Using the file system as a database seems like it's easy, but it's
actually a real pain. I'd like to use LMDB to store the prompts locally
so we have more control. We can always add an export option, but I want
the source of truth to be somewhere other than the file system.
So far, I have a PromptStore which is global to the application and can
be initialized on startup. Then there's a `PromptLibrary` which is
intended to be the root of a new kind of Zed window. I haven't actually
seen pixels yet, but I've sketched out the basics needed to create a new
prompt, save, etc.
Still lots to figure out but the foundations of being backed by a DB and
rendering in an independent window are in place.
/cc @iamnbutler @as-cii
Release Notes:
- N/A
---------
Co-authored-by: Antonio Scandurra <me@as-cii.com>
When running the tests for linux, I found a lot of benign errors getting
logged. This PR cuts down some of the noise from unnecessary workspace
serialization and SVG renders
Release Notes:
- N/A
Data migration plan:
- [X] Make a duplicate table of `copilot_events`
- Name: `inline_completion_events`
- Omit `suggestion_id` column
- [X-reverted-skipping] In collab, continue to match on copilot_events,
but simply stuff their data into inline_completion_events, to forward it
to the new table
- [skipping] Once collab is deployed, ensure no events are being sent to
copilot_events, migrate `copilot_events` to new table via a transaction
- [skipping] Delete `copilot_events` table
---
- [X] Locally test that copilot events sent from old clients get put
into inline_completions_table
- [X] Locally test that copilot events and supermaven events sent from
new clients get put into inline_completions_table
---
- [X] Why are discard events being spammed?
- A:
8d4315712b/crates/editor/src/editor.rs (L2147)
This will throw off the past results for accepted / dismissed that I was
wanting to use to evaluate Supermaven quality, by comparing its rate
with copilot's rate.
I'm not super thrilled with this fix, but I think it'll do. In the
`supermaven_completions_provider`, we check if there's a `completion_id`
before sending either an accepted or discard completion event. I don't
see a similar construct in the `copilot_completions_provider` to
piggyback off of, so I begrudgingly introduced
`should_allow_event_to_send` and had it follow the same pattern that
`completion_id` does. Maybe there's a better way?
---
Adds events to supermaven suggestions. Makes "CopilotEvents" generic ->
"InlineCompletionEvents".
Release Notes:
- N/A
Adds a supermaven provider for completions. There are various other
refactors amidst this branch, primarily to make copilot no longer a
dependency of project as well as show LSP Logs for global LSPs like
copilot properly.
This feature is not enabled by default. We're going to seek to refine it
in the coming weeks.
Release Notes:
- N/A
---------
Co-authored-by: Antonio Scandurra <me@as-cii.com>
Co-authored-by: Nathan Sobo <nathan@zed.dev>
Co-authored-by: Max <max@zed.dev>
Co-authored-by: Max Brunsfeld <maxbrunsfeld@gmail.com>
This PR adds support for replacing the current preview tab when using
GoToDefinition. Previously a tab, that was navigated away from, was
converted into a permanent tab and the new tab was opened as preview.
Without `enable_preview_from_code_navigation`:
https://github.com/zed-industries/zed/assets/53836821/99840724-d6ff-4738-a9c4-ee71a0001634
With `enable_preview_from_code_navigation`:
https://github.com/zed-industries/zed/assets/53836821/8c60efcb-d597-40bf-b08b-13faf5a289b6
Note: In the future I would like to improve support for the navigation
history, because right now tabs that are not "normal" project items,
e.g. FindAllReferences cannot be reopened
Release Notes:
- Added support for replacing the current preview tab when using code
navigation (`enable_preview_from_code_navigation`)
This pull request introduces the new
`ElementContext::request_autoscroll(bounds)` and
`ElementContext::take_autoscroll()` methods in GPUI. These new APIs
enable container elements such as `List` to change their scroll position
if one of their children requested an autoscroll. We plan to use this in
the revamped assistant.
