This PR updates the `Thread::is_streaming` method so that it includes
tool use in the "streaming" state.
This will prevent the streaming indicator from disappearing when we're
doing tool use.
Release Notes:
- N/A
This PR updates the `ToolWorkingSet` to reduce the amount of locking we
need to do.
A number of the methods have had corresponding versions moved to the
`ToolWorkingSetState` so that we can take out the lock once and do a
number of operations without needing to continually acquire and release
the lock.
Release Notes:
- N/A
This PR adds a new `ExtensionEvents` event bus that can be used to
listen for extension-related events throughout the app.
Today you need to have a handle to the `ExtensionStore` (which entails
depending on `extension_host`) in order to listen for extension events.
With this change subscribers only need to depend on `extension`, which
has a leaner dependency graph.
Release Notes:
- N/A
This PR refactors the `ContextMenu::render` method to extract a couple
smaller methods from it.
The existing `render` method was suffering from its size, with some of
the `match` arms not being able to be formatted with `rustfmt`.
Release Notes:
- N/A
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
This PR fixes some language links in the docs.
The Shell Script page wasn't being linked from `SUMMARY.md`, so no page
was being generated.
There were also some differences in the language lists in the sidebar
and on the top-level languages page.
Release Notes:
- N/A
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 updates the `ExtensionBuilder` to require that grammar names are
written in snake_case.
The grammar names are used to construct identifiers, so we need them to
be valid C identifiers.
Release Notes:
- N/A
This PR removes a number of `.unwrap`s in the Bedrock provider.
We must not `.unwrap` in situations where it is not provably safe to do
so, which it was not in any of these cases.
Release Notes:
- Fixed some potential panics in the AWS Bedrock model provider.
This PR finishes up the support for generating commit messages using an
LLM.
We're shelling out to `git diff` to get the diff text, as it seemed more
efficient than attempting to reconstruct the diff ourselves from our
internal Git state.
https://github.com/user-attachments/assets/9bcf30a7-7a08-4f49-a753-72a5d954bddd
Release Notes:
- Git Beta: Added support for generating commit messages using a
language model.
---------
Co-authored-by: Conrad Irwin <conrad.irwin@gmail.com>
This PR adds support for capabilities for the extension process API.
In order to use the process API, an extension must declare which
commands it wants to use, with arguments:
```toml
[[capabilities]]
kind = "process:exec"
command = "echo"
args = ["hello!"]
```
A `*` can be used to denote a single wildcard in the argument list:
```toml
[[capabilities]]
kind = "process:exec"
command = "echo"
args = ["*"]
```
And `**` can be used to denote a wildcard for the remaining arguments:
```toml
[[capabilities]]
kind = "process:exec"
command = "ls"
args = ["-a", "**"]
```
Release Notes:
- N/A
---------
Co-authored-by: Conrad Irwin <conrad.irwin@gmail.com>
This PR adds two new tools to Assistant 2:
- `list-worktrees` - Lists the worktrees in a project
- `read-file` - Reads a file at the given path in the project
I don't see `list-worktrees` sticking around long-term, as when we have
tools for listing files those will include the worktree IDs along with
the path, but making this tool available allows the model to utilize
`read-file` when it otherwise wouldn't be able to.
Release Notes:
- N/A
This PR changes the default value when no input is provided with a tool
use from `null` to `{}`.
This fixes an issue I was seeing where tools that didn't accept input
were not being called correctly.
Release Notes:
- N/A
This PR fixes the language model selector.
I tried to piece together the state prior to #25697 (the state it was in
at 11838cf89e) while retaining unrelated
changes that happened since then.
Release Notes:
- Fixed an issue where language models would not be authenticated until
after the model selector was opened (Preview only).
This PR makes it so the copy button on Markdown code blocks stays
absolutely positioned even when scrolled:
<img width="1297" alt="Screenshot 2025-03-04 at 5 28 48 PM"
src="https://github.com/user-attachments/assets/b0d0fae9-ccd6-43c1-bef3-44d8d3c3e669"
/>
We achieve this by inserting a new parent element around both the copy
button and the code block itself so we can position the copy button
absolutely within that element.
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 adds a new `restrict_scroll_to_axis` style to allow consumers to
opt-in to the scrolling behavior found on the web.
When this is enabled the behavior will be such that:
- Scrolling using the mouse wheel will only scroll the Y axis
- Scrolling using the mouse wheel with <kbd>Shift</kbd> held will only
scroll the X axis
This behavior is useful in scenarios where you have some
vertically-scrollable content that is interspersed with
horizontally-scrollable elements, as otherwise the scroll will be
constantly hijacked by the horizontally-scrollable elements while trying
to scroll up and down in the vertically-scrollable container.
I think that this behavior should be the default, but it's a bit of a
sweeping change to make all at once, so for now it remains opt-in.
Release Notes:
- N/A
This PR adds initial support for displaying tables to the `markdown`
crate.
This allows us to render tables in Assistant 2:
| Before | After |
|
----------------------------------------------------------------------------------------------------------------------------------------------------
|
----------------------------------------------------------------------------------------------------------------------------------------------------
|
| <img width="1309" alt="Screenshot 2025-03-03 at 1 39 39 PM"
src="https://github.com/user-attachments/assets/ad7ada01-f35d-4fcf-a20c-deb42b55b34e"
/> | <img width="1297" alt="Screenshot 2025-03-03 at 3 38 21 PM"
src="https://github.com/user-attachments/assets/9b771126-30a0-479b-8c29-f5f572936f56"
/> |
There are a few known issues that should be addressed as follow-ups:
- The horizontal scrolling within a table is linked with the scrolling
of the parent container (e.g., the Assistant 2 thread)
- Cells are currently cut off entirely when they are too wide, would be
nice to truncate them with an ellipsis
Release Notes:
- N/A
This PR changes how we persist threads in Assistant2 to use `serde_json`
instead of `bincode` for the representation.
