Sometimes, inside the edit agent, Sonnet thinks that it's doing a tool
call and closes its response with `</parameter></invoke>` instead of
properly closing </new_text>.
A better but more labor-intensive way of fixing this would be switching
to streaming tool calls for LLMs that support it.
Closes#39921
Release Notes:
- Fixed Sonnet's regression with inserting `</parameter></invoke>`
sometimes
This PR makes it clearer that you can click on the file path to open the
corresponding file in the agent panel's "edit bar", which is the element
that shows up in the panel as soon as agent-made edits happen.
Release Notes:
- agent panel: Improved the "go to file" affordance in the edit bar.
When we introduced the ACP-based agent panel, the condition that the
"review" | "reject" | "keep" buttons observed to be displayed got
mismatched between the panel and the pane (when in the single file
review scenario). In the panel, the buttons appear as soon as there are
changed buffers, whereas in the pane, they appear when response
generation is done.
I believe that making them appear at the same time, observing the same
condition, is the desired behavior. Thus, I think the panel behavior is
more correct, because there are loads of times where agent response
generation isn't technically done (e.g., when there's a command waiting
for permission to be run) but the _file edit_ has already been performed
and is in a good state to be already accepted or rejected.
So, this is what this PR is doing; effectively removing the "generating"
state from the agent diff, and switching to `EditorState::Reviewing`
when there are changed buffers.
Release Notes:
- Improved agent edit single file reviews by making the "reject" and
"accept" buttons appear at the same time.
Some housekeeping updates:
- Update hardcoded actions/keybindings so they're pulled from the repo
- Mention settings window when useful
- Add more info about agent panel's font size
- Break sentences in individual lines
Release Notes:
- N/A
These have been migrated to the README.md
[here](https://github.com/zed-industries/release_notes). These don't
need to be public. Putting them in the same repo where we draft
(`release_notes`) means less jumping around and allows us to include
additional information we might not want to make public.
Release Notes:
- N/A
Fixes the Debugpy toolchain detection bug in #40324
When detecting what toolchain (venv) to use in the Debugpy configuration
stage, we used to only base it off of the current working directory
argument passed to the config. This is wrong behavior for cases like
mono repos, where the correct virtual environment to use is nested in
another folder.
This PR fixes this issue by adding the program and module fields as
fallbacks to check for virtual environments. We also added support for
program/module relative paths as well when cwd is not None.
Release Notes:
- debugger: Improve mono repo virtual environment detection with Debugpy
---------
Co-authored-by: Remco Smits <djsmits12@gmail.com>
Adds a `convert-example` subcommand to the zeta cli that converts eval
examples from/to `json`, `toml`, and `md` formats.
Release Notes:
- N/A
---------
Co-authored-by: Max Brunsfeld <maxbrunsfeld@gmail.com>
Release Notes:
- N/A
---
Previously, agent markdown rendering used hardcoded font sizes
(TextSize::Default and TextSize::Small) which ignored the
agent_ui_font_size and agent_buffer_font_size settings. This updates the
markdown style to respect these settings.
This pull request adds support for customizing the font size of code
blocks in agent responses, making it possible to set a distinct font
size for code within the agent panel. The changes ensure that if the new
setting is not specified, the font size will fall back to the agent UI
font size, maintaining consistent appearance.
(I am a frontend developer without any Rust knowledge so this is
co-authored with Claude Code)
**Theme settings extension:**
* Added a new `agent_buffer_code_font_size` setting to
`ThemeSettingsContent`, `ThemeSettings`, and the default settings JSON,
allowing users to specify the font size for code blocks in agent
responses.
[[1]](diffhunk://#diff-a3bba02a485aba48e8e9a9d85485332378aa4fe29a0c50d11ae801ecfa0a56a4R69-R72)
[[2]](diffhunk://#diff-aed3a9217587d27844c57ac8aff4a749f1fb1fc5d54926ef5065bf85f8fd633aR118-R119)
[[3]](diffhunk://#diff-42e01d7aacb60673842554e30970b4ddbbaee7a2ec2c6f2be1c0b08b0dd89631R82-R83)
* Updated the VSCode import logic to recognize and import the new
`agent_buffer_code_font_size` setting.