As a drive-by, we also:
- Renamed `Element::before_layout` to `Element::request_layout`
- Renamed `Element::after_layout` to `Element::prepaint`
- Introduced a new `List::splice_focusable` method to splice focusable
elements into the list, which enables rendering offscreen elements that
are focused.
Release Notes:
- N/A
---------
Co-authored-by: Nathan <nathan@zed.dev>
This PR implements the preview tabs feature from VSCode.
More details and thanks for the head start of the implementation here
#6782.
Here is what I have observed from using the vscode implementation ([x]
-> already implemented):
- [x] Single click on project file opens tab as preview
- [x] Double click on item in project panel opens tab as permanent
- [x] Double click on the tab makes it permanent
- [x] Navigating away from the tab makes the tab permanent and the new
tab is shown as preview (e.g. GoToReference)
- [x] Existing preview tab is reused when opening a new tab
- [x] Dragging tab to the same/another panel makes the tab permanent
- [x] Opening a tab from the file finder makes the tab permanent
- [x] Editing a preview tab will make the tab permanent
- [x] Using the space key in the project panel opens the tab as preview
- [x] Handle navigation history correctly (restore a preview tab as
preview as well)
- [x] Restore preview tabs after restarting
- [x] Support opening files from file finder in preview mode (vscode:
"Enable Preview From Quick Open")
I need to do some more testing of the vscode implementation, there might
be other behaviors/workflows which im not aware of that open an item as
preview/make them permanent.
Showcase:
https://github.com/zed-industries/zed/assets/53836821/9be16515-c740-4905-bea1-88871112ef86
TODOs
- [x] Provide `enable_preview_tabs` setting
- [x] Write some tests
- [x] How should we handle this in collaboration mode (have not tested
the behavior so far)
- [x] Keyboard driven usage (probably need workspace commands)
- [x] Register `TogglePreviewTab` only when setting enabled?
- [x] Render preview tabs in tab switcher as italic
- [x] Render preview tabs in image viewer as italic
- [x] Should this be enabled by default (it is the default behavior in
VSCode)?
- [x] Docs
Future improvements (out of scope for now):
- Support preview mode for find all references and possibly other
multibuffers (VSCode: "Enable Preview From Code Navigation")
Release Notes:
- Added preview tabs
([#4922](https://github.com/zed-industries/zed/issues/4922)).
---------
Co-authored-by: Conrad Irwin <conrad.irwin@gmail.com>
Refs #9647
Fixes https://github.com/zed-industries/zed/issues/9792
This pull request moves the computation of scrollbar markers off the
main thread, to prevent them from grinding the editor to a halt when we
have a lot of them (e.g., when there are lots of search results on a
large file). With these changes we also avoid generating multiple quads
for adjacent markers, thus fixing an issue where we stop drawing other
primitives because we've drawn too many quads in the scrollbar.
Release Notes:
- Improved editor performance when displaying lots of search results,
diagnostics, or symbol highlights in the scrollbar
([#9792](https://github.com/zed-industries/zed/issues/9792)).
---------
Co-authored-by: Antonio Scandurra <me@as-cii.com>
Co-authored-by: Nathan <nathan@zed.dev>
This pull request introduces a new `InlineCompletionProvider` trait,
which enables making `Editor` copilot-agnostic and lets us push all the
copilot functionality into the `copilot_ui` module. Long-term, I would
like to merge `copilot` and `copilot_ui`, but right now `project`
depends on `copilot`, which makes this impossible.
The reason for adding this new trait is so that we can experiment with
other inline completion providers and swap them at runtime using config
settings.
Please, note also that we renamed some of the existing copilot actions
to be more agnostic (see release notes below). We still kept the old
actions bound for backwards-compatibility, but we should probably remove
them at some later version.
