This PR is a follow-up to
https://github.com/zed-industries/zed/pull/22911 to further improve the
registration of code action providers for the Assistant in order to
prevent duplicates.
The `CodeActionProvider` trait now has an `id` method that is used to
return a unique ID for a code action provider. We use this to prevent
registering duplicates of the same provider.
The registration of the code action providers for Assistant1 and
Assistant2 have also been reworked. Previously we were not call the
registration function—and thus setting up the subscriptions—until we
resolved the feature flags. However, this could lead to the registration
happening too late for existing workspace items.
We now perform the registration right away and then remove the undesired
code action providers once the feature flags have been resolved.
Release Notes:
- N/A
Also:
* Adds `impl_internal_actions!` for deriving the `Action` trait without
registering.
* Removes some deserializers that immediately fail in favor of
`#[serde(skip)]` on fields where they were used. This also omits them
from the schema.
Release Notes:
- Keymap settings file now has more JSON schema information to inform
`json-language-server` completions and info, particularly for actions
that take input.
This PR fixes the duplicated `Fix with Assistant` code actions that were
being shown in the code actions menu.
This fix isn't 100% ideal, as there is an edge case in buffers that are
already open when the workspace loads, as we may not observe the feature
flags in time to register the code action providers by the time we
receive the event that an item was added to the workspace.
Closes https://github.com/zed-industries/zed/issues/22400.
Release Notes:
- Fixed duplicate "Fix with Assistant" entries showing in the code
action list.
This one seems to be triggered when the assistant's
`View<ContextEditor>` is leased during the call into
`NavHistory::for_each_entry`, which then tries to read it again through
the `ItemHandle` interface. Fix it by skipping entries that can't be
read in the history iteration.
Release Notes:
- N/A
Previously, to use a green and red shade with `TintColor` you'd need to
pass `Positive` and `Negative`, respectively. This terminology always
tripped me up, because, for example, I'd often try to use something
like:
```
Button::new("icon_color", "Negative")
style(ButtonStyle::Tinted(TintColor::Negative))
.color(Color::Error)
.icon_color(Color::Error)
.icon(IconName::Trash),
)
```
...and due to `icon_color` taking `Color::Error`, I'd always get
`TintColor` wrong at a first try, because I would, out of muscle memory,
write `TintColor::Error`, which wouldn't compile. That's exactly the
change in this PR—`TintColor` now takes `Success` and `Error` instead of
`Positive` and `Negative`, for more consistency.
Release Notes:
- N/A
- [x] Rewrite worktree git handling
- [x] Fix tests
- [x] Fix `test_propagate_statuses_for_repos_under_project`
- [x] Replace `WorkDirectoryEntry` with `WorkDirectory` in
`RepositoryEntry`
- [x] Add a worktree event for capturing git status changes
- [x] Confirm that the local repositories are correctly updating the new
WorkDirectory field
- [x] Implement the git statuses query as a join when pulling entries
out of worktree
- [x] Use this new join to implement the project panel and outline
panel.
- [x] Synchronize git statuses over the wire for collab and remote dev
(use the existing `worktree_repository_statuses` table, adjust as
needed)
- [x] Only send changed statuses to collab
Release Notes:
- N/A
---------
Co-authored-by: Cole Miller <cole@zed.dev>
Co-authored-by: Mikayla <mikayla@zed.com>
Co-authored-by: Nathan <nathan@zed.dev>
Removed a settings update that should have been removed in the 0.148.0
release.
I am not sure if there is a tracking issue, but I identified this check
for outdated settings that should not be needed anymore. I investigated
a bit and did not find any conflicts or UB as a result of removing this
code.
Release Notes:
- N/A
---------
Co-authored-by: Marshall Bowers <elliott.codes@gmail.com>
This means that `workspace::ToggleRightDock` will open the assistant if
no right-dock panel has been manually activated, instead of the chat as
before. Also cleans up the `active_panel_index` logic a bit.
cc @nathansobo
Release Notes:
- Make `workspace::ToggleRightDock` open the assistant panel if no
right-dock panel has yet been activated
* Remove unnecessary WindowContext and ViewContext '_ lifetimes
* Removed some cases where WindowContext has a different name than `cx`.
Release Notes:
- N/A
Closes https://github.com/zed-industries/zed/issues/19214
Closes https://github.com/zed-industries/zed/pull/22443
Adds `resolved` property into Zed completion item data, to ensure we
resolve every completion item exactly once.
There are 2 paths for singplayer Zed, and corresponding 2 analogues for
multi player code, where resolve may happen:
* completions menu display & selection, that ends up using
`resolve_completions` in `lsp_store.rs`
* applying a completion menu entry, that ends up using
`apply_additional_edits_for_completion` in `lsp_store.rs`
Now, all local counterparts check `enabled` field before resolving and
set it to true afterwards, and reuse the same `resolve_completion_local`
method for resolving the items.
