`toggle_action` to the correct keymap. The existing `ctrl-`` shortcut
now shows up perfectly in the hover tooltip. Good catch on the existing
shortcuts, keeps this much cleaner.

Co-authored-by: Michael Smolkin <msmolkin@users.noreply.github.com>

## Context

Ensure subagent threads build requests with the Subagent intent instead
of UserPrompt. This allows us to properly
attribute this as a tool call for certain providers instead of a user
request.

## Self-Review Checklist

<!-- Check before requesting review: -->
- [x] I've reviewed my own diff for quality, security, and reliability
- [x] Unsafe blocks (if any) have justifying comments
- [x] The content is consistent with the [UI/UX
checklist](https://github.com/zed-industries/zed/blob/main/CONTRIBUTING.md#uiux-checklist)
- [x] Tests cover the new/changed behavior
- [x] Performance impact has been considered and is acceptable

Release Notes:

- copilot_chat: Fix subagent requests being marked as user requests.

## Context

- Updates docs for how to update edit prediction bindings
- Walks back how often we use tab to accept prediction where we
previously didn't; now just when in leading whitespace, tab does not
accept completion when completions menu is open
- migration for keymaps using the old `edit_prediction_conflict` key
context

## How to Review

The only code worth reviewing is the migration
<!-- Help reviewers focus their attention:
- For small PRs: note what to focus on (e.g., "error handling in
foo.rs")
- For large PRs (>400 LOC): provide a guided tour — numbered list of
files/commits to read in order. (The `large-pr` label is applied
automatically.)
     - See the review process guidelines for comment conventions -->

## Self-Review Checklist

<!-- Check before requesting review: -->
- [x] I've reviewed my own diff for quality, security, and reliability
- [x] Unsafe blocks (if any) have justifying comments
- [x] The content is consistent with the [UI/UX
checklist](https://github.com/zed-industries/zed/blob/main/CONTRIBUTING.md#uiux-checklist)
- [x] Tests cover the new/changed behavior
- [x] Performance impact has been considered and is acceptable

Release Notes:

- N/A or Added/Fixed/Improved ...

---------

Co-authored-by: Max <max@zed.dev>

Release Notes:

- N/A

This PR fixes test flakiness introduced in
https://github.com/zed-industries/zed/pull/52111

The flakiness was because the test was using its own RNG, rather than
using gpui's built-in support for consistently-seeded RNGs in tests. The
fix was just to adjust the test to use the same logic for computing row
and column count as was introduced in the above PR.

Release Notes:

- N/A

Related to #31504

## Problem

If you drag a terminal into the center editor area as a tab, pressing
`Cmd+N` still opens the new terminal in the terminal panel instead of
next to the tab you're looking at.

## Solution

`Cmd+N` now checks whether the center pane is focused and its active tab
is a terminal. If both are true, the new terminal opens there. All other
behavior is unchanged.



https://github.com/user-attachments/assets/01036cb4-5f1f-456e-900e-eed2e53d61f5


## Test plan

- [ ] Drag a terminal into the center editor area. Focus it. Press
Cmd+N. New terminal opens as a center tab.
- [ ] Toggle terminal panel with Ctrl+`. Press Cmd+N. Opens in panel as
before.
- [ ] Focus a regular code file. Press Cmd+N. Opens in panel as before.
- [ ] Have a terminal in the center, but focus the terminal panel. Press
Cmd+N. Opens in panel as before.

Release Notes:

- Improved `terminal: New Terminal` (Cmd+N) to open in the center pane
when a terminal tab there is focused, instead of always opening in the
terminal panel.

This PR adds an `is_personal` field to the `Organization` DTO returned
from Cloud.

We're already returning this field in production.

Part of CLO-569.

Release Notes:

- N/A

Release Notes:

- N/A

Setting up the infra for unit evals on the new tool. For now,
replicating the eval setup as close as possible to the old tool.


