See #27808. `font_id_for_cosmic_id` was another path updated
`loaded_fonts_store` but did not push to `features_store`. Solution is
just to have one `Vec` with fields rather than relying on the indices
matching up
Release Notes:
- N/A
Closes#30526.
This PR makes the CacheKey used by raster_bounds and rasterize_glyph the
same, as they had not used the same sub pixel shift previously. Fixing
this resolves both the alignment and text-rendering issues introduced in
`ddf8d07`.
Release Notes:
- Fixed text rendering issues on Linux.
Fixes#15752.
- Updated `cosmic_text` to 0.14.0
- Made a basic implementation for setting font features.
#12176 is not fixed by this PR.
Release Notes:
- Added initial support for `font_features` on Linux
- Adds a new smoke test for the use of the read_file tool by the agent
in an SSH project
- Fixes the SSH shutdown sequence to use a timer from the app's executor
instead of always using a real timer
- Changes the main executor loop for tests to advance the clock
automatically instead of panicking with `parked with nothing left to
run` when there is a delayed task
Release Notes:
- N/A
This pull request introduces a new tool for streaming edits. The
short-term goal is for this tool to replace the existing `EditFileTool`,
but we want to get this out the door as soon as possible so that we can
start testing it.
`StreamingEditFileTool` is mutually exclusive with `EditFileTool`. It
will be enabled by default for anyone who has the `agent-stream-edits`
feature flag, as well as people that set `assistant.stream_edits` to
`true` in their settings.
### Implementation
Streaming is achieved by requesting a completion while the `edit_file`
tool gets called. We invoke the model by taking the existing
conversation with the agent and appending a prompt specifically tailored
for editing. In that prompt, we ask the model to produce a stream of
`<old_text>`/`<new_text>` tags. As the model streams text in, we
incrementally parse it and start editing as soon as we can.
### Evals
Note that, as part of this pull request, I also defined some new evals
that I used to drive the behavior of the recursive LLM call. To run
them, use this command:
```bash
cargo test --package=assistant_tools --features eval -- eval_extract_handle_command_output
```
Or comment out the `#[cfg_attr(not(feature = "eval"), ignore)]` macro.
I recommend running them one at a time, because right now we don't
really have a way of orchestrating of all these evals. I think we should
invest into that effort once the new agent panel goes live.
Release Notes:
- N/A
---------
Co-authored-by: Nathan Sobo <nathan@zed.dev>
Co-authored-by: Bennet Bo Fenner <bennetbo@gmx.de>
Co-authored-by: Oleksiy Syvokon <oleksiy.syvokon@gmail.com>
Closes:
https://github.com/zed-industries/zed/pull/29177#issuecomment-2823359242
Removes dependency on
[quininer/x11-clipboard](https://github.com/quininer/x11-clipboard) as
it is in [maintenance
mode](https://github.com/quininer/x11-clipboard/issues/19).
X11 clipboard functionality is now built-in to GPUI which was
accomplished by stripping the non-x11-related code/abstractions from
[1Password/arboard](https://github.com/1Password/arboard) and extending
it to support all image formats already supported by GPUI on wayland and
macos.
A benefit of switching over to the `arboard` implementation, is that we
now make an attempt to have an X11 "clipboard manager" (if available -
something the user has to setup themselves) save the contents of
clipboard (if the last copy operation was within Zed) so that the copied
contents can still be pasted once Zed has completely stopped.
Release Notes:
- Linux(X11): Add support for pasting images from clipboard
Closes#28533
Release Notes:
- Linux: Improved parsing of `ZED_DEVICE_ID` environment variable in an
attempt to fix some cases where it erroneously failed to parse. The
`ZED_DEVICE_ID` is now expected to always be a 4 digit hexadecimal
number (as it is in the output of `lcpci`) with an optional `0x` or `0X`
prefix.
Closes#29117
This makes `window_min_size` work by using the `WM_GETMINMAXINFO` window
message.
Release Notes:
- N/A
---------
Co-authored-by: 张小白 <364772080@qq.com>
This PR adds a new `PlatformKeyboardLayout` trait with two methods:
`id(&self) -> &str` and `name(&self) -> &str`. The `id()` method returns
a unique identifier for the keyboard layout, while `name()` provides a
human-readable name. This distinction is especially important on
Windows, where the `id` and `name` can be quite different. For example,
the French layout has an `id` of `0000040C`, which is not
human-readable, whereas the `name` would simply be `French`. Currently,
the existing `keyboard_layout()` method returns what's essentially the
same as `id()` in this new design.
This PR implements the `name()` method for both Windows and macOS. On
Linux, for now, `name()` still returns the same value as `id()`.
Release Notes:
- N/A
This fixes the panic that happened in debug builds in Wayland when
focusing/defocusing window in the edit mode:
```
"Thread "main" panicked with "CursorStyle::None should be handled separately in the client" at crates/gpui/src/platform/linux/wayland.rs:40:17"
```
Full log:
[stacktrace.txt](https://github.com/user-attachments/files/19814411/stacktrace.txt)
@smitbarmase, you seem to have worked on this code. Tagging you for
visibility :)
Release Notes:
- N/A
While working on implementing `add_recent_documents` for Windows, I
found that the process is significantly more complex compared to macOS.
