Turns out that the existing CI step for Nightly did create the licenses
and they have been baked into X86 builds ever since, because our
builders are stateful.
On ARM machines, the licenses wouldn't exist in the binary because we
called `script/generate-licenses` too late in `scripts/bundle-linux`,
after the binary had been created.
This removes the duplication and generates the licenses once, before the
binary is created.
Fixes#14302.
Release Notes:
- Fixed "View Dependency Licenses" (or `zed: open licenses`) crashing on
Linux ARM machines.
([#14302](https://github.com/zed-industries/zed/issues/14302)
This restores https://github.com/zed-industries/zed/pull/13943 which was
reverted in #13974 because it was possible to get in a state where focus
could not be restored on a window.
In this PR there's an additional change: `FocusIn` and `FocusOut` events
are always handled, even if the `event.mode` is not "NORMAL". In my
testing, `alt-tabbing` between windows didn't produce `FocusIn` and
`FocusOut` events when we had that check. Now, with the check removed,
it's possible to switch focus between two windows again with `alt-tab`.
Release Notes:
- N/A
---------
Co-authored-by: Conrad <conrad@zed.dev>
Co-authored-by: Conrad Irwin <conrad.irwin@gmail.com>
This reverts #13943 and reopens#13897 since the fix in #13943 comes
with a regression:
Sometimes Zed loses keyboard focus and can't be restored. I haven't
figured out yet exactly when and how this happens and can't reliably
reproduce it yet, but there's something off with focus handling.
One reliable way to reproduce _one_ of the problems:
1. Open two zed windows
2. Focus one Zed window
3. Hover with the mouse over the other
4. Try to type in the window that should still be focused
So, to be careful, I'm going to revert the PR first, since I couldn't
find an obvious fix yet. If we do find a fix, we can unrevert.
Release Notes:
- N/A
This addresses the question in [this
comment](https://github.com/zed-industries/zed/issues/5830#issuecomment-2211942554)
by adding support for `classAttributes` to the settings.
Meaning that the following Zed `settings.json` now works:
```jsonc
{
"lsp": {
"tailwindcss-language-server": {
"settings": {
"classAttributes": [
"class",
"className",
"ngClass",
// add styles so will give intellisense to styles constant.
"styles"
]
// Optional:
// "includeLanguages": {
// "erb": "html",
// "ruby": "html"
// },
// "experimental": {
// "classRegex": ["\\bclass:\\s*['\"]([^'\"]*)['\"]"]
// }
}
}
}
}
```
Release Notes:
- Added support for setting `classAttributes` in the configuration for
`tailwindcss-language-server`. Example: `{ "lsp": {
"tailwindcss-language-server": { "settings": { "classAttributes": [
"class", "className", "ngClass", "styles" ] } } } }`
Introduces the `ZED_WINDOW_DECORATIONS` env variable.
- Not set, defaulting to client-side decorations
- Value is "client": client-side decorations
- Value is "server": server-side decorations
I think it's good to have this escape-hatch next to all possible
detection mechanisms.
Release Notes:
- N/A
This fixes the issue of a tooltip covering the thing that the button has
revealed.
It also mirrors what other UI frameworks do. Chrome on Linux behaves the
same, and Safari does the same thing on macOS.
It fixes this:
Release Notes:
- N/A
This has been bugging me for a while. If you create a new file and then
save it, the dialogue would show the home directory and not the folder
that you were in.
This fixes it.
Release Notes:
- N/A
With the new window decorations resizing was _really_ laggy on my X11
machine.
Before:
- Click on window border (hitbox doesn't work properly, but that's
another issue)
- Drag and resize
- 4-5s nothing happens
- Window is resized
After:
- Click on window border
- Drag and resize
- Window is resized
I'm still not 100% sure on why this happens on my machine and not
Conrad's/Mikayla's, but seems like that GTK_EDGE_CONSTRAINTS atom is
sent when resizing.
The other thing that I can't explain is that we get a `ConfigureNotify`
when resizing, with the right size, but maybe not often enough?
Anyway, for now we'll go with this.
Release Notes:
- N/A
This changes the implementation of the X11 client to use `mio`, as a
polling mechanism, and a custom run loop instead of `calloop` and its
callback-based approach.
We're doing this for one big reason: more control over how we handle
events.
With `calloop` we don't have any control over which events are processed
when and how long they're processes for. For example: we could be
blasted with 150 input events from X11 and miss a frame while processing
them, but instead of then drawing a new frame, calloop could decide to
work off the runnables that were generated from application-level code,
which would then again cause us to be behind.
We kinda worked around some of that in
https://github.com/zed-industries/zed/pull/12839 but the problem still
persists.
So what we're doing here is to use `mio` as a polling-mechanism. `mio`
notifies us if there are X11 on the XCB connection socket to be
processed. We also use its timeout mechanism to make sure that we don't
wait for events when we should render frames.
On top of `mio` we now have a custom run loop that allows us to decide
how much time to spend on what — input events, rendering windows, XDG
events, runnables — and in what order we work things off.