Also, as a drive-by, we added new methods to the `Global` trait that let
you read or mutate a global directly, e.g.:
```rs
MyGlobal::update(cx, |global, cx| {
});
```
Release Notes:
- Renamed the `copilot::Suggest` action to
`editor::ShowInlineCompletion`
- Renamed the `copilot::NextSuggestion` action to
`editor::NextInlineCompletion`
- Renamed the `copilot::PreviousSuggestion` action to
`editor::PreviousInlineCompletion`
- Renamed the `editor::AcceptPartialCopilotSuggestion` action to
`editor::AcceptPartialInlineCompletion`
---------
Co-authored-by: Nathan <nathan@zed.dev>
Co-authored-by: Kyle <kylek@zed.dev>
Co-authored-by: Kyle Kelley <rgbkrk@gmail.com>
See https://zed.dev/channel/gpui-536
Fixes https://github.com/zed-industries/zed/issues/9010
Fixes https://github.com/zed-industries/zed/issues/8883
Fixes https://github.com/zed-industries/zed/issues/8640
Fixes https://github.com/zed-industries/zed/issues/8598
Fixes https://github.com/zed-industries/zed/issues/8579
Fixes https://github.com/zed-industries/zed/issues/8363
Fixes https://github.com/zed-industries/zed/issues/8207
### Problem
After transitioning Zed to GPUI 2, we started noticing that interacting
with the mouse on many UI elements would lead to a pretty annoying
flicker. The main issue with the old approach was that hover state was
calculated based on the previous frame. That is, when computing whether
a given element was hovered in the current frame, we would use
information about the same element in the previous frame.
However, inspecting the previous frame tells us very little about what
should be hovered in the current frame, as elements in the current frame
may have changed significantly.
### Solution
This pull request's main contribution is the introduction of a new
`after_layout` phase when redrawing the window. The key idea is that
we'll give every element a chance to register a hitbox (see
`ElementContext::insert_hitbox`) before painting anything. Then, during
the `paint` phase, elements can determine whether they're the topmost
and draw their hover state accordingly.
We are also removing the ability to give an arbitrary z-index to
elements. Instead, we will follow the much simpler painter's algorithm.
That is, an element that gets painted after will be drawn on top of an
element that got painted earlier. Elements can still escape their
current "stacking context" by using the new `ElementContext::defer_draw`
method (see `Overlay` for an example). Elements drawn using this method
will still be logically considered as being children of their original
parent (for keybinding, focus and cache invalidation purposes) but their
layout and paint passes will be deferred until the currently-drawn
element is done.
With these changes we also reworked geometry batching within the
`Scene`. The new approach uses an AABB tree to determine geometry
occlusion, which allows the GPU to render non-overlapping geometry in
parallel.
### Performance
Performance is slightly better than on `main` even though this new
approach is more correct and we're maintaining an extra data structure
(the AABB tree).
Release Notes:
- Fixed a bug that was causing popovers to flicker.
---------
Co-authored-by: Nathan Sobo <nathan@zed.dev>
Co-authored-by: Thorsten <thorsten@zed.dev>
We were seeing panics due to callers assuming they had valid
excerpt_ids, but that cannot easily be guaranteed across await points as
anyone may remove an excerpt.
Release Notes:
- Fixed a panic when hovering in a multibuffer
Before this change, enabling LSP trace checkbox closed the panel and
toggled the server logs on.
Now, the newly enabled trace logs are shown instead.
Release Notes:
- Improved LSP logs checkbox behavior
This PR adds **internal** ability to run arbitrary language servers via
WebAssembly extensions. The functionality isn't exposed yet - we're just
landing this in this early state because there have been a lot of
changes to the `LspAdapter` trait, and other language server logic.
## Next steps
* Currently, wasm extensions can only define how to *install* and run a
language server, they can't yet implement the other LSP adapter methods,
such as formatting completion labels and workspace symbols.
* We don't have an automatic way to install or develop these types of
extensions
* We don't have a way to package these types of extensions in our
extensions repo, to make them available via our extensions API.
