This PR greatly improves our handling of remote threads in the sidebar.
One primary issue was that many parts of the sidebar were only looking
at a thread's path list and not its remote connection information. The
fix here is to use `ProjectGroupKey` more consistently throughout the
sidebar which also includes remote connection information.
The second major change is to extend the MultiWorkspace with the ability
to initiate the creation of remote workspaces when needed. This involved
refactoring a lot of our remote workspace creation paths to share a
single code path for better consistency.
Release Notes:
- (Preview only) Fixed remote project threads appearing as a separate
local project in the sidebar
---------
Co-authored-by: Anthony Eid <anthony@zed.dev>
Co-authored-by: Anthony Eid <hello@anthonyeid.me>
Co-authored-by: Max Brunsfeld <maxbrunsfeld@gmail.com>
Remove the unused `ConversationView::delete_history_entry` method and
its now-unused `ThreadMetadataStore` import. The method had zero callers
— the same functionality is covered by
`ThreadHistoryView::remove_thread` and
`ThreadsArchiveView::delete_thread`.
Release Notes:
- N/A
Don't move the branch when making WIP commits during archiving; instead
make detached commits via `write-tree` + `commit-tree`.
(No release notes because this isn't stable yet.)
Release Notes:
- N/A
---------
Co-authored-by: Anthony Eid <anthony@zed.dev>
Connect the git API and archived worktree data model to the sidebar's
archive/unarchive flow:
- Add `thread_worktree_archive` module: orchestrates the full archive
cycle (WIP commits, DB records, git refs, worktree deletion) and restore
cycle (detached worktree creation, reset to recover staged/unstaged
state, branch restoration)
- Integrate into sidebar: `archive_thread` now persists worktree state
before cleanup; `activate_archived_thread` restores worktrees via git
with targeted path replacement for multi-root threads
- Add pending worktree restore UI state with spinner and cancel button
- Show toast on restore failure instead of silent log
- Switch to surviving workspace when archiving active thread whose
workspace will be deleted
- Deserialize persisted `project_group_keys` on window restore
- Guard `cleanup_empty_workspaces` against dropped entities
- Await rollback DB operations instead of fire-and-forget
Part 3 of 3 in the persist-worktree stack. Stacked on #53214. This wires
up parts 1 and 2.
Release Notes:
- Added worktree state preservation when archiving agent threads,
automatically restoring staged and unstaged changes when the thread is
unarchived
---------
Co-authored-by: Anthony Eid <anthony@zed.dev>
When toggling fast mode, the setting is now written to `settings.json`
under `agent.default_model.speed`, so new threads start with the same
speed. This follows the same pattern as `enable_thinking` and `effort`.
The `speed` field uses the existing `Speed` enum (`"fast"` /
`"standard"`) rather than a boolean, to leave room for future speed
tiers.
Example settings:
```json
{
"agent": {
"default_model": {
"provider": "zed.dev",
"model": "claude-sonnet-4",
"speed": "fast"
}
}
}
```
cc @benbrandt
Release Notes:
- N/A
---------
Co-authored-by: Ben Brandt <benjamin.j.brandt@gmail.com>
Self-Review Checklist:
- [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:
- (preview only) sidebar: Fixed issue where tool confirmation indicator
would not show up when subagent asks for permissions
We were incorrectly wrapping new terminal auth methods in double ssh
calls.
Only affected ACP beta users, but important for testing and stabilizing
the feature.
We moved the ssh wrapping to be only added in the acp process creation
where it was needed.
Self-Review Checklist:
- [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: Bennet Bo Fenner <bennetbo@gmx.de>
Follow up to https://github.com/zed-industries/zed/pull/53349 as I was
still seeing cases where content would overlap with buttons.
<img width="500" height="266" alt="Screenshot 2026-04-08 at 6 53@2x"
src="https://github.com/user-attachments/assets/e58742d4-a736-4509-9a29-35c135ddb980"
/>
I believe this PR makes the whole message editor more solid against
these cases because we're relying on `flex_basis` now to set the
max-width, which runs before we do any editor mode calculations. Tested
this a bunch and it seems to be pretty stable.
Release Notes:
- N/A
We now always show a notification for a thread if it is not the active
one (even if the sidebar is open).
