Instead of a menagerie of macros for implementing `Action`, now there
are just two:
* `actions!(editor, [MoveLeft, MoveRight])`
* `#[derive(..., Action)]` with `#[action(namespace = editor)]`
In both contexts, `///` doc comments can be provided and will be used in
`JsonSchema`.
In both contexts, parameters can provided in `#[action(...)]`:
- `namespace = some_namespace` sets the namespace. In Zed this is
required.
- `name = "ActionName"` overrides the action's name. This must not
contain "::".
- `no_json` causes the `build` method to always error and
`action_json_schema` to return `None`
and allows actions not implement `serde::Serialize` and
`schemars::JsonSchema`.
- `no_register` skips registering the action. This is useful for
implementing the `Action` trait
while not supporting invocation by name or JSON deserialization.
- `deprecated_aliases = ["editor::SomeAction"]` specifies deprecated old
names for the action.
These action names should *not* correspond to any actions that are
registered. These old names
can then still be used to refer to invoke this action. In Zed, the
keymap JSON schema will
accept these old names and provide warnings.
- `deprecated = "Message about why this action is deprecation"`
specifies a deprecation message.
In Zed, the keymap JSON schema will cause this to be displayed as a
warning. This is a new feature.
Also makes the following changes since this seems like a good time to
make breaking changes:
* In `zed.rs` tests adds a test with an explicit list of namespaces. The
rationale for this is that there is otherwise no checking of `namespace
= ...` attributes.
* `Action::debug_name` renamed to `name_for_type`, since its only
difference with `name` was that it
* `Action::name` now returns `&'static str` instead of `&str` to match
the return of `name_for_type`. This makes the action trait more limited,
but the code was already assuming that `name_for_type` is the same as
`name`, and it requires `&'static`. So really this just makes the trait
harder to misuse.
* Various action reflection methods now use `&'static str` instead of
`SharedString`.
Release Notes:
- N/A
<img width="1728" alt="image"
src="https://github.com/user-attachments/assets/a63925a7-8e13-4d48-bd31-33f434209ea6"
/>
Diagnostics UI elements (underlines, popovers, hovers) are quite noisy
by themselves and get even more so with the git background colors.
Release Notes:
- Stopped showing diagnostics in the diff-related editors
This PR moves the UI-dependent logic in the `agent` crate into its own
crate, `agent_ui`. The remaining `agent` crate no longer depends on
`editor`, `picker`, `ui`, `workspace`, etc.
This has compile time benefits, but the main motivation is to isolate
our core agentic logic, so that we can make agents more
pluggable/configurable.
Release Notes:
- N/A
This PR adds an optional `clone: bool` argument to
`workspace::MoveItemToPane` and `workspace::MoveItemToPaneInDirection`
which causes the item to be cloned into the destination pane rather than
moved. It provides similar functionality to
`workbench.action.splitEditorToRightGroup` in vscode.
This PR supercedes #25030.
Closes#24889
Release Notes:
- Add optional `clone: bool` (default: `false`) to
`workspace::MoveItemToPane` and `workspace::MoveItemToPaneInDirection`
which causes the item to be cloned into the destination pane rather than
moved.
Removed an additional label from the modal empty state as I figured we
don't need it, given we already have a version of the description in
place for when the extension itself doesn't return any. So, ultimately,
having both the label and the description was redundant.
Release Notes:
- N/A
There may be cases where we're needing to pass a button just so it is
dismissible, so I figured this out help! It also helps when you want to
have two buttons, one to perform an action and another to dismiss and
cancel.
Release Notes:
- N/A
This PR addresses to fix (#31308) a race condition where auto-indent (in
buffer.cs) and on-type-formatting (in lsp_store.rs) concurrently
calculate indentation using the same buffer snapshot.
Previous Solution (Abandoned):
https://github.com/zed-industries/zed/pull/31340
Final Solution:
Delay applying on-type-formatting until auto-indent is complete.
Issue:
If AutoindentMode finishes first, formatting works correctly. If
"Formatting on typing" starts before AutoindentMode completes, it
results in double indentation.
Closes#31308
Release Notes:
- Fixed a race condition resulting in incorrect buffer contents when combining auto-indent and on-type-formatting
Before this PR force quitting Zed would leave hanging debug adapter
processes and not allow debug adapters to clean up their sessions
properly.
This PR fixes this problem by sending a disconnect/terminate to all
debug adapters and force shutting down their processes after they
respond.
Co-authored-by: Cole Miller \<cole@zed.dev\>
Release Notes:
- debugger: Shutdown and clean up debug processes when force quitting
Zed
---------
Co-authored-by: Conrad Irwin <conrad.irwin@gmail.com>
Co-authored-by: Remco Smits <djsmits12@gmail.com>
This PR is in preparation for doing automatic retries for certain
errors, e.g. Overloaded. It doesn't change behavior yet (aside from some
granularity of error messages shown to the user), but rather mostly
changes some error handling to be exhaustive enum matches instead of
`anyhow` downcasts, and leaves some comments for where the behavior
change will be in a future PR.