* The Rust extension API crate, `zed-extension-api` has not yet been
published to crates.io, because we still consider the API a work in
progress.
Release Notes:
- N/A
---------
Co-authored-by: Marshall <marshall@zed.dev>
Co-authored-by: Nathan <nathan@zed.dev>
Co-authored-by: Marshall Bowers <elliott.codes@gmail.com>
I would like to keep diagnostics open on one side, and process them on
the other.
Release Notes:
- Added `editor::OpenExcerptsSplit` (bound to `cmd-k enter`) to open the
selected excerpts in the adjacent pane
- vim: Added `ctrl-w d`, `ctrl-w shift-d` and `ctrl-w space` for
`editor::GoTo{,Type}Definition` and `editor::OpenExcerptsSplit`
This PR adds the initial support for loading extensions in Zed.
### Extensions Directory
Extensions are loaded from the extensions directory.
The extensions directory has the following structure:
```
extensions/
installed/
extension-a/
grammars/
languages/
extension-b/
themes/
manifest.json
```
The `manifest.json` file is used internally by Zed to keep track of
which extensions are installed. This file should be maintained
automatically, and shouldn't require any direct interaction with it.
Extensions can provide Tree-sitter grammars, languages, and themes.
Release Notes:
- N/A
---------
Co-authored-by: Marshall <marshall@zed.dev>
Based on the great work in
https://github.com/zed-industries/zed/pull/7130 , now sends this data
```
[crates/lsp/src/lsp.rs:588] ClientInfo { name: name.to_string(), version: Some(version.to_string()) } = ClientInfo {
name: "Zed Dev",
version: Some(
"0.122.0",
),
}
```
with every LSP server initialization.
Release Notes:
- Added Zed name and version to LSP InitializeParams requests
This PR renames the `h_stack` and `v_stack` to `h_flex` and `v_flex`,
respectively.
We were previously using `h_stack` and `v_stack` to match SwiftUI, but
`h_flex` and `v_flex` fit better with the web/flexbox terminology that
the rest of GPUI uses.
Additionally, we were already calling the utility functions used to
implement `h_stack` and `v_stack` by the new names.
Release Notes:
- N/A
This commit mostly fixes invalid URLs in docstrings. It also
encapsulates crates we reexport (serde stuff + linkme) into a public
module named "private" in order to reduce the API surfaced through docs.
Moreover, I fixed up a bunch of crates that were pulling serde_json in
through gpui explicitly instead of using Cargo manifest.
Release Notes:
- N/A
This PR fixes an issue where certain tabs were not using the correct
color for their labels when they were inactive.
Release Notes:
- Fixed an issue where some inactive tabs were not using the correct
label color.
This is a PR I built for a friend of a friend at StrangeLoop, who is
making a much better LSP for elixir that elixir folks want to experiment
with. This PR also improves the our debug log viewer to handle LSP
restarts.
TODO:
- [ ] Make sure NextLS binary loading works.
Release Notes:
- Added support for the experimental Next LS for Elxir, to enable it add
the following field to your settings to enable:
```json
"elixir": {
"next": "on"
}
```
Copilot sends multiple events about its LSP server readiness, not necessarily recreating the server from scratch (e.g. due to re-sign in action).
Avoid re-adding same log subscriptions on the same LSP server, which
causes panics.
This is still WIP, mostly pending styling. I added a pretty rudimentary
text field and no buttons whatsoever other than that. I am targeting a
Preview of 09.13, as I am gonna be on PTO for the next week.
I dislike the current implementation slightly because of `regex`'s crate
syntax and lack of support of backreferences. What strikes me as odd wrt
to syntax is that it will just replace a capture name with empty string
if that capture is missing from the regex. While this is perfectly fine
behaviour for conditionally-matched capture groups (e.g. `(foo)?`), I
think it should still error out if there's no group with a given name
(conditional or not).
Release Notes:
- Added "Replace" functionality to buffer search.