This makes the format more flexible to work with (and will allow for
using things like `#[serde(default)]`) if the schema changes over time.
Note: We have to bump the LMDB database version for this, so any threads
created before now will be gone.
Release Notes:
- N/A
This PR updates the extension CLI to make the use of
`scrollbar_thumb.background` in a theme a hard error.
We're working to eradicate usage of this theme property, so this will
prevent new extensions from being published that use it.
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 comments in the `default.json` file consistently use
`//`.
Some comments were using `///`, which doesn't make sense in JSONC.
Release Notes:
- N/A
This PR moves the `PopoverButton` component into the `ui` crate.
The `popover_button` crate only depended on `ui`, so there doesn't seem
to be a need for it to live in its own crate and add another step in the
crate graph.
Release Notes:
- N/A
This PR combines two disjoint conditions for the same value into one.
This makes it so the type checker can accurately reason about the
branches.
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
This PR adds a new `send_to_model` method to the `Thread` to encapsulate
more of the thread-specific capabilities.
We then call this in `MessageEditor::send_to_model`.
Release Notes:
- N/A
This PR fixes the generation of summaries for threads when tools are
being used.
Previously we were including the tool uses in the summarization request,
but this would result in invalid messages being sent to the model and
summaries not being generated.
We now exclude any tool uses or results from the model when summarizing
a thread.
Release Notes:
- N/A
This PR makes `auto` an alias for the `subtle` edit prediction mode.
Right now I'm in a state where I can't have valid settings in both
development and Nightly because the settings values are disparate.
Release Notes:
- N/A
This is a follow-up to https://github.com/zed-industries/zed/pull/25573.
We were still using the spend for a particular model when determining if
the user was over their maximum monthly spend instead of looking at the
usage across all models.
Release Notes:
- N/A
This PR fixes an issue introduced in #25530 that broke the notifications
that inform the user that a Zed update is required to continue using
edit prediction.
The issue is that the `Workspace` stored on the `Editor` is set _after_
the point we initialize Zeta, so capturing the `Workspace` at
construction time leads to it being `None`.
@ConradIrwin suggested that we could obtain the `Workspace` from the
`Window`, which does indeed do the trick.
I tested it both with and without this change by mocking the error
response, like so:
```rs
let response: Result<PredictEditsResponse, anyhow::Error> =
Err(anyhow!(ZedUpdateRequiredError {
minimum_version: SemanticVersion::new(0, 1, 0),
}));
```
Release Notes:
- N/A
This PR updates the client-side checks to give Zed AI users access to
Claude 3.7 Sonnet.
Requires https://github.com/zed-industries/zed/pull/25576 to be
deployed.
Release Notes:
- Added support for Claude 3.7 Sonnet to Zed AI.
This PR adjusts the usage checks for the LLM free tier.
Previously we would limit the usage on a per-model basis, meaning the
user would get $10/mo free for each model they had access to.
We now have usage for all models count towards the free tier limit.
Release Notes:
- N/A
This PR removes the dependents of the `language_models` crate.
The following types have been moved from `language_models` to
`language_model` to facilitate this:
- `LlmApiToken`
- `RefreshLlmTokenListener`
- `MaxMonthlySpendReachedError`
- `PaymentRequiredError`
With this change only `zed` now depends on `language_models`.
Release Notes:
- N/A
This PR moves the `report_assistant_event` function from the
`language_models` crate to the `language_model` crate.
This allows us to drop some dependencies on `language_models`.
Release Notes:
- N/A
This PR removes the dependencies on the individual model provider crates
from the `language_model` crate.
The various conversion methods for converting a `LanguageModelRequest`
into its provider-specific request type have been inlined into the
various provider modules in the `language_models` crate.
The model providers we provide via Zed's cloud offering get to stay, for
now.
Release Notes:
- N/A
This PR changes the color used for `@variable` syntax highlights in the
Gruvbox themes to be less intense.
We now use the same color as `editor.foreground`.
| Language | Before | After |
| -------- |
-----------------------------------------------------------------------------------------------------------------------------------------------------
|
-----------------------------------------------------------------------------------------------------------------------------------------------------
|
| Rust | <img width="1410" alt="Screenshot 2025-02-24 at 10 08 41 AM"
src="https://github.com/user-attachments/assets/9a34964d-9fdc-4deb-ac30-4a1c9e6fb531"
/> | <img width="1410" alt="Screenshot 2025-02-24 at 10 55 18 AM"
src="https://github.com/user-attachments/assets/c245d0fd-28af-42b8-93f6-48cb14671d94"
/> |
| Python | <img width="1410" alt="Screenshot 2025-02-24 at 10 08 38 AM"
src="https://github.com/user-attachments/assets/8f8d111e-1d50-4229-a333-eb29b6ce9f4f"
/> | <img width="1410" alt="Screenshot 2025-02-24 at 10 55 20 AM"
src="https://github.com/user-attachments/assets/010b661e-dc9e-4ccb-8e52-ee10c8eb8342"
/> |
In #25333 and #25331 the highlight used for identifiers in Rust and
Python, respectively, was changed to `@variable`, which resulted in the
intense colors you see in the "Before" screenshots above.
We considered reverting the highlight query changes to those languages,
but after taking a look at our other languages, they already use similar
queries. Instead we're adjusting the theme to make these cases less
visually intense.
Release Notes:
- Gruvbox themes: Changed the color used for `@variable` syntax
highlights to be less intense.
Marshall Bowers
