@agu-z and paired on trying out a "one tool call per edit" approach for
editing files. (The previous approach is still available, it's just
unchecked by default for now.)
Release Notes:
- N/A
---------
Co-authored-by: Agus <agus@zed.dev>
We decided to take this out for now. It doesn't seem necessary, and it
complicates the code a lot. We can always put it back later if desired.
Release Notes:
- N/A
Also includes some fixes for how the Lua tool was being generated.
<img width="644" alt="Screenshot 2025-03-21 at 6 26 18 PM"
src="https://github.com/user-attachments/assets/51bd1685-5b3f-4ed3-b11e-6fa8017847d4"
/>
Release Notes:
- N/A
---------
Co-authored-by: Ben <ben@zed.dev>
Co-authored-by: Agus Zubiaga <agus@zed.dev>
Co-authored-by: Joseph T. Lyons <JosephTLyons@gmail.com>
Co-authored-by: Danilo Leal <daniloleal09@gmail.com>
Co-authored-by: Ben Kunkle <ben.kunkle@gmail.com>
Closes#25671
Release Notes:
- Added support for `claude-3-7-sonnet-thinking` in the assistant panel
---------
Co-authored-by: Danilo Leal <daniloleal09@gmail.com>
Co-authored-by: Antonio Scandurra <me@as-cii.com>
Co-authored-by: Agus Zubiaga <hi@aguz.me>
Ensuring that text between the "you" messages align with text in the
assistant response. This also creates a nice subtle hierarchy effect
where the "you" message card is wider than the message, making it
slightly easier to tell them apart.
<img
src="https://github.com/user-attachments/assets/616c1776-ca51-454e-9d52-e480bf26c843"
width="600px" />
Release Notes:
- N/A
This is the core change:
https://github.com/zed-industries/zed/pull/26758/files#diff-044302c0d57147af17e68a0009fee3e8dcdfb4f32c27a915e70cfa80e987f765R1052
TODO:
- [x] Use AsyncFn instead of Fn() -> Future in GPUI spawn methods
- [x] Implement it in the whole app
- [x] Implement it in the debugger
- [x] Glance at the RPC crate, and see if those box future methods can
be switched over. Answer: It can't directly, as you can't make an
AsyncFn* into a trait object. There's ways around that, but they're all
more complex than just keeping the code as is.
- [ ] Fix platform specific code
Release Notes:
- N/A
When the model reads file, we'll track the version it read, and let it
know if the user makes edits to the buffer. This helps prevent edit
failures because it'll know to re-read the file before.
Release Notes:
- N/A
When the user attached context in the thread, the editor model request
would fail because its tool use wouldn't be removed properly leading to
an API error.
Also, after an edit, we'd keep the old file snapshot in the context.
This would make the model think that the edits didn't apply and make it
go in a loop.
Release Notes:
- N/A
Release Notes:
- Multibuffers now use less vertical space for excerpt boundaries.
Additionally the expand up/down arrows are hidden at the start and end
of the buffers
---------
Co-authored-by: Nate Butler <iamnbutler@gmail.com>
Co-authored-by: Zed AI <claude-3.5-sonnet@zed.dev>
Exposes a new "edit files" tool that the model can use to apply
modifications to files in the project. The main model provides
instructions and the tool uses a separate "editor" model (Claude 3.5 by
default) to generate search/replace blocks like Aider does:
````markdown
mathweb/flask/app.py
```python
<<<<<<< SEARCH
from flask import Flask
=======
import math
from flask import Flask
>>>>>>> REPLACE
```
````
The search/replace blocks are parsed and applied as they stream in. If a
block fails to parse, the tool will apply the other edits and report an
error pointing to the part of the input where it occurred. This should
allow the model to fix it.
Release Notes:
- N/A
---------
Co-authored-by: Antonio Scandurra <me@as-cii.com>
This PR decouples the scripting tool from the `Tool` trait while still
allowing it to be used as a tool from the model's perspective.