This custom run loop is consciously "dumb": we render all windows at the
highest frame rate right now, because we want to keep things predictable
for now while we test this approach more. We can then always switch to
more granular timings. But considering that our loop runs and checks for
windows to be redrawn whenever there's an event, this is more an
optimization than a requirement.
One reason for why we're doing this for X11 but not for Wayland is due
to how peculiar X11's event handling is: it's asynchronous and by
default X11 generates synthetic events when a key is held down. That can
lead to us being flooded with input events if someone keeps a key
pressed.
So another optimization that's in here is inspired by [GLFW's X11 input
handling](https://github.com/glfw/glfw/blob/b35641f4a3c62aa86a0b3c983d163bc0fe36026d/src/x11_window.c#L1321-L1349):
based on a heuristic we detect whether a `KeyRelease` event was
auto-generated and if so, we drop it. That essentially halves the amount
of events we have to process when someone keeps a key pressed.
Release Notes:
- N/A
---------
Co-authored-by: Conrad Irwin <conrad.irwin@gmail.com>
Co-authored-by: Conrad <conrad@zed.dev>
Took me a while to figure out that I can't run
cargo run -p gpui --example animation
and that it has to run in the `gpui` crate.
So I thought I'd fix this.
Release Notes:
- N/A
This is a small change that aims to address frames being dropped when we
get a ton of X11 input events.
What it does, in short, is to first read all X11 input events and then
prioritize the rendering.
In my testing, it causes less frames to be dropped when the system is
under heavy load and lots of input events are being created.
Release Notes:
- N/A
This reverts commit f69c8ca74e after it
has already been partially reverted in
https://github.com/zed-industries/zed/pull/13458.
Why the revert?
The changes in that commit/PR fix one type of problem — dropping of
frames when being blasted with input events — but trades it for another
one that I can't explain yet: when the system is under load, then input
becomes _laggy_ and input events seem to be delayed.
Two examples of how that shows up:
1. When the system is under load* and you hold down the `down` key to
scroll, then lift the finger, the cursor stops sometimes. If you then
produce another input event by jiggling the mouse cursor you'll see more
`down`-key events coming up and the cursor moving down. It feels as if
the event loop is not being woken up even though there are still events.
I suspect it might have something to do with XIM, because if it's
disabled, it seems as if problems become less severe.
2. When the system is under load* and you click-and-drag a selection in
the editor, you can see how the selection is delayed and takes 500ms-1s
to catch up to where the cursor is.
* system under load: start Zed, then in another terminal window create a
release build of Zed, for example.
With the changes reverted, the failure mode looks different: we skip
frames. But that, I think, is the better of two bad options, because
skipping frames means that you see what's happening vs. input events
seemingly still coming in seconds after you stopped using the keyboard.
Release Notes:
- N/A
This fixes#4432 by ensuring that we scan & watch the `.zed` folder,
just like we watch the `.git`, for changes.
Release Notes:
- Settings are now loaded from local `.zed/settings.json` files even if
they are `.gitignore`d.
([#4432](https://github.com/zed-industries/zed/issues/4432)).
Co-authored-by: Bennet <bennet@zed.dev>
This change ensures that we always render a window according to its
refresh rate, even if there are a lot of X11 events.
We're working around some limitations of `calloop`. In the future, we
think we should revisit how the event loop is implemented on X11, so
that we can ensure proper prioritization of input events vs. rendering.
Release Notes:
- N/A
Co-authored-by: Antonio <me@as-cii.com>
Our dev builds don't have updates and will never have updates, so
instead of polling our servers every time we start a dev instance, let's
disable it for the dev channel.
Release Notes:
- N/A
This adds two new actions to `editor`:
- `editor::SelectPageUp`
- `editor::SelectPageDown`
On Linux they're bound by default to `shift-pageup` and
`shift-pagedown`, which matches VS Code and JetBrains.
Release Notes:
- N/A
This matches the behavior of VS Code and JetBrains.
(Of course I implemented both actions myself before figuring out that we
already have actions to scroll a line up and down.)
Release Notes:
- N/A
This fixes `ctrl-w` and `ctrl-e` not working in the terminal pane but
instead triggering Zed actions ("close pane" and "search project files"
respectively).
I've added both because I think they're pretty commonly used in
terminals, since they're default Emacs-style keybindings.
But I also didn't want to add more, since it's relatively easy for users
to define themselves which keybindings should be forwarded to the
terminal and which not.