Also made clicking on the notification actually open the corresponding
thread (previously we would just changed the workspace)
Self-Review Checklist:
- [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
- added a shared helper `take_active_draft_initial_content` to resolve
initial content
- `draft_prompt` is copied over to a new thread and is persisted in old
thread with this behavior
Self-Review Checklist:
- [x] I've reviewed my own diff for quality, security, and reliability
- [x] no Unsafe blocks
- [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
- [ ] Performance impact has been considered and is acceptable
Closes#38028
Release Notes:
- persist draft messages in ai thread when a new thread is created so we
dont use any messages or draft prompts mistakenly
Question:
need some guidance on how to persist a thread when the `draft_prompt` is
present but the message is unsent because currently having a draft
prompt in the agent input box and thenmaking a new agent thread copies
the prompt over to the new thread(changes from this pr) but on going
back to the old thread new thread is destroyed
This PR adjusts how project group removal works in the sidebar to be
project group aware. Removal is now a batched operation, and supports a
contextually appropriate fallback active workspace fallback. In the
process of removing this, I had to chase down and fix other broken code:
- Removes the "move to new window" actions
- Changed the sidebar to store things in "most recently added" order
- The recent project UI was still using workspaces, instead of project
groups
- Adjusted the archive command to use the new APIs to remove the
workspace associated with that worktree (cc: @rtfeldman)
- The property tests where still using indexes and workspaces instead of
project groups
Self-Review Checklist:
- [ ] 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
Closes #ISSUE
Release Notes:
- N/A
After a user creates successfully creates a new git worktree we reset
the agent thread picker state to avoid accidentally creating a new
worktree for a thread.
This also fixes some bugs where a past agent thread picker state was
invalid because it became out of sync after a new worktree was created,
or when a mutating git operation occurred.
Self-Review Checklist:
- [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 ...
Quick follow up to https://github.com/zed-industries/zed/pull/52730 with
some adjustments. Ended up cleaning some unused UI elements already
(i.e., the "Recent" history items in the agent panel's empty state).
Release Notes:
- N/A
This PR adjust the name of the worktree within the agent panel: we
display `main` if you're in the main worktree and then the worktree name
when in a linked one. Additionally, I also took the opportunity to fix
something that confused me in the general worktree picker: for the list
item that represented the main worktree, we were using the current
checked out branch name for that item's name. I think that's a bit
incorrect, given the _name_ of that worktree isn't the name of the
branch you have active at that moment. For now, I'm using simply `main`
as the worktree name. Lastly, also added a little blue check icon in the
worktree _and branch_ picker to highlight which worktree/branch is
currently active.
| main worktree | linked worktree | main worktree in picker |
|--------|--------|--------|
| <img width="802" height="338" alt="Screenshot 2026-04-07 at 10 32
2@2x"
src="https://github.com/user-attachments/assets/aad993e6-adc3-4c8d-b01a-abb2597182e7"
/> | <img width="804" height="332" alt="Screenshot 2026-04-07 at 10
32@2x"
src="https://github.com/user-attachments/assets/e59c5b86-8cee-476c-ab7b-32ae87249c1a"
/> | <img width="804" height="856" alt="Screenshot 2026-04-07 at 10
36@2x"
src="https://github.com/user-attachments/assets/282d89a8-570e-4e62-b703-853bed9b21ff"
/> |
Release Notes:
- Git: Fixed display of the main worktree name in the worktree picker.
### Summary
Keep the prompt editor’s max-width centering, but move the footer
controls out of the same flexible wrapper as the full-height editor and
add `min_h_0()` shrink guards so the editor no longer overlaps the
footer in zoomed/full-screen mode.
#### Before
<img width="1624" height="1061" alt="Screenshot 2026-04-07 at 8 39
03 PM"
src="https://github.com/user-attachments/assets/db889fcd-af8a-4297-a1f9-0acb84e6ead3"
/>
#### After
<img width="1624" height="976" alt="Screenshot 2026-04-07 at 9 14 52 PM"
src="https://github.com/user-attachments/assets/ddae7fd4-e720-432d-b84d-45f667aa7dce"
/>
Self-Review Checklist:
- [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
Closes #ISSUE
Release Notes:
- N/A or Added/Fixed/Improved ...