This will allow us to evolve the scripting tool as more of a first-class
citizen while still retaining the ability to have the model call it as a
regular tool.
Release Notes:
- N/A
We decided to expose scripting as tools again. We are aware of the UX
downsides of doing so, but we want to focus on getting it working well
first, and the model seems to make better use of it as an actual tool.
In the future, the tools API might support streaming. If it doesn't and
we need to ship, we can consider reverting this.
Release Notes:
- N/A
This PR makes refactors the scripting functionality to be a first-class
concept of the assistant instead of a generic tool, which will allow us
to build a more customized experience.
- The tool prompt has been slightly tweaked and is now included as a
system message in all conversations. I'm getting decent results, but now
that it isn't in the tools framework, it will probably require more
refining.
- The model will now include an `<eval ...>` tag at the end of the
message with the script. We parse this tag incrementally as it streams
in so that we can indicate that we are generating a script before we see
the closing `</eval>` tag. Later, this will help us interpret the script
as it arrives also.
- Threads now hold a `ScriptSession` entity which manages the state of
all scripts (from parsing to exited) in a centralized way, and will
later collect all script operations so they can be displayed in the UI.
- `script_tool` has been renamed to `assistant_scripting`
- Script source now opens in a regular read-only buffer
Note: We still need to handle persistence properly
Release Notes:
- N/A
---------
Co-authored-by: Marshall Bowers <git@maxdeviant.com>
This PR factors out a new `Thread::all_tools_finished` method to
encapsulate some of the boilerplate in the `ThreadEvent::ToolFinished`
event handler.
This should make this event handler easier to replicate for the eval
use-case.
Release Notes:
- N/A
This PR adds two new helper methods to the `Thread` for dealing with
tool use:
- `use_pending_tools` - This uses all of the tools that are pending
- The reason we aren't calling this directly in `stream_completion` is
that we still might need to have a way for users to confirm that they
want tools to be run, which would need to happen at the UI layer in the
`ActiveThread`.
- `send_tool_results_to_model` - This encapsulates inserting a new user
message that contains the tool results and sending them up to the model.
Release Notes:
- N/A
This PR decouples the `Tool` trait from the `Workspace` (and from the
UI, in general).
`Tool::run` now takes a `WeakEntity<Project>` instead of a
`WeakEntity<Workspace>` and a `Window`.
Release Notes:
- N/A
This PR adds a new rounded/corner border token: `rounded_md` with a
value of 6px.
I feel like I was wanting to use 6px border radius a lot but avoiding
due to it being an arbitrary value... so, not anymore! It's also cool to
have this be consistent with Tailwind v4.
Follow on to the prior renames:
- `rounded_sm` -> `rounded_xs`:
https://github.com/zed-industries/zed/pull/26221
- `rounded_md` -> `rounded_sm`:
https://github.com/zed-industries/zed/pull/26228
Release Notes:
- N/A
---------
Co-authored-by: Marshall Bowers <git@maxdeviant.com>
Fine-tuning the visuals (namely, reducing font and keybinding size) and
passing `on_click` handlers to the Cancel & Regenerate actions.
Release Notes:
- N/A
This PR adds the ability for Markdown code blocks and tables to be made
horizontally scrollable.
This is a feature that the caller can opt in to.
Right now we're using it for the rendered Markdown in the Assistant 2
panel.
Release Notes:
- N/A
This PR makes it so we include some textual content in the user messages
that as used to send up tool results.
I observed that when sending up the tool results with no text, it would
lead the model to start replying with no text, which would then result
in an error when attaching later tool results.
I think there's a deeper issue at play here, but for now we just include
some text to keep the model on track.
Release Notes:
- N/A
This PR reworks how we store tool results and refer to them later.
We now maintain a mapping of the tool uses to their corresponding
results, with separate mappings for the messages and the tool uses they
correspond to.
Release Notes:
- N/A
This PR updates the tool use flow in Assistant 2 to automatically
respond to the model with tool results when the tools have finished
running.
Release Notes:
- N/A