A logic for re-generating docs and item labels was moved out from the
`resolve_completion_local` method into a separate method, as
`apply_additional_edits_for_completion` does not need that, but needs
the rest of the logic for resolving.
During the extraction, I've noted that multiplayer clients are not
getting the item labels, regenerated after the resolve — as the Zed
protocol-based flow is not the exact copy of the local resolving.
To improve that, `resolve_completion_remote` needs to be adjusted, but
this change is omitted to avoid bloating the PR.
Release Notes:
- Fixed autocomplete inserting multiple imports
This PR makes it so `assistant` and `assistant2` both use the same
action for inline assist (`zed_actions::InlineAssist`).
This makes it so the keybindings to deploy the inline assist seamlessly
swap based on the feature flag without needing to rebind them.
One minor caveat: if you're using `assistant2` the action name in the
command palette will be `assistant: inline assist`.
Release Notes:
- N/A
This PR adds a `text_ellipsis` method to `Label`s.
This can be used to truncate the text with an ellipsis without needing
to wrap the `Label` in another element.
Release Notes:
- N/A
`CodeContextMenu` is always accessed on one thread, so only `Rc`s and
`Rc<RefCell<_>>` are needed. There should be tiny performance benefits
from this. The main benefit of this is that when seeing code accessing a
`RwLock` it would be reasonable to wonder whether it will block. The
only potential downside is the potential for panics due to overlapping
borrows of the RefCells. I think this is an acceptable risk because most
errors of this nature will be local or will be caught by clippy via the
check for holding a RefCell reference over an `await`.
Release Notes:
- N/A
When using inline assist, after successfully generating a transformation
it's not possible to generate a new transformation. Currently, you have
to modify the prompt (i.e. add a `<SPACE>` and hit `<ENTER>`) to
regenerate.
So, I changed the restart button to be visible after a successful
transformation. And in that case I map it to a different keyboard
shortcut because `menu::Confirm` is mapped to accept the current
suggestion.
Now, I can invoke a series of transforms back to back until I get what I
want!
It might also be desired to keep the accept button visible after
modifying the prompt (before submitting it). In that case we'll need to
remap accept to an alternate key (perhaps the same alt-shift-enter I am
using for restart. That wouldn't be too insane to remember. But maybe
someone has a better idea.
I don't care what the shortcut is, I just want the ability to regenerate
without adding/deleting spaces.
## Before
**Two choices** after a suggestions is presented. Also, a great example
of why I would want to regenerate the suggestion, it left some tokens
`<rewrite_this>`!
## After
**Three choices** after a suggestion is presented, the circular icon is
for regenerate, just like you see if you modify the prompt text.
## Release Notes:
- Added Restart Button to Inline Assistant When Prompt Is Unchanged
---------
Co-authored-by: Danilo Leal <daniloleal09@gmail.com>
- Adds the Switch component
- Updates `Selected`, `Selectable` -> `ToggleState`, `Toggleable`
- Adds `checkbox` and `switch` functions to align better with other
elements in our layout system.
We decided not to merge Switch and Checkbox. However, in a followup I'll
introduce a Toggle or AnyToggle enum so we can update
`CheckboxWithLabel` -> `ToggleWithLabel` as this component will work
exactly the same with either a Checkbox or a Switch.
Release Notes:
- N/A
Noticed that whitespace-only insertions are really hard to make out, so
this changes it to make them visible by giving them a green background.
Release Notes:
- N/A
---------
Co-authored-by: Antonio <antonio@zed.dev>
While working on Assistant2, I noticed that the `LanguageModelSelector`
was recreating its `Picker` view on every single render.
This PR makes it so we create the view once and hold onto it in the
parent view.
Release Notes:
- N/A
Still TODO:
* [x] Factor out `start_language_server` so we can call it on register
(instead of on detect language)
* [x] Only call register in singleton editors (or when
editing/go-to-definition etc. in a multibuffer?)
* [x] Refcount on register so we can unregister when no buffer remain
* [ ] (maybe) Stop language servers that are no longer needed after some
time
Release Notes:
- Fixed language servers starting when doing project search
- Fixed high CPU usage when ignoring warnings in the diagnostics view
---------
Co-authored-by: Piotr Osiewicz <24362066+osiewicz@users.noreply.github.com>
Co-authored-by: Cole <cole@zed.dev>
This PR updates the `gpui::prelude` to not export the `Context` trait
named.
This prevents some naming clashes in downstream consumers.
Release Notes:
- N/A
Ensuring it is consistent with the buffer inline assistant. Just thought
of not having "Transform" here as that felt it made less sense for
terminal-related prompts, where arguably more frequently, one would be
suggesting for actual commands rather than code transformation.