## Self-Review Checklist

<!-- Check before requesting review: -->
- [x] I've reviewed my own diff for quality, security, and reliability
- [x] Unsafe blocks (if any) have justifying comments
- [x] The content is consistent with the [UI/UX
checklist](https://github.com/zed-industries/zed/blob/main/CONTRIBUTING.md#uiux-checklist)
- [x] Tests cover the new/changed behavior
- [x] Performance impact has been considered and is acceptable

Release Notes:

- N/A

* Update `Vim::register` to correctly re-activate when either `vim_mode`
  or `helix_mode` are enabled, when they were previously disabled.
* Update `settings_ui::page_data::keymap_page` such that, when enabling
  or disabling `vim_mode` or `helix_mode`, the other setting is
  disabled, in case it is set. This ensures that only one mode is
  enabled at a time, as it should be.
  
Closes #51704 

Release Notes:

- Fixed switching between vim and helix mode needing multiple calls to
  have an effect while an editor is already open
- Update Settings UI such that, enabling vim or helix mode, now disables
  the other if it was previously active

---------

Co-authored-by: dino <dinojoaocosta@gmail.com>

Release Notes:

- The agent panel now has a flexible width, similar to the center panes
of the workspace.

Closes CLO-339

This PR moves the organization menu items inside the user menu, instead
of a standalone button within the title bar. We did it so because 1)
we're not fully sure how useful _always_ seeing in which org you are in
will be, and 2) seeing in which org you're in and seeing in which _plan_
you are in are very similar in terms of use case, and we don't currently
display the plan in the title bar... thus, that served as argument for
the same level of visibility.

---

- [x] I've reviewed my own diff for quality, security, and reliability
- [x] Unsafe blocks (if any) have justifying comments
- [x] The content is consistent with the [UI/UX
checklist](https://github.com/zed-industries/zed/blob/main/CONTRIBUTING.md#uiux-checklist)
- [x] Tests cover the new/changed behavior
- [x] Performance impact has been considered and is acceptable

Release Notes:

- N/A

---------

Co-authored-by: Gaauwe Rombouts <mail@grombouts.nl>

## Context

This fixes a bug where switching to a worktree workspace sometimes
briefly flashes it as a top-level project in the sidebar. This happened
because the sidebar would rebuild its entries whenever the multi
workspace emits the added workspace event.

The fix is checking from the root workspace git linked worktree list as
well to see if newly added workspaces are linked.

## Self-Review Checklist

<!-- Check before requesting review: -->
- [x] I've reviewed my own diff for quality, security, and reliability
- [x] Unsafe blocks (if any) have justifying comments
- [x] The content is consistent with the [UI/UX
checklist](https://github.com/zed-industries/zed/blob/main/CONTRIBUTING.md#uiux-checklist)
- [x] Tests cover the new/changed behavior
- [x] Performance impact has been considered and is acceptable

Release Notes:

- N/A

## Context

We're seeing issues where discarding a prediction (e.g. `editor::Cancel`
while a prediction is being shown) seems to just hide it temporarily.
This is becoming a larger issue with #51842 as we now expect people to
be dismissing predictions in order to insert literal tabs in certain
contexts.
The logic changed in this PR is part of the problem, but the model
generating the same prediction multiple times is also likely
contributing. That will be solved as-needed later.


## Self-Review Checklist

<!-- Check before requesting review: -->
- [x] I've reviewed my own diff for quality, security, and reliability
- [x] Unsafe blocks (if any) have justifying comments
- [x] The content is consistent with the [UI/UX
checklist](https://github.com/zed-industries/zed/blob/main/CONTRIBUTING.md#uiux-checklist)
- [x] Tests cover the new/changed behavior
- [x] Performance impact has been considered and is acceptable

Release Notes:

- Fixed an issue where dismissing an edit prediction would not fully
discard it, causing it to re-appear

Increases timeouts, simplifies queries, and makes other changes needed to
pull data more efficiently.