All that's required is adding something like this to the keymap:
```json
{
"context": "Terminal",
"bindings": {
"ctrl-n": ["terminal::SendKeystroke", "ctrl-n"],
"ctrl-p": ["terminal::SendKeystroke", "ctrl-p"]
}
}
```
cc @mikayla-maki
Release Notes:
- N/A
We saw this panic come up:
```
called `Result::unwrap()` on an `Err` value: IoError(Custom { kind: Other, error: UnknownError })
core::panicking::panic_fmt
core::result::unwrap_failed
<gpui::platform::linux::x11::window::X11Window as core::ops::drop::Drop>::drop
core::ptr::drop_in_place<gpui::platform::linux::x11::window::X11Window>
core::ptr::drop_in_place<gpui::window::Window>
gpui::app::AppContext::shutdown
gpui::app::AppContext::new::{{closure}}
gpui::platform::linux::platform::<impl gpui::platform::Platform for P>::run
gpui::app::App::run
zed::main
std::sys_common::backtrace::__rust_begin_short_backtrace
std::rt::lang_start::{{closure}}
std::rt::lang_start_internal
main
__libc_start_call_main
__libc_start_main_impl
_start
```
I'm not sure where exactly that error comes from, except from the X11
stuff. So let's be defensive and log error and only then tear down
everything.
I _think_ that if the error is repeatable that means we won't close the
window but instead just log errors, but I do think that's better than
panicking right now.
Release Notes:
- N/A
zbus, naga, and some parts of blade are pretty noisy at the INFO level.
zbus especially dumps large debug dumps into the logs.
So on Linux, when logging to a file, we reduce that noise. That means
one still gets the full firehose when doing `RUST_LOG=info cargo run`,
but not in the logs.
Release Notes:
- N/A
This changes the breadcrumb header from "untitled" to "Last 1000 lines
in <location of log file>".
Reason is that it's been incredibly frustrating not seeing the location
of the log file there and not seeing that it's actually a truncated
version of the logs.
This is one remedy for that.
Other options considered:
1. Opening the actual log file. Turns out that is huge. On Linux right
now it's 5 megabyte after 5 minutes.
2. Opening the file and adding it on the buffer. That is tricky and
weird, because you have to modify the underlying buffer and set the
file, after having to add it to the workspace and getting its entry,
etc.
3. Setting a `display_file_path` on Buffer. That would require also
adding it on `BufferSnapshot` and then threading that through so that it
gets returned by the multi-buffer and then in the editor. And ultimately
this here is a "view concern", so we thought we just add it as such.
So yes, not the best change possible, but it's not that invasive and
makes it clear that it's a view-only concern.
Release Notes:
- N/A
Co-authored-by: Kirill <kirill@zed.dev>
This fixes everything but the main Zed window (GPUI examples, prompt
library, etc.) not being closable by clicking on the X in X11.
We had a dangling reference before: we would remove the window from the
X11 state, but GPUI itself would still have the window in its
references.
In order to fix this we have to call `window.close()`, which ends up
calling `cx.remove_window()`, which removes the reference.
That in turn then causes the reference to be dropped, which cleans up
the X11 state for the window.
Release Notes:
- N/A
Turns out we still get FocusOut and UnmapNotify events after the window
has been destroyed, which resulted in error messages popping up because
we can't find the window anymore that we want to mark as unfocused.
Release Notes:
- N/A
I noticed that when I use my mouse wheel, we get a ton of the
`XkbStateNotify` events, but the modifiers don't change, so we add a ton
of useless input events for the window.
Release Notes:
- N/A
This helps with the problem of keyboard input feeling laggy when the
event loop is under load.
What would previously happen is:
- N events from X11 arrive
- N events get forwarded to XIM
- N events are handled in N iterations of the event loop (sadly, yes: we
only seem to be getting back one `ClientMessage` per poll from XCB
connection)
- Each event is pushed into the channel
- N event loop iterations are needed to get the events off the channel
and handle them
With this change, we get rid of the last 2 steps: instead of pushing the
event onto a channel, we store it on the XIM handler itself, and then
work it off synchronously.
Usually one shouldn't block the event loop, but I think in this case -
user input! - it's better to handle the events directly instead of
re-enqueuing them again in a channel, where they can accumulate and need
multiple iterations of the loop to be worked off.
This does *not* fix the problem of input feeling choppy/slower when the
system is under load, but it makes the behavior now feel exactly the
same as when XIM is disabled.
I also think the code is easier to understand since it's more
straightforward.
Release Notes:
- N/A
This change ensures that the event loop prioritizes enqueueing another
render or handling user input over executing runnables.
It's a subtle change as a result of a week of digging into performance
on X11. It's also not perfect: ideally we'd get rid of the intermediate
channel here and had more control over when and how we run runnables vs.
X11 events, but I think short of rewriting how we use an event loop,
this is good cost/benefit change.
To illustrate:
Before this change, it was possible to block the app from rendering for
a long time by just creating a ton of futures that were executed on the
"main" thread (we don't have a "main" thread on Linux, but we have a
single thread in which we run the event loop).
That was relatively easy to reproduce by opening the `zed` repository
and starting `rust-analyzer`: at some point `rust-analyzer` sends us so
many notifications, that are all handled in futures, that the event loop
is busy just working off the runnables, never getting to the events that
X11 sends us or our own timer to re-enqueue another render.