Self-Review Checklist:
- [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
Closes #ISSUE
Release Notes:
- N/A or Added/Fixed/Improved ...
It's happening.
Self-Review Checklist:
- [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
Closes#39532
## Summary
Move the long-running agent "generating" spinner into a dedicated
`GeneratingSpinner` view.
Previously, the generating state rendered `SpinnerLabel` inline as part
of the full thread view. This change keeps the same spinner styling, but
isolates it into its own small view so the animation no longer needs to
live directly inside the larger thread UI subtree while the agent is
running.
before:
<img width="398" height="448" alt="image"
src="https://github.com/user-attachments/assets/8a56c394-f7a6-4835-b877-fbc1500d2d03"
/>
after:
<img width="524" height="597" alt="image"
src="https://github.com/user-attachments/assets/bd2e5804-dba8-49c0-a421-7dc5ba600348"
/>
## Testing
- `cargo run -j 4 ~/Downloads/WorkSpace/zed/`
- `cargo check -p agent_ui`
Release Notes:
- Fixed high GPU usage from the agent panel's generating spinner while
an agent is running.
---------
Co-authored-by: Anthony Eid <anthony@zed.dev>
Add the persistence layer for tracking archived git worktrees:
- `ArchivedGitWorktree` struct with `staged_commit_hash` and
`unstaged_commit_hash` fields to precisely identify WIP commits
- DB migrations for `archived_git_worktrees` and
`thread_archived_worktrees` (join table) tables
- CRUD operations: create, link to thread, query by thread, delete, mark
as restored
- `Column` impl for deserializing `ArchivedGitWorktree` from DB rows
- Tests for two-SHA round-trip, create/retrieve, delete, restore flag,
multi-thread linking, and multiple worktrees per thread
Part 2 of 3 in the persist-worktree stack. Stacked on #53213. This is
pure data model — nothing is wired up yet.
Release Notes:
- N/A
---------
Co-authored-by: Anthony Eid <anthony@zed.dev>
When a project mixes main repo worktrees (e.g. `extensions`) with linked
git worktrees (e.g. `zed4`, a linked worktree of `zed`), two things were
broken:
1. **Threads appeared under the wrong sidebar heading.** A thread
created in a workspace with both `extensions` and `zed4` would show
under the "zed" group instead of the "extensions, zed" group. This
happened because `main_worktree_paths` in `ThreadMetadata` was only
populated with paths from linked worktrees, omitting regular repos
entirely. The fix uses `project_group_key()` to normalize all visible
worktrees to their main repo paths — the same normalization the sidebar
uses for group headers.
2. **The worktree chip tooltip only showed linked worktree paths**,
missing main repo paths like `extensions`. This happened because
`worktree_info_from_thread_paths` filtered out main worktree paths
entirely. The fix introduces a `WorktreeKind` enum (`Main` / `Linked`)
on `ThreadItemWorktreeInfo`, so all worktrees are included in the data
model. Chips still only render for `Linked` worktrees (main worktrees
are redundant with the group header), but the tooltip now shows all
paths.
Release Notes:
- N/A
After chat functionality was removed, this panel became redundant. It
only displayed three notification types: incoming contact requests,
accepted contact requests, and channel invitations.
This PR moves those notifications into the collab experience by adding
toast popups and a badge count to the collab panel. It also removes the
notification-panel-specific settings, documentation, and Vim command.
Before you mark this PR as ready for review, make sure that you have:
- [ ] 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:
- Removed the notification panel from Zed
This PR improves the recent projects picker in the context of
multi-workspace:
- The currently active project now appears in the "This Window" section
with a checkmark indicator. Clicking it simply dismisses the picker,
since there's nothing to switch to. This feels like a better UX because
it gives you visual confirmation of where you are.
- The remove button is hidden for the current project entry, both in the
row and the footer, to prevent accidentally removing the workspace
you're actively using.
- The "Add to Workspace" button now uses a more descriptive icon
(`FolderOpenAdd`) and shows a meta tooltip clarifying that it adds the
project as a multi-root folder project.
The primary click/enter behavior remains unchanged—it opens the selected
project in the current window's multi-workspace. The "Open in New
Window" action continues to be available via the icon button or
shift+enter.