<img width="700" alt="Screenshot 2024-12-02 at 09 11 00"
src="https://github.com/user-attachments/assets/ad96d03e-0366-46e8-8056-581066712d59">
Release Notes:
- N/A
The `/fetch` command is naturally already accessible via the completion
menu when you type / in the assistant panel, but it wasn't on the "Add
Context" command selector. I think it should! It's a super nice/powerful
one, and I've seen folks not knowing it existed. Side-note: maybe, in
the near future, it'd be best to rename it to "`/web`, as that's an
easier name to parse and assume what it does.
<img width="700" alt="Screenshot 2024-11-28 at 16 52 07"
src="https://github.com/user-attachments/assets/37134e1c-c788-48ca-81ae-d7416e8c8706">
Release Notes:
- N/A
Closes https://github.com/zed-industries/zed/issues/4351
Applies the same splitting mechanism, as Zed's central pane has, to the
terminal panel.
Similar navigation, splitting and (de)serialization capabilities are
supported.
Notable caveats:
* zooming keeps the terminal splits' ratio, rather expanding the
terminal pane
* on macOs, central panel is split with `cmd-k up/down/etc.` but `cmd-k`
is a "standard" terminal clearing keybinding on macOS, so terminal panel
splitting is done via `ctrl-k up/down/etc.`
* task terminals are "split" into regular terminals, and also not
persisted (same as currently in the terminal)
Seems ok for the initial version, we can revisit and polish things
later.
Release Notes:
- Added the ability to split the terminal panel
This PR restructures the storage of the tool uses and results in
`assistant2` so that they don't live on the individual messages.
It also introduces a `LanguageModelToolUseId` newtype for better type
safety.
Release Notes:
- N/A
This PR adds rudimentary support for using tools to `assistant2`. There
are currently no visual affordances for tool use.
This is gated behind the `assistant-tool-use` feature flag.
<img width="1079" alt="Screenshot 2024-11-25 at 7 21 31 PM"
src="https://github.com/user-attachments/assets/64d6ca29-c592-4474-8e9d-c344f855bc63">
Release Notes:
- N/A
This PR factors the tool definitions out of the `assistant` crate so
that they can be shared between `assistant` and `assistant2`.
`ToolWorkingSet` now lives in `assistant_tool`. The tool definitions
themselves live in `assistant_tools`, with the exception of the
`ContextServerTool`, which has been moved to the `context_server` crate.
As part of this refactoring I needed to extract the
`ContextServerSettings` to a separate `context_server_settings` crate so
that the `extension_host`—which is referenced by the `remote_server`—can
name the `ContextServerSettings` type without pulling in some undesired
dependencies.
Release Notes:
- N/A
This reverts #20824 and #20899. After adding them last week we came to
the conclusion that the hints are too distracting in everyday use, see
#21128 for more details.
Release Notes:
- N/A
This PR factors the language model selector out into its own
`language_model_selector` crate so that it can be reused in
`assistant2`.
Also renamed it from `ModelSelector` to `LanguageModelSelector` to be a
bit more specific.
Release Notes:
- N/A
This PR overhauls extension registration in order to make it more
modular.
The `extension` crate now contains an `ExtensionHostProxy` that can be
used to register various proxies that the extension host can use to
interact with the rest of the system.
There are now a number of different proxy traits representing the
various pieces of functionality that can be provided by an extension.
The respective crates that provide this functionality can implement
their corresponding proxy trait in order to register a proxy that the
extension host will use to register the bits of functionality provided
by the extension.
Release Notes:
- N/A
See ["mtime comparison considered
harmful"](https://apenwarr.ca/log/20181113) for details of why
comparators other than equality/inequality should not be used with
mtime.
Release Notes:
- N/A
This PR adds a new `language_models` crate to house the various language
model providers.
By extracting the provider definitions out of `language_model`, we're
able to remove `language_model`'s dependency on `editor`, which improves
incremental compilation when changing `editor`.
Release Notes:
- N/A
This PR fixes an issue where slash commands in the output of other slash
commands were not being evaluated when configured to do so.
Closes https://github.com/zed-industries/zed/issues/20820.
Release Notes:
- Fixed slash commands from other slash commands (like `/default`) not
being evaluated (Preview only).
Co-Authored-by: Thorsten <thorsten@zed.dev>
Co-Authored-by: Antonio <antonio@zed.dev>
Screenshot:
TODO:
- [x] docs
Release Notes:
- Added inline hints that guide users on how to invoke the inline
assistant and open the assistant panel. (These hints can be disabled by
setting `{"assistant": {"show_hints": false}}`.)
---------
Co-authored-by: Thorsten <thorsten@zed.dev>
Co-authored-by: Antonio <antonio@zed.dev>
Co-authored-by: Thorsten Ball <mrnugget@gmail.com>
Meant to avoid the excessive use of "Here's a concise 3-7 word title..."
and "Title:" instances we've been seeing lately.
Follow up to: https://github.com/zed-industries/zed/pull/19530
Release Notes:
- Improve prompt for generating title summaries, avoiding preambles