**Font size application in agent UI:**
* Modified the agent UI rendering logic in `thread_view.rs` to use the
new `agent_buffer_code_font_size` for code blocks, and to fall back to
the agent UI font size if unset.
[[1]](diffhunk://#diff-f73942e8d4f8c4d4d173d57d7c58bb653c4bb6ae7079533ee501750cdca27d98L5584-R5584)
[[2]](diffhunk://#diff-f73942e8d4f8c4d4d173d57d7c58bb653c4bb6ae7079533ee501750cdca27d98L5596-R5598)
* Implemented a helper method in `ThemeSettings` to retrieve the code
block font size, with fallback logic to ensure a value is always used.
* Updated the settings application logic to propagate the new code block
font size setting throughout the theme system.
### Example Screenshots
---------
Co-authored-by: Danilo Leal <daniloleal09@gmail.com>
Currently Zed only displays agent notifications (e.g. when the agent
completes a task) if the user has switched apps and Zed is not in the
foreground. This adds PR supports the scenario where the agent finishes
a long-running task and the user is busy coding within Zed on something
else.
Releases Note:
- If agent notifications are turned on, they will now also be displayed
when the agent panel is hidden, in complement to them showing when the
Zed window is in the background.
---------
Co-authored-by: Danilo Leal <daniloleal09@gmail.com>
Closes#41458
Dynamically position mode selector tooltip to prevent clipping.
Position tooltip on the right when panel is docked left, otherwise on
the left. This ensures the tooltip remains visible regardless of panel
position.
**Note:** The tooltip currently vertically aligns with the bottom of the
menu rather than individual items. Would be great if it can be aligned
with the option it explains. But this doesn't seem trivial to me to
implement and not sure if it's important enough atm?
Before:
<img width="431" height="248" alt="Screenshot 2025-10-30 at 22 21 09"
src="https://github.com/user-attachments/assets/073f5440-b1bf-420b-b12f-558928b627f1"
/>
After:
<img width="632" height="158" alt="Screenshot 2025-10-30 at 17 26 52"
src="https://github.com/user-attachments/assets/e999e390-bf23-435e-9df0-3126dbc14ecb"
/>
<img width="685" height="175" alt="Screenshot 2025-10-30 at 17 27 15"
src="https://github.com/user-attachments/assets/84efca94-7920-474b-bcf8-062c7b59a812"
/>
Release Notes:
- Improved the agent panel's mode selector by preventing it to go
off-screen in case the panel is docked to the left.
Currently, if a commit operation takes some time, there's no visual
feedback in the UI that anything's happening.
This PR changes the colour of the text on the button to the
`Color::Disabled` colour when a commit operation is pending.
Release Notes:
- Improved UI feedback when a commit is in progress
---------
Co-authored-by: Danilo Leal <daniloleal09@gmail.com>
Closes#38345, #34882, #33280
Debugpy has four distinct configuration scenarios, which are:
1. launch
2. attach with process id
3. attach with listen
4. attach with connect
Spawning Debugpy directly works with the first three scenarios but not
with "attach with connect". Which requires host/port arguments being
passed in both with an attach request and when starting up Debugpy. This
PR passes in the right arguments when spawning Debugpy in an attach with
connect scenario, thus fixing the bug.
The VsCode extension comment that explains this:
98f5b93ee4/src/extension/debugger/adapter/factory.ts (L43-L51)
Release Notes:
- debugger: Fix Python attach-based sessions not working with `connect`
or `port` arguments
Currently, this only applies to long-running individually selected
unstaged files in the git panel. Next up I would like to make this work
for `Stage All`/`Unstage All` however this will most likely require
pushing `PendingOperation` into `GitStore` (from the `GitPanel`).
Release Notes:
- N/A
We missed making extensions that provide agent servers fill the
`provides` field with `agent-servers`, and thus, filtering for this type
of extension in both the app and site wouldn't return anything.