Release Notes:
- N/A
Custom arguments replace any arguments that we normally pass to the DAP.
For interpreted languages, they are passed to the interpreter after the
DAP path or module. They can be combined with a custom binary, or you
can omit `dap.binary` and just customize the arguments to the DAPs we
download.
This doesn't take care of updating the extension API to support custom
arguments.
Release Notes:
- debugger: Implemented support for passing custom arguments to a debug
adapter binary using the `dap.args` setting.
- debugger: Fixed not being able to use the `dap` setting in
`.zed/settings.json`.
Implements a workaround which removes the `WS_EX_LAYERED` style from the
window right before closing it which seems to fix the window close
animation not playing.
Release Notes:
- N/A
CC: @danilo-leal Do you have thoughts on this? I found myself typing
chat messages after a long thread and then deciding I would be better
served by restarting from a summary -- and then "poof" the contents of
my chat box was lost.
Release Notes:
- agent: "New From Summary" now preserves any unsent content in the chat
box.
Previously, if editing a long previous user message in the thread, you'd
have a double scroll situation because the editor used in that case had
its max number of lines capped. To solve that, I made the `max_lines` in
the editor `AutoHeight` mode optional, allowing me to not pass any
arbitrary number to the previous user message editor, and ultimately,
solving the double scroll problem by not having any scroll at all.
Release Notes:
- agent: Fixed double scroll that happened when editing a long previous
user message.
@ConradIrwin adding you as a reviewer as I'm touching editor code
here... want to be careful. :)
Closes#31903
Release Notes:
- agent: Fix an issue where an error would occur when MCP servers
specified tools with the same name
---------
Co-authored-by: Ben Brandt <benjamin.j.brandt@gmail.com>
I don't have a concrete misbehavior from it, but this update of
`insert_range` doesn't make sense for two reasons:
* If the resolved completion doesn't have a new `text_edit` it would
clear out the `insert_range`.
* It doesn't update the completion if it has already been resolved,
except this update of `insert_range` happened before that.
Guessing it was written this way because this field needed to only be
mutated within the `CompletionSource::Lsp` case and this was a
convenient match below.
Release Notes:
- N/A
Co-authored-by: Smit <smit@zed.dev>
Closes#29956
Unlike GNOME Keyring, KeePassXC locks individual secrets in addition to
the entire database when configured to ask for confirmation for access
requests by DBus clients. As such, before the secret is read it should
be unlocked by the client.
Tested against both KeePassXC and GNOME Keyring, and with this patch Zed
successfully logs in and fetches the API keys from the Secret Service.
Release Notes:
- Fixed KeePassXC integration via org.freedesktop.Secrets
This PR contains a set of changes for improving FreeBSD support (#15309,
#29550) and is a kind of follow up to the PR #20480 which added an
initial support for FreeBSD.
A summary of changes is as follows:
- Add some more freebsd conditionals which seem missing in the previous
PR.
- Implement `anonymous_fd()` and `current_path()` functions for FreeBSD.
- Improve detection of FreeBSD in telemetry and GPU detection.
- Temporarily disable LiveKit/WebRTC support to make build succeed.
- Remove support for flatpak since it is Linux-only packaging format.
Adding `RUSTFLAGS="-C link-dead-code"` does not seem necessary anymore.
It builds fine without the flag.
Known issues:
- Integrated terminal is painfully laggy and virtually unusable in my
environment. This might be specific to my setup.
- I cannot input Japanese using IME. When I type characters, they appear
on the screen. But when I hit return key, they disappears. Seems the
same issue as #15409.
My environment is MATE desktop on X11 on FreeBSD 14.2 on Intel Core
i5-7260U integrated graphics.
P.S. For those who might be interested, a work-in-progress FreeBSD port
and binary packages are available at
https://github.com/tagattie/FreeBSD-Zed
Release Notes:
- N/A
---------
Co-authored-by: Peter Tripp <peter@zed.dev>
Closes#33106
Instead of directly using `filter_text` in `StringMatchCandidate`, which
yields better results for fuzzy matching but messes up with highlighting
letters in bold, as `positions` list generated by fuzzy crate are now of
`filter_text` and not `label` shown to the user.
This PR fixes it by keeping use of `filter_range` in
`StringMatchCandidate`, which is range w.r.t to `label` shown to user.
And actually generating this `filter_range` at source by using
`filter_range` if exists.
- [x] Tests
Release Notes:
- Fixed issue where incorrect letters are marked as bold in completions.
- [x] Show disabled MCP servers in the list so you can enable them again
- [x] If MCP is not present in the settings, add it
Closes #ISSUE
Release Notes:
- agent: Allow to enable/disable context servers permanently in the
agent configuration view
---------
Co-authored-by: Danilo Leal <daniloleal09@gmail.com>
Did some bit cleanup of code for loading models for settings as that is
not required as we are fetching all the models from openrouter so it's
better to maintain one source of truth
Release Notes:
- Add thinking support to OpenRouter provider
Having `Thread::last_usage` as an override of the initially fetched
usage could cause the initial usage to be displayed when the current
thread is empty or in text threads. Fix is to just store last usage info
in `UserStore` and not have these overrides
Release Notes:
- Agent: Fixed request usage display to always include the most recently
known usage - there were some cases where it would show the initially
requested usage.
