This PR adds word/character diff for expanded diff hunks that have both
a deleted and added section, as well as a setting `word_diff_enabled` to
enable/disable word diffs per language.
- `word_diff_enabled`: Defaults to true. Whether or not expanded diff
hunks will show word diff highlights when they're able to.
### Preview
<img width="1502" height="430" alt="image"
src="https://github.com/user-attachments/assets/1a8d5b71-449e-44cd-bc87-d6b65bfca545"
/>
### Architecture
I had three architecture goals I wanted to have when adding word diff
support:
- Caching: We should only calculate word diffs once and save the result.
This is because calculating word diffs can be expensive, and Zed should
always be responsive.
- Don't block the main thread: Word diffs should be computed in the
background to prevent hanging Zed.
- Lazy calculation: We should calculate word diffs for buffers that are
not visible to a user.
To accomplish the three goals, word diffs are computed as a part of
`BufferDiff` diff hunk processing because it happens on a background
thread, is cached until the file is edited, and is only refreshed for
open buffers.
My original implementation calculated word diffs every frame in the
Editor element. This had the benefit of lazy evaluation because it only
calculated visible frames, but it didn't have caching for the
calculations, and the code wasn't organized. Because the hunk
calculations would happen in two separate places instead of just
`BufferDiff`. Finally, it always happened on the main thread because it
was during the `EditorElement` layout phase.
I used Zed's
[`diff_internal`](https://github.com/zed-industries/zed/blob/02b2aa6c50c03d3005bec2effbc9f87161fbb1e8/crates/language/src/text_diff.rs#L230-L267)
as a starting place for word diff calculations because it uses
`Imara_diff` behind the scenes and already has language-specific
support.
#### Future Improvements
In the future, we could add `AST` based word diff highlights, e.g.
https://github.com/zed-industries/zed/pull/43691.
Release Notes:
- git: Show word diff highlight in expanded diff hunks with less than 5
lines.
- git: Add `word_diff_enabled` as a language setting that defaults to
true.
---------
Co-authored-by: David Kleingeld <davidsk@zed.dev>
Co-authored-by: Cole Miller <cole@zed.dev>
Co-authored-by: cameron <cameron.studdstreet@gmail.com>
Co-authored-by: Lukas Wirth <lukas@zed.dev>
At some point, rust-analyzer started including the snippet expression
for some completions in the description field, but that can look like a
bug:
<img width="1570" height="578" alt="CleanShot 2025-11-27 at 20 39 49@2x"
src="https://github.com/user-attachments/assets/5a87c9fe-c0a8-472f-8d83-3bc9e9e00bbc"></img>
In these cases, we will now inline the tab stops as an ellipsis
character:
<img width="1544" height="428" alt="CleanShot 2025-12-01 at 10 01 18@2x"
src="https://github.com/user-attachments/assets/4c550891-4545-47cd-a295-a5eb07e78e92"></img>
You may also notice that we now syntax highlight the pattern closer to
what it looks like after accepted.
Alternatively, when the tab stop isn't just one position, it gets
highlighted as a selection since that's what it would do when accepted:
<img width="1558" height="314" alt="CleanShot 2025-12-01 at 10 04 37@2x"
src="https://github.com/user-attachments/assets/ce630ab2-da22-4072-a996-7b71ba21637d"
/>
Release Notes:
- rust: Display completion tab stops inline rather than as a raw LSP
snippet expression
Gist is we only need to block the foreground thread for reparsing if
immediate language changes are useful to the user. That is usually only
the case when they edit the buffer
Release Notes:
- Improved performance of large project searches and project diffs
Co-authored by: David Kleingeld <david@zed.dev>
Closes#9648#9755
Release Notes:
- Added way to configure ESLint's working directories in settings. For
example:
`{"lsp":{"eslint":{"settings":{"workingDirectories":["./client","./server"]}}}}`
---------
Co-authored-by: Piotr Osiewicz <24362066+osiewicz@users.noreply.github.com>
Zeta evals now include a character n-gram metric adapted for multi-edit diffs (“delta chrF”). It works as follows:
1. Reconstruct the original, golden (expected), and actual texts from unified diffs.
- "original": the text before any edits
- "golden": the text after applying the expected edits
- "actual": the text after applying the actual edits
2. Compute n-gram count deltas between original→golden and original→actual.
- n-grams are computed as in chrF (max n=6, whitespace ignored).
3. Compare these deltas to assess how well the actual edits match the expected edits.
- As in standard chrF, classify n-grams as true positives, false positives, and false negatives, and report the F-beta score with beta=2.
Release Notes:
- N/A
It is easy for us to get the two fields out of sync causing weird
problems, there is no reason to have both here so.
Release Notes:
- N/A *or* Added/Fixed/Improved ...
Co-authored by: Antonio Scandurra <antonio@zed.dev>
We've realized that a lot of the logic within an
`EditPredictionProvider` is not specific to a particular edit prediction
model / service. Rather, it is just the generic state management
required to perform edit predictions at all in Zed. We want to move to a
setup where there's one "built-in" edit prediction provider in Zed,
which can be pointed at different edit prediction models. The only logic
that is different for different models is how we construct the prompt,
send the request, and parse the output.
This PR also changes the behavior of the staff-only `zeta2` feature flag
so that in only gates your *ability* to use Zeta2, but you can still use
your local settings file to choose between different edit prediction
models/services: zeta1, zeta2, and sweep.
This PR also makes zeta1's outcome reporting and prediction-rating
features work with all prediction models, not just zeta1.
To do:
* [x] remove duplicated logic around sending cloud requests between
zeta1 and zeta2
* [x] port the outcome reporting logic from zeta to zeta2.