When you put print statements into the code to show when which event was
handled, you'd see something like this **before this change**:
```
[ ... hundreds of runnable.run() ... ]
runnable.run()
runnable.run()
runnable.run()
runnable.run()
runnable.run()
runnable.run()
runnable.run()
runnable.run()
runnable.run()
runnable.run()
runnable.run()
runnable.run()
runnable.run()
runnable.run()
runnable.run()
runnable.run()
runnable.run()
new render tick timer. lag: 56.942049ms
X11 event
new render tick timer. lag: 9.668µs
X11 event
new render tick timer. lag: 9.955µs
X11 event
runnable.run()
runnable.run()
runnable.run()
runnable.run()
new render tick timer. lag: 12.462µs
X11 event
new render tick timer. lag: 14.868µs
X11 event
new render tick timer. lag: 11.234µs
X11 event
new render tick timer. lag: 11.681µs
X11 event
new render tick timer. lag: 13.926µs
X11 event
```
Note the `lag: 56ms`: that's the difference between when we wanted to
execute the callback that enqueues another render and when it ran.
Longer lags are possible, this is just the first one I grabbed from the
logs.
Now, compare this with the logs **after this change**:
```
runnable.run()
runnable.run()
runnable.run()
runnable.run()
runnable.run()
runnable.run()
runnable.run()
runnable.run()
runnable.run()
runnable.run()
runnable.run()
runnable.run()
runnable.run()
runnable.run()
runnable.run()
runnable.run()
runnable.run()
runnable.run()
runnable.run()
runnable.run()
runnable.run()
runnable.run()
new render tick timer. lag: 36.051µs
runnable.run()
runnable.run()
runnable.run()
runnable.run()
runnable.run()
runnable.run()
runnable.run()
runnable.run()
runnable.run()
X11 event
runnable.run()
runnable.run()
runnable.run()
runnable.run()
runnable.run()
runnable.run()
runnable.run()
runnable.run()
runnable.run()
runnable.run()
runnable.run()
runnable.run()
runnable.run()
runnable.run()
```
In-between many `runnable.run()` we'll always handle events.
So, in essence, what this change does is to introduce 2 priorities into
the X11 event queue:
- high: X11 events (user events, render events, ...), render tick, XIM
events, ...
- low: all async rust code
I've tested this with a debug build and release build and I think the
app now feels more responsive. It doesn't feel perfect still, especially
in the slow debug builds, but I couldn't observe 10s lockups anymore.
Since it's a pretty small change, I think we should go for it and see
how it behaves.
Thanks to @maan2003 this now also includes the same change to Wayland.
Release Notes:
- N/A
---------
Co-authored-by: maan2003 <manmeetmann2003@gmail.com>
This fixes#12728 as much as I can tell.
The problem was that inside ERB files, when inside Ruby code, we didn't
treat `-` as part of the word, which broke completions.
So, with the change in here, and the following Zed settings, it works.
```json
{
"languages": {
"Ruby": {
"language_servers": ["tailwindcss-language-server", "solargraph"]
}
},
"lsp": {
"tailwindcss-language-server": {
"settings": {
"includeLanguages": {
"erb": "html",
"ruby": "html"
},
"experimental": {
"classRegex": ["\\bclass:\\s*['\"]([^'\"]*)['\"]"]
}
}
}
}
```
This enabled `tailwindcss-language-server` for Ruby files and tells the
language server to look for classes inside `class: ""` strings.
See demo video.
Release Notes:
- Fixed `tailwindcss-language-server` not being activated inside Ruby
strings (inside `.erb`)
([#12728](https://github.com/zed-industries/zed/issues/12728)).
Demo video:
https://github.com/zed-industries/zed/assets/1185253/643343b4-d64f-4c4e-98a1-d10df0b24e31
Co-authored-by: Max Brunsfeld <max@zed.dev>
On X11 I was unable to type ä ü and other umlauts in files with
autoclose enabled, because typing ä requires me to hit
- compose key
- `"`
- `a`
When the `"` was typed, Zed would insert a matching `"` because it had a
selection around the dead-key that was inserted by the compose key.
We ran into a similar issue in #7611, but in the case of the Brazilian
keyboard, the `"` is the compose key so we didn't trigger the matching
`"`, because we didn't have a selection yet.
What this does is it fixes the issue by making the
surround-selection-with-quotes-or-brackets also depend on the autoclose
settings, which is didn't do before. This is a breaking change for users
of a Brazilian keyboard layout in which `"` cannot be used to surround
an existing selection with quotes anymore.
That _might_ be a change that users notice, but I can't think of
scenario for that where the user wants, say, `"` to be NOT autoclosed,
but work with selections. (Example is Markdown, for which autoclose for
`"` is disabled. Do we want that but allow surrounding with quotes?)
So it fixes the issue and makes the behavior slightly more consistent,
in my eyes.
Release Notes:
- Changed the behavior of surrounding selections with brackets/quotes to
also depend on the auto-close settings of the language. This is a
breaking change for users of a Brazilian keyboard layout in which `"`
cannot be used to surround an existing selection with quotes anymore.