Release Notes:
- Improved the recent projects picker to show the currently active
project in the "This Window" section with a checkmark indicator.
- `language_model` no longer depends on provider-specific crates such as
`anthropic` and `open_ai` (inverted dependency)
- `language_model_core` was extracted from `language_model` which
contains the types for the provider-specific crates to convert to/from.
- `gpui::SharedString` has been extracted into its own crate (still
exposed by `gpui`), so `language_model_core` and provider API crates
don't have to depend on `gpui`.
- Removes some unnecessary `&'static str` | `SharedString` -> `String`
-> `SharedString` conversions across the codebase.
- Extracts the core logic of the cloud `LanguageModelProvider` into its
own crate with simpler dependencies.
Release Notes:
- N/A
---------
Co-authored-by: John Tur <john-tur@outlook.com>
This PR adds a configurable max-width to the agent panel. This will be
particularly useful when opting into an agentic-first layout where the
thread will be at the center of the UI (with the panel most likely
full-screen'ed, which is why I'm also adding here the button to make it
full screen in the toolbar). The default max-width is 850, which is a
bit bigger than the one generally considered as a standard (~66
characters wide, which usually sums up to 750 pixels).
Release Notes:
- Agent: Added a max-width to the thread view for better readability,
particularly when the panel is zoomed in.
This PR allows users to create a new thread based off a git worktree
that already exists or has a custom name. User's can also choose what
branch they want the newly generated worktree to be based off of.
The UI still needs some polish, but I'm merging this early to get the
team using this before our preview launch. I'll be active today and
tomorrow before launch to fix any nits we have with the UI.
Functionality of this feature works! And I have a basic test to prevent
regressions
Self-Review Checklist:
- [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
Closes #ISSUE
Release Notes:
- N/A or Added/Fixed/Improved ...
---------
Co-authored-by: cameron <cameron.studdstreet@gmail.com>
Self-Review Checklist:
- [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
- [x] Tests cover the new/changed behavior
- [x] Performance impact has been considered and is acceptable
Closes#52239
## Summary
Note - The code in this fix is AI generated.
OpenCode ACP chats lost slash-command support after Zed restarted and a
thread was reopened. The UI no longer showed `/ for commands`, and slash
commands like `/help` were treated as unsupported.
## Root Cause
ACP available commands were treated as transient UI state instead of
durable thread state.
- `AcpThread` handled `AvailableCommandsUpdate` but did not retain the
commands on the thread
- restored thread views rebuilt `SessionCapabilities` with an empty
`available_commands` list
- the message-editor placeholder started in the wrong state for restored
threads
- live command updates could be applied to the wrong thread view
## Fix
- persisted `available_commands` on `AcpThread`
- restored `SessionCapabilities` from thread state
- reused the same command augmentation logic for restore and live update
paths
- updated live command handling to target the correct thread view
- initialized the message-editor placeholder from current command
availability
- added a regression test for the restore path
## Verification
- `cargo test -p agent_ui
conversation_view::tests::test_restored_threads_keep_available_commands
-- --exact --nocapture`
- `./script/clippy -p agent_ui --tests`
Release Notes:
- Fixed ACP slash commands disappearing after reopening restored
threads.
---------
Co-authored-by: Ben Brandt <benjamin.j.brandt@gmail.com>
Self-Review Checklist:
- [ ] 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
Follow up to https://github.com/zed-industries/zed/pull/49807
Previously, when there were multiple conflicts across the codebase, we
would pop a toast at the bottom right corner of the UI. A toast seemed
like a functional idea because it'd be visible from any state of the app
and thus it'd be a good place to expose the button that allows you to
quickly prompt the agent to resolve all conflicts, as opposed to
creating a thread for each individual one. However, the toast was met
with some negative (and correct) feedback, mostly because it is
interruptive, and thus can sometimes block very relevant surfaces, like
either the agent panel itself or the Git commit area.
Therefore, in this PR, I'm removing the toast and adding a button in the
status bar instead; a bit more minimal, not interruptive, and a common
place for other items that might require your attention. The status bar
can be quite busy these days, though; we can display diagnostics, LSP
status, and file names in there; conscious of that. But it felt like it
could work given this button is such a transient one that you can either
easily manually dismiss or wait for it to be auto-dismissed as you or
the agent resolves the merge conflicts.
