After a lot of back-and-forth, this is a small attempt to implement
solutions (1) and (3) in
https://github.com/zed-industries/zed/issues/7902. The goal is to have a
minimal change that helps users get started with Zed, until we have
extensions ready.
Release Notes:
- Added detection of user-installed `gopls` to Go language server
adapter. If a user has `gopls` in `$PATH` when opening a worktree, it
will be used.
- Added detection of user-installed `zls` to Zig language server
adapter. If a user has `zls` in `$PATH` when opening a worktree, it will
be used.
Example:
I don't have `go` installed globally, but I do have `gopls`:
```
~ $ which go
go not found
~ $ which gopls
/Users/thorstenball/code/go/bin/gopls
```
But I do have `go` in a project's directory:
```
~/tmp/go-testing φ which go
/Users/thorstenball/.local/share/mise/installs/go/1.21.5/go/bin/go
~/tmp/go-testing φ which gopls
/Users/thorstenball/code/go/bin/gopls
```
With current Zed when I run `zed ~/tmp/go-testing`, I'd get the dreaded
error:
But with the changes in this PR, it works:
```
[2024-02-23T11:14:42+01:00 INFO language::language_registry] starting language server "gopls", path: "/Users/thorstenball/tmp/go-testing", id: 1
[2024-02-23T11:14:42+01:00 INFO language::language_registry] found user-installed language server for Go. path: "/Users/thorstenball/code/go/bin/gopls", arguments: ["-mode=stdio"]
[2024-02-23T11:14:42+01:00 INFO lsp] starting language server. binary path: "/Users/thorstenball/code/go/bin/gopls", working directory: "/Users/thorstenball/tmp/go-testing", args: ["-mode=stdio"]
```
---------
Co-authored-by: Antonio <antonio@zed.dev>
Before:
```
[2024-02-21T18:55:55+02:00 INFO language::language_registry] starting language server "eslint", path: "/Users/someonetoignore/Downloads/eslint-configs-demo", id: 2
[2024-02-21T18:55:56+02:00 INFO lsp] 2 unhandled notification window/logMessage:
{
"type": 3,
"message": "ESLint server running in node v18.15.0"
}
[2024-02-21T18:55:56+02:00 INFO lsp] 2 unhandled notification eslint/confirmESLintExecution:
{
"scope": "local",
"uri": "file:///Users/someonetoignore/Downloads/eslint-configs-demo/index.js",
"libraryPath": "/Users/someonetoignore/Downloads/eslint-configs-demo/node_modules/eslint/lib/api.js"
}
```
After:
```
[2024-02-21T18:57:31+02:00 INFO language::language_registry] starting language server "eslint", path: "/Users/someonetoignore/Downloads/eslint-configs-demo", id: 2
[2024-02-21T18:57:32+02:00 INFO lsp] Language server with id 2 sent unhandled notification window/logMessage:
{
"type": 3,
"message": "ESLint server running in node v18.15.0"
}
[2024-02-21T18:57:32+02:00 INFO project::prettier_support] Fetching default prettier and plugins: [("prettier-plugin-tailwindcss", "0.5.11"), ("prettier", "3.2.5")]
[2024-02-21T18:57:32+02:00 INFO lsp] Language server with id 2 sent unhandled notification eslint/confirmESLintExecution:
{
"scope": "local",
"uri": "file:///Users/someonetoignore/Downloads/eslint-configs-demo/index.js",
"libraryPath": "/Users/someonetoignore/Downloads/eslint-configs-demo/node_modules/eslint/lib/api.js"
}
```
We have to pass a name there too, but the problem here is that the
unhandled message callback is created very early, along with the binary,
but the server name is received from the LSP initialize response, which
is a totally separate piece of code.
I plan to refactor that code next, but so far, improve the logs at least
slightly.
Release Notes:
- N/A
This should help us debug more failures because we can now see what
exactly was started.
Release Notes:
- N/A
Co-authored-by: Nathan <nathan@zed.dev>
Co-authored-by: Max <max@zed.dev>
Zed's LSP support expects request messages to have integer ID, Metals
LSP uses string. According to specification, both is acceptable:
interface RequestMessage extends Message {
/**
* The request id.
*/
id: integer | string;
...
This pull requests modifies the types and serialization/deserialization
so that string IDs are accepted.
Release Notes:
- Make Zed LSP request ids compliant to the LSP specification
The current LSP headers reader implementation assumes a specific order
(i.e., `Content-Length` first, and then `Content-Type`). Unfortunately,
this assumption is not always valid, as no specification enforces the
rule. @caius and I encountered this issue while implementing the
Terraform LSP, where `Content-Type` comes first, breaking the
implementation in #6929.
This PR introduces a `read_headers` function, which asynchronously reads
the incoming pipe until the headers' delimiter (i.e., '\r\n\r\n'),
adding it to the message buffer, and returning an error when delimiter's
not found.
I added a few tests but only considered scenarios where headers are
delivered at once (which should be the case?). I'm unsure if this
suffices or if I should consider more scenarios; I would love to hear
others' opinions.
Release Notes:
- N/A
---------
Co-authored-by: Caius <caius@caius.name>
Previously we would run these loops indefinitely when a language server
closed its stdout/stderr and the `read_until` returned `0` bytes read.
Easy to reproduce: start Zed with LSP attached, `kill -9` the LSP, see
logs accumulate.
Release Notes:
- Fix high CPU usage when a language server crashes (or closes its
stdout/stderr on purpose).
Co-authored-by: Julia <julia@zed.dev>
Co-authored-by: Mikayla <mikayla@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 sorts the dependency lists in our `Cargo.toml` files so that
they are in alphabetical order.
This should make them easier to visually scan when looking for a
dependency.
Apologies in advance for any merge conflicts 🙈
Release Notes:
- N/A
- [x] Fill in GPL license text.
- [x] live_kit_client depends on live_kit_server as non-dev dependency,
even though it seems to only be used for tests. Is that an issue?
Release Notes:
- N/A
Before this change, we would send requests to language servers without
canceling them even if we never wait for their response.
Example: when doing document highlights, we'd send a request (modulo
debouncing) whenever a change was made, ignoring previously sent
requests that might still be in-flight.
With this change, we now send a Cancel request (from the LSP spec) to
the language server in case no one listens to the response anymore
(which is what happens when the `Future` returned by `request_internal`)
is dropped.
Adding the typos crate to our CI will take some doing, as we have
several tests which rely on typos in various ways (e.g. checking state
as the user types), but I thought I'd take a first stab at fixing what
it finds.
Release Notes:
- N/A
This PR cleans up a handful of references in doc comments in the `lsp`
crate so that `rustdoc` will link and display them correctly.
Release Notes:
- N/A
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.
Tailwind likes to throw a lot of completion data at us, this gets it to
send less. Previously it would respond to a completion with 2.5 MB JSON
blob, now it is more like 0.8 MB.
Relies on a local copy of lsp-types with the `itemDefaults` field added.
I don't have write perms to push to our fork of the crate atm, sorry :)