Before:
[Screencast from 2024-06-04
11-49-51.webm](https://github.com/zed-industries/zed/assets/1185253/6bf255b5-32e9-4ba7-8b46-1e49ace2ba7c)
After:
[Screencast from 2024-06-04
11-52-19.webm](https://github.com/zed-industries/zed/assets/1185253/3cd196fc-20ba-465f-bb54-e257f7f6d9f3)
This fixes#11236 by ignoring the `bounds.origin` values when the window
is only being resized.
The cause for the issue was that the `ConfigureNotify` event would
contain "wrong" values when the window was being resized (by dragging a
corner).
In my case it would *always* contain x:14/y:49, which is I think might
map to the origin of the top bar in GNOME.
We would then persist these wrong values when serializing the workspace.
On restart, we'd use these values and end up with the window decorations
in the wrong place.
What I still don't know:
1. What exactly the 14/49 map to, because it's not the origin of the top
bar in GNOME. I also tried the X11 TranslateCoordinates call but
couldn't get meaningful results back (even taking scale factor into
account).
2. Why the window decorations end up looking wrong vs. the window being
in the first place. But if you look at my screenshot in #11236, it looks
like the decorations are off exactly by 14/49px.
That being said, I think the solution here is a good one for now: we
don't do an additional X11 call and when we're resizing, we're not
interested in the origin changing.
Release Notes:
- N/A
Proof:
[Screencast from 2024-06-03
15-08-36.webm](https://github.com/zed-industries/zed/assets/1185253/90efccfc-8ec6-42d2-8380-1625eff57805)
I don't know whether there are any hard UI guidelines that dictate
whether this should be allowed or not, but I think it's very handy and
missed it.
I also think it makes sense to have this in a directory-centric editor
in which opening a directory creates a new window.
Release Notes:
- Added ability to create directory in open-file dialog on macOS.
When you press `V` to go into visual-line mode in Vim,
`selections.line_mode` is true and the selection contains _lines_.
But `$ZED_SELECTED_TEXT` always contained just the cursor location or
any non-line-mode selection that was previously made.
Release Notes:
- Fixed `$ZED_SELECTED_TEXT` variable in Tasks ignoring whether
visual-line-mode in Vim was used.
This fixes#4529 by allowing unsaved buffers to be formatted with
prettier.
Steps to do that:
1. Create a new buffer
2. Set language for the buffer (e.g.: `language selector: toggle` and
JSON)
3. In settings, set prettier parser for language (can't be inferred,
since we don't have filename) and allow formatting with prettier:
```json
{
"languages": {
"JSON": {
"prettier": {
"allowed": true,
"parser": "json"
}
}
}
}
```
4. Use `editor: format`
Release Notes:
- Added ability to format unsaved buffers with Prettier. Requirement is
to set a Prettier parser in the user settings. Example for JSON: `{
"languages": { "JSON": { "prettier": { "allowed": true, "parser": "json"
} } } }` ([#4529](https://github.com/zed-industries/zed/issues/4529)).
Demo:
https://github.com/zed-industries/zed/assets/1185253/d24e490b-2e2c-4a5d-95a8-fc8675523780
This fixes#10224 by handling `client/unregisterCapability` requests
that have a `workspace/didChangeWatchedFiles` method.
While debugging the issue, I found out that `gopls` seems to block
indefinitely when there's no reply to the `client/unregisterCapability`
request. Even an empty response would fix the issue.
Seems like gopls 15.x and later seem to handle nested subfolders well,
but do not handle unanswered requests.
Instead of replying with an empty response, I decided to change how we
handle file watching and keep a list of all registered paths so that we
can then unregister paths and recreate the glob patterns.
Release Notes:
- Fixed `gopls` not working correctly when the `go.mod` file was in a
subfolder and not the root folder of the project opened in Zed.
([#10224](https://github.com/zed-industries/zed/issues/10224)).
This adds the ability to configure the `tailwindcss-language-server`
integration to use a custom build of the server.
Example configuration in Zed `settings.json`:
```json
{
"lsp": {
"tailwindcss-language-server": {
"binary": {
"arguments": [
"/Users/username/tailwindcss-intellisense/packages/tailwindcss-language-server/bin/tailwindcss-language-server",
"--stdio"
]
}
}
}
}
```
This will cause Zed to use its own Node version and run it with the
given arguments.
**Note**: you need to provide `--stdio` as the second argument!
It's also possible to use a custom Node binary:
```json
{
"lsp": {
"tailwindcss-language-server": {
"binary": {
"path": "/Users/username/bin/my-node",
"arguments": [
"/Users/username/tailwindcss-intellisense/packages/tailwindcss-language-server/bin/tailwindcss-language-server",
"--stdio"
]
}
}
}
}
```
This is *super handy* when debugging the language server.