<img width="500" height="864" alt="Screenshot 2026-04-02 at 9 15@2x"
src="https://github.com/user-attachments/assets/4412a05c-77d0-4391-8ea1-25d1749b5e20"
/>
Release Notes:
- Git: Improved how we surface the affordance to resolve codebase-wide
merge conflicts with the agent in the UI.
- Agent: Added a setting to control whether or not the button to resolve
merge conflicts with the agent should be displayed.
This PR fixes an issue where an image mention would have its label reset
to just "Image", instead of persisting the original label, when the
prompt got submitted.
Closes#48564
Release Notes:
- agent: Fixed image mention labels by persisting the file name after
submitting the prompt
- agent: Fixed directory mentions being incorrectly parsed as files when
pasting into prompt editor
---------
Co-authored-by: Bennet Bo Fenner <bennetbo@gmx.de>
Currently, when pressing Shift+Tab to change Zed Agent's profile, the UI
isn't immediately updated. This PR fixes this issue so the `Change
Profile` button updates immediately after pressing Shift+Tab.
The current behavior. Observe that the `Change Profile` button doesn't
update right after Shift+Tab changes the active profile:
https://github.com/user-attachments/assets/fa1e6488-0dc3-4cc9-a4f3-7f62da48cc19
After this fix, the button text is update immediately on profile change:
https://github.com/user-attachments/assets/93261b11-037a-42c9-b1b8-0ca1e1adb851
---
Release Notes:
- Fixed Zed Agent profile selector button not visually updating when
cycled with Shift+Tab.
Signed-off-by: k4yt3x <i@k4yt3x.com>
Some model like glm-5、kimi-k2.5 support reasoning, but require
reasoning_effort parameter
This pr add support for setting reasoing_effort for openai compatible
models
Tested using the following config:
```json
{
"language_models": {
"openai_compatible": {
"My LiteLLM": {
"available_models": [
{
"name": "glm-5",
"display_name": "glm-5",
"max_tokens": 73728,
"reasoning_effort": "low"
},
{
"name": "kimi-k2.5",
"display_name": "kimi-k2.5",
"max_tokens": 262144,
"reasoning_effort": "low"
}
]
}
}
}
}
```
Release Notes:
- Added a setting to control `reasoning_effort` in custom
OpenAI-compatible models
The archive view's `fuzzy_match_positions` used `chars().enumerate()`
which produces **character indices**, not **byte indices**. When thread
titles contain multi-byte UTF-8 characters (emoji, CJK, accented
characters, etc.), these character indices don't correspond to valid
byte boundaries, causing a panic in `HighlightedLabel::new` which
asserts that highlight indices are valid UTF-8 boundaries.
The fix switches to `char_indices()` and `eq_ignore_ascii_case()` to
produce correct byte positions, matching the approach already used by
the sidebar's version of the same function.
Release Notes:
- Fixed a crash when searching archived threads whose titles contain
emoji or other non-ASCII characters.
Release Notes:
* [x] It's possible to get into a state where agent panel shows a thread
that is archived
- N/A
---------
Co-authored-by: Eric Holk <eric@zed.dev>
Co-authored-by: Mikayla Maki <mikayla.c.maki@gmail.com>
Follow up to https://github.com/zed-industries/zed/pull/53017
This PR does some significant refactoring of the `follow_tail` feature
in the GPUI list. That's only used by the agent panel's thread view and
given to the height-changing nature of streaming agent responses, we
were seeing some scroll snapping bugs upon scrolling while the thread is
generating. In the process of fixing it, we introduced a
`remeasure_items` method as an alternative to `splice` so that we could
get the remeasurement fix without scroll position changes. We already
had a `remeasure` method that did that for all of the indexes, but we
needed something more scoped out for the agent panel case, so as to not
remeasure the entire list's content on every new streamed token.
Effectively, this ends up reverting what the PR linked above introduced,
but it improved the API in the process.