Release Notes:
- N/A
Just tidying it up by removing the unnecessary eye icon buttons in all
list items and adding that action in the form of a button in the footer,
closer to all other actions. Also reordering the footer buttons so that
the likely most common action is in the far right.
Release Notes:
- N/A
Given agent servers will soon be a thing, I'm adding Claude Code, Gemini
CLI, and Codex CLI as included agents in case anyone comes first to
search them as extensions before looking up on the agent panel.
Release Notes:
- N/A
Closes#17524
This PR adds a button to the bottom right corner of the ollama settings
ui. It resets the available ollama models, also resets the "Connected"
state in the process. This means it can be used to check if the
connection is still valid as well. It's a question whether we should
clear the available models on ALL `fetch_models` calls, since these only
happen during auth anyway.
Ollama is a local model provider which means clicking the refresh button
often only flashes the "not connected" state because the latency of the
request is so low. This accentuates changes in the UI, however I don't
think there's a way around this without adding some rather cumbersome
deferred ui updates.
I've attached the refresh button to the "Connected" `ButtonLike`, since
I don't think automatic UI spacing should separate these elements. I
think this is okay because the "Connected" isn't actually something that
the user can interact with.
Before:
<img width="211" height="245" alt="image"
src="https://github.com/user-attachments/assets/ea90e24a-b603-4ee2-9212-2917e1695774"
/>
After:
<img width="211" height="250" alt="image"
src="https://github.com/user-attachments/assets/be9af950-86a2-4067-87a0-52034a80a823"
/>
Alternative approach: There was also a suggestion to simply add a entry
to the command palette, however none of the other providers have this
ability currently either so I went with this approach. The current
approach also makes it more discoverable to the user.
Release Notes:
- Added a button for refreshing available ollama models
---------
Co-authored-by: Bennet Bo Fenner <bennetbo@gmx.de>
Closes#41478
Release Notes:
- Fixed#41478
<img width="459" height="916" alt="Screenshot 2025-10-29 at 1 31 26 PM"
src="https://github.com/user-attachments/assets/1d5b9fdf-9800-44e4-bdd5-f0964f93625f"
/>
> caused by using haiku 4.5 from the anthropic provider and then
swapping to sonnet 3.7 through zed, doing this does mess with prompt
caching but a model swap already invalidates that so it shouldn't have
any cost impact on end users
Update the behavior of the `zed_actions::agent::AddSelectionToThread`
action so that, after the selecitons are added to the current thread,
the editor automatically scrolls to the cursor's position, fixing an
issue where the inserted selection's UI component could wrap the cursor
to the next line below, leaving it outside the viewable area.
Closes#39694
Release Notes:
- Improved the `agent: add selection to thread` action so as to
automatically scroll to the cursor's position after selections are
inserted
Update `Vim::activate` to ensure that the `Vim.focused` method is only
called if the associated editor is also focused.
This ensures that the `VimEvent::Focused` event is only emitted when the
editor is actually focused, preventing a bug where, after starting Zed,
Vim's mode indicator would show that the mode was `Insert` even though
it was in `Normal` mode in the main editor.
Closes#41353
Release Notes:
- Fixed vim's mode being shown as `Inserted` right after opening Zed
Co-authored-by: Conrad Irwin <conrad.irwin@gmail.com>
When rerunning a task, our process id fetching seems to sometimes return
the previous terminal's process id when respawning the task, causing us
to kill the new terminal once the previous one drops as we spawn a new
one, then drop the old one. This results in rerun sometimes spawning a
blank task as the terminal immediately exits. The fix here is simple, we
actually want to kill the process running inside the terminal process,
not the terminal process itself when we exit in the terminal.
No relnotes as this was introduced yesterday in
https://github.com/zed-industries/zed/pull/41562
Release Notes:
- N/A *or* Added/Fixed/Improved ...
Closes#41125
Release Notes:
- Fixed `SwitchToHelixNormalMode` to keep selection
- Added default keybinds for `SwitchToHelixNormalMode` when in Helix
mode