Release Notes:
- Added ability to configure own build of `tailwindcss-language-server`
in Zed settings. Example:
`{"lsp":{"tailwindcss-language-server":{"binary":{"arguments":["/absolute/path/to/tailwindcss-language-server/bin/tailwindcss-language-server",
"--stdio" ]}}}}`
Without this, we'd get constant errors when typing something with ESLint
enabled:
[2024-05-16T10:32:30+02:00 WARN project] Generic lsp request to node
failed: Request textDocument/codeAction failed with message: Cannot read
properties of undefined (reading 'disableRuleComment')
[2024-05-16T10:32:30+02:00 ERROR util]
crates/project/src/project.rs:7023: Request textDocument/codeAction
failed with message: Cannot read properties of undefined (reading
'disableRuleComment')
[2024-05-16T10:32:31+02:00 WARN project] Generic lsp request to node
failed: Request textDocument/codeAction failed with message: Cannot read
properties of undefined (reading 'disableRuleComment')
[2024-05-16T10:32:31+02:00 ERROR util]
crates/project/src/project.rs:7023: Request textDocument/codeAction
failed with message: Cannot read properties of undefined (reading
'disableRuleComment')
This is fixed by changing the default settings for ESLint language
server to have those fields.
I don't think we need to make these configurable yet. These are defaults
that multiple other plugins also use:
- vscode-eslint:
https://sourcegraph.com/github.com/microsoft/vscode-eslint@4d9fc40e71c403d359beaccdd4a6f8d027031513/-/blob/client/src/client.ts?L702-703
- nvim-lspconfig:
https://sourcegraph.com/github.com/neovim/nvim-lspconfig@a27179f56c6f98a4cdcc79ee2971b514815a4940/-/blob/lua/lspconfig/server_configurations/eslint.lua?L94-101
- coc-eslitn:
https://sourcegraph.com/github.com/neoclide/coc-eslint@70eb10d294e068757743f9b580c724e92c5b977d/-/blob/src/index.ts?L698:17-698:35
Release Notes:
- Changed the default ESLint configuration to include the following in
order to silence warnings/errors: `{"codeAction": {
"disableRuleComment": { "enable": true, "location": "separateLine", },
"showDocumentation": { "enable": true } }}`
This fixes the issue mentioned here:
https://github.com/zed-industries/zed/issues/5830#issuecomment-2111947083
In order for other languages to work, we need to pass the following
settings along to the Tailwind language server.
With the following Zed settings, it then also works for Scala:
```json
{
"languages": {
"Scala": {
"language_servers": ["tailwindcss-language-server"]
},
},
"lsp": {
"tailwindcss-language-server": {
"settings": {
"includeLanguages": {
"scala": "html"
},
"experimental": {
"classRegex": ["[cls|className]\\s\\:\\=\\s\"([^\"]*)"]
}
}
}
}
}
```
Release Notes:
- Added ability to configure settings for `tailwindcss-language-server`,
namely the `includeLanguages` and `experimental` objects.
**NOTE**: I have only tested that the language server boots up for Scala
files and that the settings are forwarded correctly. I don't have a
Scala+Tailwind project with which to test that the actual completions
also work.
cc @nguyenyou
Previously, we had an instance buffer pool that could only allocate
buffers with a fixed size (hardcoded to 2mb). This caused certain scenes
to render partially, e.g. when showing tens of thousands of glyphs on a
big screen.
With this commit, when `MetalRenderer` detects that a scene would be too
large to render using the current instance buffer size, it will:
- Clear the existing instance buffers
- Allocate new instance buffers that are twice as large
- Retry rendering the scene that failed with the newly-allocated buffers
during the same frame.
This fixes#11615.
Release Notes:
- Fixed rendering issues that could arise when having large amounts of
text displayed on a large display. Fixed by dynamically increasing the
size of the buffers used on the GPU.
([#11615](https://github.com/zed-industries/zed/issues/11615)).
Before:
https://github.com/zed-industries/zed/assets/1185253/464463be-b61c-4149-a417-01701699decb
After:
https://github.com/zed-industries/zed/assets/1185253/4feacf5a-d862-4a6b-90b8-317ac74e9851
Co-authored-by: Antonio <me@as-cii.com>
This fixes#8794 and other related problems.
The problem, in short, is this: `$SHELL` might be outdated. This code
ensures that we update `$SHELL` to what we can deem the newest version,
if we're started as a desktop application.
The background is that you can get the user's preferred shell in two
ways:
1. Read the `SHELL` env variable
2. Read the `/etc/passwd` file and check which shell is set
Most applications should and do prefer (1) over (2).
Why is it preferred? Reading `SHELL` means that processes can inherit
the variable from each other. And you can do something like
`SHELL=/bin/cool-shell ./my-cool-app`
But what happens if the application was launched from the desktop? Which
SHELL env does it inherit then?
It inherits the env from the process that launched it, which is
Finder.app or launchd or GNOME or something else — these are all
long-running processes that get their environment when the user logs in.
They do *not* get a new environment unless restarted (either process
restarted or computer restarted)
That means the `SHELL` env variable they have might be outdated.
That's a problem if you, for example, change your shell with `chsh` and
then launch the app from the desktop.
That change of the default shell is not reflected in the app if the app
only reads from SHELL. Because that hasn’t been updated. Instead it
should read from passwd file to get the newest value.