Release Notes:
- N/A
Co-authored-by: Mikayla Maki <mikayla.c.maki@gmail.com>
This PR decouples `language_model`'s dependence on Zed-specific
implementation details. In particular
* `credentials_provider` is split into a generic `credentials_provider`
crate that provides a trait, and `zed_credentials_provider` that
implements the said trait for Zed-specific providers and has functions
that can populate a global state with them
* `zed_env_vars` is split into a generic `env_var` crate that provides
generic tooling for managing env vars, and `zed_env_vars` that contains
Zed-specific statics
* `client` is now dependent on `language_model` and not vice versa
Release Notes:
- N/A
Fixes#52656
Self-Review Checklist:
- [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
Closes#52656
Release Notes:
- Added keybindings for scrolling in agent view
---------
Co-authored-by: Oleksiy Syvokon <oleksiy.syvokon@gmail.com>
This PR fixes an issue where, sometimes, you couldn't scroll all the way
to the bottom of the thread's content. The scrollbar would show up at
the bottom of the scrollable container but the content was visibly cut
off. Turns out that's a consequence of the top-down thread generation
introduced in https://github.com/zed-industries/zed/pull/52440, where
changing the list alignment to `Top` made it visible that sometimes, the
maximum scroll area would get underestimated because the items in the
thread view's list would have a stale height measurement. So, the way
this PR fixes the issue is by calling `splice_focusable` in the
`EntryUpdated` event, too, so that the height of the items in the
overdraw area get marked as unmeasured, triggering a list re-render and
re-measuring.
We started by writing a test at the list level that would reproduce the
regression but then later figured out that this is not an inherent list
problem; it was rather a problem with its use within the thread view
layer. Then, we explored writing a test that documented the regression,
but it turned out to be very hard to simulate this sort of set up in
which certain elements would have its height changed during streaming,
which would be how you'd get to a mismatched height situation.
Therefore, given `AcpThreadEvent::NewEntry` already called
`splice_focusable` and don't have a test for it, we figure it'd be safe
to move forward without one, too. We then introduced a helper that's now
shared between `AcpThreadEvent::NewEntry` and
`AcpThreadEvent::EntryUpdated`.
Release Notes:
- Agent: Fixed an issue where sometimes you couldn't scroll all the way
to the bottom of the thread even though there's visibly more content
below the fold.
Co-authored-by: Eric Holk <eric@zed.dev>
Self-Review Checklist:
- [X] I've reviewed my own diff for quality, security, and reliability
- [ ] Unsafe blocks (if any) have justifying comments
- [ ] 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
- [ ] Performance impact has been considered and is acceptable
Release Notes:
- Added the option for highlights from languages to specify fallbacks.
That means that if you have a pattern with the captures `@second.capture
@first.capture`, Zed will first try resolving a highlight from your
theme for the code fragment using the first capture, then look for the
second capture if no match for the first capture could be found.
---------
Co-authored-by: Kirill Bulatov <kirill@zed.dev>
Make sure we don't try and load a thread we don't have metadata for and
therefore is unlikely to exist
Self-Review Checklist:
- [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
We were inserting threads with no entries into the sidebar, when you
clicked on New Thread, typed a message and then added a new project to
the current workspace.
We now skip persisting threads when they are empty.
Self-Review Checklist:
- [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: Ben Brandt <benjamin.j.brandt@gmail.com>
This PR improves the unreliability of the new top-down flow when editing
previous user messages. Previously, sometimes the thread wouldn't
auto-scroll as you're editing older messages and re-submitting them. It
turns out, the culprit seems to be a bit of a race condition where the
Stopped event (which sets the follow_tail method to false) could still
be emitted a while after a generation has been stopped. So depending on
time, one given possibility could happen. But now, this is what we do:
- If it is still generating, that means the stop/error event is stale
relative to a newer turn, so we do not run the cleanup that disables
follow-tail.
- If it is not generating, then it’s a real completion for the active
turn, and we do the normal cleanup.
---
Release Notes:
- N/A
This PR has us collecting the main worktree path metadata for the
sidebar, allows us to query by that path, and in the interim adds in
logic for querying the main worktree paths. This means that, as this
change percolates, there shouldn't be much difference between opening a
workspace to a linked worktree, and opening a workspace to the main
worktree.
Self-Review Checklist:
- [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
### Summary
This PR starts work on adding basic hook support in the `TaskStore`. To
enable users to setup tasks that are ran when the agent panel creates a
new git worktree to start a thread in. It also adds a new task variable
called `ZED_MAIN_GIT_WORKTREE` that's the absolute path to the main repo
that the newly created linked worktree is based off of.