On macOS, simply registering the `add_recent_documents` function is
enough, as the system handles everything automatically.
On Windows, however, there are two cases to consider:
- **Files opened by the app**: These appear in the "Recent" section (as
shown in the screenshot, "test.txt") and are managed automatically by
Windows (by setting windows registry), similar to macOS.
- **Folders opened by the app**: This is more complicated because
Windows does not handle it automatically, requiring the application to
track opened folders manually.
To address this, this PR introduces a `History Manager` along with
`HistoryManagerEvent::Update` and `HistoryManagerEvent::Delete` events
to simplify the process of managing recently opened folders.
https://github.com/user-attachments/assets/a2581c15-7653-4faf-96b0-7c48ab1dcc8d
Release Notes:
- N/A
---------
Co-authored-by: Mikayla Maki <mikayla@zed.dev>
Move the SetLayeredWindowAttributes call to immediately after window
construction, and initialize it with per-pixel transparency settings, no
color key and no global blending. The render pipeline will perform alpha
blending during compositing.
Cleaned up the DWM acrylic API calls some, to explicitly set to the
three appropriate modes depending on opaque, transparent or blurred
settings. The API internally hides versioning concerns from the caller.
Set the window class background color to black, this prevents a
flashbang on slow startup, e.g. debug builds on a heavily loaded system.
The outcome is that the window no longer receives paint demands for
underlying window updates, while also having per-pixel transparency -
opaque theme elements are now correctly opaque. The transparency
settings are now portable across windows and macOS having mostly similar
outcomes (modulo palette differences). Small fonts may still appear to
be alpha blended - this seems to be in the glyph atlas, their pixels are
not actually opaque. Larger fonts (or higher DPIs) don't suffer this and
are as opaque as expected. Layering the window atop one that is
rendering at 120fps, the editor window can drop to its 8fps idle state,
while still being composited with 120fps alpha blend in the background,
in both blur and transparent modes.
Updates #20400
Release Notes:
- Improved transparency on Windows to be more efficient, support fully
opaque elements and more closely match other platforms.
Seems to be very similar to
https://github.com/zed-industries/zed/pull/28059
Edit: Updated the reproduction-steps as I missed something.
The method without a check currently causes my debug-builds to crash on
the regular if I:
- Run a debug build and open it fullscreen in a dedicated space on my
Mac.
- Work on any of the built-in languages (e.g. remove some content from
any `highlights.scm`)
- Reopen the workspace with the debug-build.
- Crash.
~~We might actually be able to revert the changes made in
https://github.com/zed-industries/zed/pull/21510 and just add the
null-check. Then again, I am not at all sure whether that would work.~~
See comment below.
Release Notes:
- N/A
Use the `NSOperatingSystemVersion` struct provided by the cocoa crate
instead of our own. Additionally we can directly use
`isOperatingSystemAtLeastVersion` instead of manually implementing
version comparison logic.
The `isOperatingSystemAtLeastVersion` instance method has been available
since MacOS 10.10, which released a decade ago.
Documentation for `isOperatingSystemAtLeastVersion `:
https://developer.apple.com/documentation/foundation/nsprocessinfo/1414876-isoperatingsystematleastversion
Release Notes:
- N/A
Reverts the error behavior introduced in #27558. Upper-case keys in
keybindings no longer generate errors, instead they are transformed into
`shift-{KEY}`
e.g. `ctrl-N` becomes `ctrl-shift-n`
The behavior introduced in #27558 where "special" keys such as function
keys, `control`, `shift`, etc. Are parsed case-insensitively is
preserved.
Release Notes:
- Improved how upper-case characters are handled in keybinds. "special"
keys such as the function keys, `control`, `shift`, etc. are now parsed
case-insensitively, so for example `F8`, `CTRL`, `SHIFT` are now
acceptable alternatives to `f8`, `ctrl`, and `shift` when declaring
keybindings. Additionally, upper-case (ascii) characters will now be
converted explicitly to `shift` + the lowercase version of the
character, to match the Vim behavior.
NOTE: Release notes above should replace the release notes from #27558
While building my own application using GPUI, I found that the `key_up`
event doesn't fire on Windows or macOS, with each platform failing for
different reasons. These events aren't used anywhere in Zed yet, so it
makes sense that the issue hasn't already been caught.
I don't have a Linux machine set up right now, so I don't know if these
events fire correctly on Linux or not.