Release Notes:
- Fixed SHELL being outdated if Zed was launched via Finder or Raycast
or other desktop launchers.
([#8794](https://github.com/zed-industries/zed/issues/8794))
This fixes#10871.
The introduction of #11412 broke Vue.js language support, since it made
Zed rely more heavily on correct language name -> language ID mappings,
which the Vue.js extension didn't have.
Release Notes:
- N/A
This fixes#10185 by not keeping snippet state around when already at
the end of the snippet and the tabstop is empty (i.e. it's not a
selection) and we're already on it.
The reason for the fix is outlined in the comments of #10185 but to
repeat:
1. `gopls` sends completions with type "snippet" even when suggesting
single word completions that don't contain tabstops
2. We use a default behavior and add an "end tabstop" by default so that
the cursor jumps to the end of the snippet when accepting it.
3. We'd then push the state of the snippet on the stack which is where
it would stay, with the cursor already at the end and the user unable to
get rid of the tabstop state.
This fixes the issue by not pushing snippet state when the tabstop we
accepted is the "end tabstop".
Release Notes:
- Fixed completions inside snippets breaking the jump-to-next-tabstop
behaviour when using Go/`gopls`
([#10185](https://github.com/zed-industries/zed/issues/10185)).
Demo:
https://github.com/zed-industries/zed/assets/1185253/35384e5e-45c6-46ab-870d-b48e56d8786b
This adds a new option to the right click menu for git blame entries in
the gutter: "Open permalink". If there is a URL for the code host, then
this will open it.
Release Notes:
- Added "Open permalink" option to right-click menu of git blame entries
in gutter.
Demo:
I ran into this when trying to get #11550 working: the VCS menu would
open repositories on its owned, based on paths, instead of going through
the worktree on which we already store the git repositories.
Release Notes:
- N/A
This fixes restart after updates not working on Linux.
On Linux we can't reliably get the binary path after an update, because
the original binary was deleted and the path will contain ` (deleted)`.
See: https://github.com/rust-lang/rust/issues/69343
We *could* strip ` (deleted)` off, but that feels nasty. So instead we
save the original binary path, before we do the installation, then
restart.
Later on, we can also change this to be a _new_ binary path returned by
the installers, which we then have to start.
Release Notes:
- N/A
This implements `app_version` on Linux by using an optional,
compile-time `RELEASE_VERSION` env var that can be set.
We settled on the `RELEASE_VERSION` as the name, since it's similar to
`RELEASE_CHANNEL` which we use in Zed.
cc @ConradIrwin @mikayla-maki
Release Notes:
- N/A
Co-authored-by: Bennet <bennetbo@gmx.de>
This fixes a race-condition that showed up when trying to restart
Nightly/Preview/...
When running with these release channels, Zed tries to ensure that
there's only one instance of Zed running.
It does that by listening on a TCP socket to which other instances can
connect on start. If the other instance receives a message, it knows
that another Zed instance is running and exits.
On Linux, though, we ran into a race condition:
1. `kill -0`, which checks whether a process is still running, returns
an error, signalling that the old Zed process has exited
2. BUT: the process was still listening on the TCP port.
It seems like that on Linux, process resources aren't guaranteed to be
cleaned up as soon as signal handling stops working for a process.
The fix is to wait until the process is no longer listening on any TCP
sockets.
There's a slight downside to this: GPUI processes that never listen on
any TCP sockets now have to pay the cost of an additional `lsof` call
when restarting. We do think that it's a reasonable tradeoff for now
though, since the other options (extending the platform interface to
provide callbacks, sharing the listening port in the framework, ...)
seem wider-reaching only to fix a very local bug.
Release Notes:
- N/A
Co-authored-by: Bennet <bennetbo@gmx.de>
This downloads Nightly/Preview releases on Linux and copies the contents
the `zed-<channel>.app` to `~/.local`.
What's missing:
- Check if we're not installed in ~/.local and abort
- Update `.desktop` file
Release Notes:
- N/A
This undoes the changes from #11333 and uses the path of the `.desktop`
file instead.
According ot the spec
(https://specifications.freedesktop.org/desktop-entry-spec/desktop-entry-spec-latest.html)
the filename and path of the `desktop` file are used to determine the
desktop file ID.
This is enough to match the windows (which have the same WMClass/app-id)
with the desktop entry.
Release Notes:
- N/A
Since we do want to have different versions of Zed running on the same
Linux install, we need to give them different application IDs so they're
not grouped together as the same application.
This changes the app_id depending on the releaes channel and, crucially,
it also matches them up with the bundle identifiers that we use on
macOS.
Release Notes:
- N/A
This allows function call (i.e. snippet) completion with
`typescript-language-server`. So far that didn't work, because
`typescript-language-server` doesn't respond with `insertText` when
getting the completions, but only when then sending
`completionItem/resolve` requests. See:
https://github.com/hrsh7th/nvim-cmp/issues/646#issuecomment-992765479
What this PR does is to support text edits in the response to
`completionItem/resolve`, which means updating the completion item.