Release Notes:

- N/A

---------

Co-authored-by: Ben Kunkle <ben@zed.dev>

CLOSES: #50002 

When using indent-based folding (the default, `document_folding_ranges =
Off`),
collapsed folds in brace-delimited languages displayed the closing
delimiter on
a separate line:

fn b() {⋯
}

This extends the fold range in `crease_for_buffer_row` to include the
trailing
newline and leading whitespace before a closing `}`, `)`, or `]`,
producing
the expected single-line display:

fn b() {⋯}

Whitespace-sensitive languages like Python are unaffected — their
terminating
lines don't start with a closing delimiter, so the existing behavior is
preserved.

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)

Screenshots:
Before:
<img width="676" height="584" alt="image"
src="https://github.com/user-attachments/assets/0779233b-e287-495a-bab4-d2a97829e1c0"
/>


After:
<img width="1498" height="824" alt="image"
src="https://github.com/user-attachments/assets/ec6b4cb0-dac1-4db0-beed-38131a27b5c8"
/>



Release Notes:

- Fixed indent-based code folding to display the closing delimiter (`}`,
`)`, `]`) on the same line as the fold placeholder instead of on a
separate line
([#50002](https://github.com/zed-industries/zed/issues/50002)).

Closes #47007 

After #50974, symlinked `settings.json` can be opened and saved, but the
worktree ends up rooted at the file itself rather than its parent
directory, which causes this error to be logged when changing settings
via the settings GUI:

```
2026-03-18T22:24:05+01:00 ERROR [worktree] ignoring event
"C:\\Users\\Bryan\\dev\\bryanpth\\dotfiles\\config\\zed\\settings.json~RF1ab7ae0.TMP" outside of root path
"C:\\Users\\Bryan\\dev\\bryanpth\\dotfiles\\config\\zed\\settings.json"
```

Fixed by canonicalizing the symlink path before passing to `open_paths`.
Tested with a symlinked settings.json, both directly and through WSL.

Release Notes:
- Fixed incorrect worktree root when settings.json is a symlink on
Windows

Fixes #51733

The previous fix in #32590 only filtered `.vscode` tasks at the modal
presentation layer, so they still appeared in other code paths and in
the used-tasks list. This moves the filtering into
`worktree_scenarios()` in the task inventory, so all consumers see the
correct behavior.

Changes:
- `InventoryFor::worktree_scenarios()` now skips `.vscode` entries when
`.zed` entries exist for the same worktree
- `used_and_current_resolved_tasks()` filters previously spawned
`.vscode` tasks from the used-tasks list
- Removes the now-redundant `.vscode` filter from the task picker modal

Release Notes:

- Fixed .vscode/tasks.json still being used when .zed/tasks.json is
present

Co-authored-by: moktamd <moktamd@users.noreply.github.com>

## Context

Sometimes the model likes to respond with `"[...]"` for the edits array,
instead of just `[...]`, which causes deserialisation to fail:
```
Failed to receive tool input: invalid type: string "[{"old_text": ..., "new_text": ...}]", expected a sequence
```

## How to Review

<!-- Help reviewers focus their attention:
- For small PRs: note what to focus on (e.g., "error handling in
foo.rs")
- For large PRs (>400 LOC): provide a guided tour — numbered list of
files/commits to read in order. (The `large-pr` label is applied
automatically.)
     - See the review process guidelines for comment conventions -->

## Self-Review Checklist

<!-- Check before requesting review: -->
- [x] I've reviewed my own diff for quality, security, and reliability
- [x] Unsafe blocks (if any) have justifying comments
- [x] The content is consistent with the [UI/UX
checklist](https://github.com/zed-industries/zed/blob/main/CONTRIBUTING.md#uiux-checklist)
- [x] Tests cover the new/changed behavior
- [x] Performance impact has been considered and is acceptable

Release Notes:

- N/A

---------

Co-authored-by: Gaauwe Rombouts <mail@grombouts.nl>