---
Without this fix, a simple layout like the following:
```rust
div()
.on_key_down(cx.listener(|_, event, _, _| println!("Key down: {:?}", event)))
.on_key_up(cx.listener(|_, event, _, _| println!("Key up: {:?}", event)));
```
...would result in the following logs if the 'a' key was pressed:
```text
Key down: KeyDownEvent { keystroke: Keystroke { modifiers: Modifiers { control: false, alt: false, shift: false, platform: false, function: false }, key: "a", key_char: Some("a") }, is_held: false }
<eof>
```
With this fix, the `key_up` event fires correctly, resulting in the
following logs:
```text
Key down: KeyDownEvent { keystroke: Keystroke { modifiers: Modifiers { control: false, alt: false, shift: false, platform: false, function: false }, key: "a", key_char: Some("a") }, is_held: false }
Key up: KeyUpEvent { keystroke: Keystroke { modifiers: Modifiers { control: false, alt: false, shift: false, platform: false, function: false }, key: "a", key_char: None } }
<eof>
```
---
I've made the assumption that the `key_char` field shouldn't be set on
the `key_up` event since, unlike the `key_down` event, it's not an event
that may produce a character.
Happy to make any changes to this PR as required. If it would be
preferable to test this on Linux as well before it's merged, let me know
and I'll sort something out.
Hopefully this makes the experience of building new applications on GPUI
smoother, and potentially saves the Zed team some time if this event is
ever used in the future.
Release Notes:
- N/A
Swift bindings BEGONE
Release Notes:
- Switched from using the Swift LiveKit bindings, to the Rust bindings,
fixing https://github.com/zed-industries/zed/issues/9396, a crash when
leaving a collaboration session, and making Zed easier to build.
---------
Co-authored-by: Conrad Irwin <conrad.irwin@gmail.com>
Co-authored-by: Michael Sloan <michael@zed.dev>
Workaround for users affected by #25899
Thanks to the work done by @kvark in
https://github.com/kvark/blade/pull/210, we have the ability to tell
Vulkan (through blade) a specific GPU to use.
This will hopefully allow some of the users affected by #25899 to use
Zed by allowing them to use a specific GPU, if the primary/default GPU
will not work
Release Notes:
- Added the ability to specify which GPU Zed uses on Linux by setting
the `ZED_DEVICE_ID` environment variable. You can obtain the device ID
of your GPU by running `lspci -nn | grep VGA` which will output each GPU
on one line like:
```
08:00.0 VGA compatible controller [0300]: NVIDIA Corporation GA104
[GeForce RTX 3070] [10de:2484] (rev a1)
````
where the device ID here is `2484`. This value is in hexadecimal, so to
force Zed to use this specific GPU you would set the environment
variable like so:
```
ZED_DEVICE_ID=0x2484
```
Make sure to export the variable if you choose to define it globally in
a `.bashrc` or similar
Closes#4461
Take 2 on https://github.com/zed-industries/zed/pull/25040.
Fixes panic caused due to using `setHiddenUntilMouseMoves` return type
to `set` cursor on macOS.
Release Notes:
- Now cursor hides when the user is typing in editor. It will stay
hidden until it is moved again. This behavior is `true` by default, and
can be configured with `hide_mouse_while_typing` in settings.
---------
Co-authored-by: Peter Tripp <peter@zed.dev>
Co-authored-by: Thomas Mickley-Doyle <thomas@zed.dev>
Co-authored-by: Agus <agus@zed.dev>
Co-authored-by: Kirill Bulatov <kirill@zed.dev>
Co-authored-by: Agus Zubiaga <hi@aguz.me>
Co-authored-by: Angelk90 <angelo.k90@hotmail.it>
Related #14222
Release Notes:
- This PR updates `cosmic-text` dependency on `gpui` crate from `0.11.2`
to 0.13.2`. This decreases RAM usage zed depending on the amount of
monospace fonts installed on the system. On Arch Linux with `nerd-fonts`
package installed (which provides around 2000 monospaced fonts), it
decreases ram usage from ~800mb to around ~300mb.
- Updated `cosmic-text` to `0.13.2` on `gpui` crate
There apparently is some amdgpu/radv bug that rendering with
multisample anti-aliasing (MSAA) results in a crash when the bounds
of a triangle list exceed 1024px, which in Zed happens with the default
buffer font size when you select a line with more than 144 characters.
This crash has been reported as #26143.
This commit introduces a workaround: you can set the
ZED_PATH_SAMPLE_COUNT=0
environment variable to disable MSAA and the error message we print
when a GPU crash is encountered with radv now suggests trying this
environment
variable as a workaround and links the respective issue.
Sidenote: MSAA was introduced in
f08b1d78ec
so you didn't run into this driver bug with versions < 0.173.8.
Release Notes:
- Added a workaround for an AMD Linux driver bug that causes Zed to
crash when selecting long lines.
Closes#4640
About the support limit of Fn:
Mac F1-F35
Win F1-F24
Linux F1-F35
Terminal F1-F20
Release Notes:
- Improved support for extended keyboards on Mac (F20-F35)
张小白
Michael Sloan
william341
peppidesu
tidely
Cole Miller
Ben Kunkle
angelrecoveryand张小白
Oleksiy Syvokon
João Marcos
Conrad Irwin
James Tucker
Finn Evers
Piotr Osiewicz
Floyd Wang
Felix Packard
Alvaro Parker
Martin Fischer
0x2CA