It then enables this feature by default for
`typescript-language-server`.
TODOs:
- [x] Make this work over collab
- [x] Test that this doesn't break existing language server support
- [x] Refactor duplicated code
Release Notes:
- Added support for function call completion when using
`typescript-language-server`. This will result in parameters being
added, which can then be changed and navigated with `<tab>`. For this to
work with `typescript-language-server`, the documentation for a given
completion item needs to be resolved, meaning that if one types very
quickly and accepts completion before `typescript-language-server` could
respond with the documentation, no full function completion is used.
Demo:
https://github.com/zed-industries/zed/assets/1185253/c23ebe12-5902-4b50-888c-d9b8cd32965d
I think the previous code was missing a `return` in there because it
always overwrote the `completion.documentation` field, even if the
`text.is_empty()` is true.
Release Notes:
- N/A
This is a follow-up to #10685 which started to hide these errors instead
of displaying them to the user.
But the errors are still noisy and not actionable, so we hide them
instead.
Release Notes:
- Removed error message being logged when `git blame` is run in a
repository without commits.
Co-authored-by: Bennet <bennetbo@gmx.de>
This fixes#9934 and does two things:
1. It registers ESLint as a secondary language server for Vue.js files
(`.vue`)
2. It registers ESLint as a _secondary_ (instead of primary) language
server for TypeScript, JavaScript and TSX.
The second point because I don't see any reason why ESLint should be
registered as a primary language server for these languages. I read
through the code in `project.rs` that uses the primary language server
and I don't think there will be any differences to how it previously
worked.
I also manually tested ESLint support in a Vue.js project, a Next.js
project and a plain old JS project — still works in all three.
Release Notes:
- Added ESLint support for Vue.js files by starting it as a language
server on `.vue` files.
([#9934](https://github.com/zed-industries/zed/issues/9934)).
This fixes#9811 by checking for the `solargraph` binary in the `$PATH`
as it's setup in the project shell.
It also adds support for configuring the path to `solargraph` manually:
```json
{
"lsp": {
"solargraph": {
"binary": {
"path": "/Users/thorstenball/bin/solargraph",
"arguments": ["stdio"]
}
}
}
}
```
## Example
Given the following setup:
- `ruby@3.3.0` used globally, no `solargraph` installed globally
- `ruby@3.2.2` used in a project, `solargraph` installed as binstub in
`$project/bin/solargraph`, `.envrc` to configure `direnv` to add
`$project/bin` to `$PATH
Which looks like this in practice:
```shell
# GLOBAL
~ $ ruby --version
ruby 3.3.0 (2023-12-25 revision 5124f9ac75) [arm64-darwin23]
~ $ which solargraph
solargraph not found
# IN PROJECT
~ $ cd work/projs/rails-proj
direnv: loading ~/work/projs/rails-proj/.envrc
direnv: export ~PATH
~/work/projs/rails-proj $ ruby --version
ruby 3.2.2 (2023-03-30 revision e51014f9c0) [arm64-darwin23]
~/work/projs/rails-proj $ which solargraph
/Users/thorstenball/work/projs/rails-proj/bin/solargraph
```
The expectation is that Zed, when opening `~/work/projs/rails-proj`,
picks up the local `solargraph`.
But with **Zed Stable** that doesn't work, as we can see in the logs:
```
2024-04-22T10:21:37+02:00 [INFO] starting language server. binary path: "solargraph", working directory: "/Users/thorstenball/work/projs/rails-proj", args: ["stdio"]
2024-04-22T10:21:37+02:00 [ERROR] failed to start language server "solargraph": No such file or directory (os error 2)
```
With the change in this PR, it uses `rails/proj/bin/solargraph`:
```
[2024-04-22T10:33:06+02:00 INFO language] found user-installed language server for Ruby. path: "/Users/thorstenball/work/projs/rails-proj/bin/solargraph", arguments: ["stdio"]
[2024-04-22T10:33:06+02:00 INFO lsp] starting language server. binary path: "/Users/thorstenball/work/projs/rails-proj/bin/solargraph", working directory: "/Users/thorstenball/work/projs/rails-proj", args: ["stdio"]
```
**NOTE**: depending on whether `mise` (or `rbenv`, `asdf`, `chruby`,
...) or `direnv` come first in the shell-rc file, it picks one or the
other, depending on what puts itself first in `$PATH`.
## Release Notes
Release Notes:
- Added support for finding the Ruby language server `solargraph` in the
user's `$PATH` as it is when `cd`ing into a project's directory.
([#9811](https://github.com/zed-industries/zed/issues/9811))
- Added support for configuring the `path` and `arguments` for
`solargraph` language server manually. Example from settings: `{"lsp":
{"solargraph": {"binary":
{"path":"/Users/thorstenball/bin/solargraph","arguments": ["stdio"]}}}}`
([#9811](https://github.com/zed-industries/zed/issues/9811))
Thorsten Ball