### This PR introduces support for adding watchers to specific
expressions (such as variable names or evaluated expressions).
This feature is useful in scenarios where many variables are in scope,
but only a few are of interest—especially when tracking variables that
change frequently. By allowing users to add watchers, it becomes easier
to monitor the values of selected expressions across stack frames
without having to sift through a large list of variables.
https://github.com/user-attachments/assets/c49b470a-d912-4182-8419-7406ba4c8f1e
------
**TODO**:
- [x] make render variable code reusable for render watch method
- [x] use SharedString for watches because of a lot of cloning
- [x] add tests
- [x] basic test
- [x] test step debugging
Release Notes:
- Debugger Beta: Add support for variable watchers
---------
Co-authored-by: Anthony Eid <hello@anthonyeid.me>
Co-authored-by: Anthony <anthony@zed.dev>
Closes#25195
In Wayland, To create buffer size (`renderer.update_drawable_size`), we
convert logical pixels to device pixels by taking the scale factor into
account. Later, we also let the compositor know the logical pixels we
want to use for our app (`viewport.set_destination`). Then, the
compositor takes our buffer and tries to scale it to fit the viewport
size we provided. If this is accurate, we see perfect rendering. If our
buffer size is not accurate (off by 1px in this case), the compositor
scales our buffer to fit the viewport size. This causes blur.
To make sure we set correct buffer size for renderer as same as what
compositor is going to use, we needs to use rounding instead of truncate
when converting logical pixels to device pixels. It's not super clear
from docs, what exact algorithm it uses but it says it uses rounding and
seems to fix issue for me if we follow that for our buffer.
From https://wayland.app/protocols/fractional-scale-v1:
> If a surface has a surface-local size of 100 px by 50 px and wishes to
submit buffers with a scale of 1.5, then a buffer of 150px by 75 px
should be used and the wp_viewport destination rectangle should be 100
px by 50 px.
>
> For toplevel surfaces, the size is **rounded halfway away from zero**.
The rounding algorithm for subsurface position and size is not defined.
Tested on:
- [x] Gnome
- [x] KDE
- [ ] ~Sway~ (Need to investigate this more for Sway)
Release Notes:
- Fixed blurry rendering on Wayland when using fractional scaling for
Gnome and KDE.
Co-authored-by: Julia Ryan p1n3appl3@users.noreply.github.com
Co-authored-by: Antonio Scandurra me@as-cii.com
Removing it for two reasons:
1. We need a better implementation that doesn't hurt caching and doesn't
distracts the agent too much (see
https://github.com/zed-industries/zed/pull/32876 for more context)
2. Current insertion point of notifications doesn't play well with
Claude Thinking models (see
https://github.com/zed-industries/zed/issues/33000#issuecomment-2991709484)
I think we should get this code back in a form of a tool. But for now,
I'm dropping it to resolve recent issues.
Closes#33000
Release Notes:
- N/A
* Restore the ability to combine --diff with other path arguments
* Restore combining --diff with --wait
There is still one defect in the current handling of `--diff`: when Zed
is already open, we'll open the diff view in your current active zed
window. It would be better to search all of the open zed windows for any
window containing the diffed paths, but implementing that is a bit
complex. Currently, the logic for *picking* an existing zed window is
coupled to the logic for opening buffers in that window. I'd like to
decouple it, but I wanted to keep this change small, so that we hotfix
it to stable without too much risk.
Release Notes:
- Fixed a bug where the `--diff` CLI flag did not work with `--wait`
Co-authored-by: Ben Brandt <benjamin.j.brandt@gmail.com>
Variable list wasn't notified when a set variable value request was
successfully. This caused the variable list and inline values to show
stale data in some cases, which this PR fixes.
Release Notes:
- debugger: Fix bug where setting a variable's value wouldn't update the
variable list or inline values
Closes#33039
This PR fixes a bug which causes the newest versions of the Biome and
Tombi extensions to not work with older Zed versions.
The bug occurs because in #32822, the type of the debug adapter and
debug locators was changed from a Vec to a BTreeMap. However, these
fields were already introduced much earlier in Zed, which now causes the
de-serialization of the `extension.toml` to fail for older Zed versions.
Any extension compiled with the newest extension CLI bumped in
https://github.com/zed-industries/extensions/pull/2866 will not work
with older Zed versions prior to v0.191.
By adding this change and bumping the extension CLI again, this could be
prevented. On de-serialization, we would just fallback to either a Vec
for versions prior to v0.190 or a BTreeMap after. Feel free to let me
know what you think here.
Release Notes:
- N/A
Follow-up of https://github.com/zed-industries/zed/pull/33028
Due to request races, we cannot predict the amount of result_id s that
we'll get. Ensure their amount grows monotonically each time.
This time, iterations=5000 did not fail for me.
Release Notes:
- N/A