### Follow Ups
- Get this hook working with the git worktree picker as well
- Make a more general approach to the hook system in `TaskStore`
- Add `ZED_PORT_{1..10}` task variables
- Migrate the task context creation code from `task_ui` to the basic
context provider
Before you mark this PR as ready for review, make sure that you have:
- [ ] 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 *or* Added/Fixed/Improved ...
---------
Co-authored-by: Remco Smits <djsmits12@gmail.com>
Co-authored-by: Richard Feldman <oss@rtfeldman.com>
TODO:
- [x] merge main
- [x] nonshrinking `set_excerpts_for_path`
- [x] Test-drive potential problem areas in the app
- [x] prepare cloud side
- [x] test collaboration
- [ ] docstrings
- [ ] ???
## Context
### Background
Currently, a multibuffer consists of an arbitrary list of
anchor-delimited excerpts from individual buffers. Excerpt ranges for a
fixed buffer are permitted to overlap, and can appear in any order in
the multibuffer, possibly separated by excerpts from other buffers.
However, in practice all code that constructs multibuffers does so using
the APIs defined in the `path_key` submodule of the `multi_buffer` crate
(`set_excerpts_for_path` etc.) If you only use these APIs, the resulting
multibuffer will maintain the following invariants:
- All excerpts for the same buffer appear contiguously in the
multibuffer
- Excerpts for the same buffer cannot overlap
- Excerpts for the same buffer appear in order
- The placement of the excerpts for a specific buffer in the multibuffer
are determined by the `PathKey` passed to `set_excerpts_for_path`. There
is exactly one `PathKey` per buffer in the multibuffer
### Purpose of this PR
This PR changes the multibuffer so that the invariants maintained by the
`path_key` APIs *always* hold. It's no longer possible to construct a
multibuffer with overlapping excerpts, etc. The APIs that permitted
this, like `insert_excerpts_with_ids_after`, have been removed in favor
of the `path_key` suite.
The main upshot of this is that given a `text::Anchor` and a
multibuffer, it's possible to efficiently figure out the unique excerpt
that includes that anchor, if any:
```
impl MultiBufferSnapshot {
fn buffer_anchor_to_anchor(&self, anchor: text::Anchor) -> Option<multi_buffer::Anchor>;
}
```
And in the other direction, given a `multi_buffer::Anchor`, we can look
at its `text::Anchor` to locate the excerpt that contains it. That means
we don't need an `ExcerptId` to create or resolve
`multi_buffer::Anchor`, and in fact we can delete `ExcerptId` entirely,
so that excerpts no longer have any identity outside their
`Range<text::Anchor>`.
There are a large number of changes to `editor` and other downstream
crates as a result of removing `ExcerptId` and multibuffer APIs that
assumed it.
### Other changes
There are some other improvements that are not immediate consequences of
that big change, but helped make it smoother. Notably:
- The `buffer_id` field of `text::Anchor` is no longer optional.
`text::Anchor::{MIN, MAX}` have been removed in favor of
`min_for_buffer`, etc.
- `multi_buffer::Anchor` is now a three-variant enum (inlined slightly):
```
enum Anchor {
Min,
Excerpt {
text_anchor: text::Anchor,
path_key_index: PathKeyIndex,
diff_base_anchor: Option<text::Anchor>,
},
Max,
}
```
That means it's no longer possible to unconditionally access the
`text_anchor` field, which is good because most of the places that were
doing that were buggy for min/max! Instead, we have a new API that
correctly resolves min/max to the start of the first excerpt or the end
of the last excerpt:
```
impl MultiBufferSnapshot {
fn anchor_to_buffer_anchor(&self, anchor: multi_buffer::Anchor) -> Option<text::Anchor>;
}
```
- `MultiBufferExcerpt` has been removed in favor of a new
`map_excerpt_ranges` API directly on `MultiBufferSnapshot`.
## 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: Conrad Irwin <conrad.irwin@gmail.com>
Co-authored-by: Piotr Osiewicz <24362066+osiewicz@users.noreply.github.com>
Co-authored-by: Jakub Konka <kubkon@jakubkonka.com>
Co-authored-by: Conrad <conrad@zed.dev>