* [x] get the "rate completions" modal working with all EP models
* [x] display edit prediction diff
* [x] show edit history events
* [x] remove the original `zeta` crate.
Release Notes:
- N/A
---------
Co-authored-by: Agus Zubiaga <agus@zed.dev>
Co-authored-by: Ben Kunkle <ben@zed.dev>
Closes https://github.com/zed-industries/zed/issues/42360
If updating a language server takes longer than 10 seconds, we now fall
back to launching the currently installed version (if exists) and
continue downloading the update in the background.
Release Notes:
- Improved language server updates for slow connection, now Zed launches
existing server if the update is taking too long.
---------
Co-authored-by: Piotr Osiewicz <24362066+osiewicz@users.noreply.github.com>
Deals with https://github.com/zed-industries/zed/issues/5259
Highlights brackets with different colors based on their depth.
Uses existing tree-sitter queries from brackets.scm to find brackets,
uses theme's accents to color them.
https://github.com/user-attachments/assets/cc5f3aba-22fa-446d-9af7-ba6e772029da
1. Adds `colorize_brackets` language setting that allows, per language
or globally for all languages, to configure whether Zed should color the
brackets for a particular language.
Disabled for all languages by default.
2. Any given language can opt-out a certain bracket pair by amending the
brackets.scm like `("\"" @open "\"" @close) ` -> `(("\"" @open "\""
@close) (#set! rainbow.exclude))`
3. Brackets are using colors from theme accents, which can be overridden
as
```jsonc
"theme_overrides": {
"One Dark": {
"accents": ["#ff69b4", "#7fff00", "#ff1493", "#00ffff", "#ff8c00", "#9400d3"]
}
},
```
Release Notes:
- Added bracket colorization (rainbow brackets) support. Use
`colorize_brackets` language setting to enable.
---------
Co-authored-by: MrSubidubi <dev@bahn.sh>
Co-authored-by: Lukas Wirth <lukas@zed.dev>
Co-authored-by: MrSubidubi <finn@zed.dev>
Co-authored-by: Lukas Wirth <me@lukaswirth.dev>
Co-authored-by: Smit Barmase <heysmitbarmase@gmail.com>
Closes#39088
Release Notes:
- Fixed AI assistant edits being scrambled when file was modified while
it was open
--
Co-authored-by: Conrad Irwin <conrad.irwin@gmail.com>
This PR introduces a new `MultiBufferOffset` new type wrapping size. The
goal of this is to make it clear at the type level when we are
interacting with offsets of a multi buffer versus offsets of a language
/ text buffer. This improves readability of things quite a bit by making
it clear what kind of offsets one is working with while also reducing
accidental bugs by using the wrong kin of offset for the wrong API.
This PR also uncovered two minor bugs due to that.
Does not yet introduce the MultiBufferPoint equivalent, that is for a
follow up PR.
Release Notes:
- N/A *or* Added/Fixed/Improved ...
We now correctly use bracket ranges from the deepest syntax layer when
finding enclosing brackets.
Release Notes:
- Fixed an issue where `MoveToEnclosingBracket` didn’t work correctly
inside Markdown code blocks.
- Fixed an issue where unmatched forward/backward Vim motions didn’t
work correctly inside Markdown code blocks.
---------
Co-authored-by: MuskanPaliwal <muskan10112002@gmail.com>
Release Notes:
- Adds support for `:[range]w {file}`
- This writes the lines in the range to the specified
- Adds support for `:[range]w`
- This replaces the current file with the selected lines
Fixes a regression caused by #42135 where LSP and DAP binaries weren't
being used from `PATH` env var
Now we absolutize the path if (path is relative AND (path has multiple
components OR path exists in worktree)).
- Relative paths with multiple components might not exist in the
worktree because they are ignored. Paths with a single component will at
least have an entry saying that they exist and are ignored.
- Relative paths with multiple components will never use the `PATH` env
var, so they can be safely absolutized
Release Notes:
- N/A
Since we removed the filtering step during context gathering, we want
the model to perform more targeted searches. This PR tweaks search tool
schema allowing the model to search within syntax nodes such as `impl`
blocks or methods.
This is what the query schema looks like now:
```rust
/// Search for relevant code by path, syntax hierarchy, and content.
#[derive(Debug, Clone, Serialize, Deserialize, JsonSchema)]
pub struct SearchToolQuery {
/// 1. A glob pattern to match file paths in the codebase to search in.
pub glob: String,
/// 2. Regular expressions to match syntax nodes **by their first line** and hierarchy.
///
/// Subsequent regexes match nodes within the full content of the nodes matched by the previous regexes.
///
/// Example: Searching for a `User` class
/// ["class\s+User"]
///
/// Example: Searching for a `get_full_name` method under a `User` class
/// ["class\s+User", "def\sget_full_name"]
///
/// Skip this field to match on content alone.
#[schemars(length(max = 3))]
#[serde(default)]
pub syntax_node: Vec<String>,
/// 3. An optional regular expression to match the final content that should appear in the results.
///
/// - Content will be matched within all lines of the matched syntax nodes.
/// - If syntax node regexes are provided, this field can be skipped to include as much of the node itself as possible.
/// - If no syntax node regexes are provided, the content will be matched within the entire file.
pub content: Option<String>,
}
```
We'll need to keep refining this, but the core implementation is ready.
Release Notes:
- N/A
---------
Co-authored-by: Ben <ben@zed.dev>
Co-authored-by: Max <max@zed.dev>
Co-authored-by: Max Brunsfeld <maxbrunsfeld@gmail.com>
Closes#40576
This PR makes Conda activation configurable and transparent by adding a
`terminal.detect_venv.on.conda_manager` setting (`"auto" | "conda" |
"mamba" | "micromamba"`, default `"auto"`), updating Python environment
activation to honor this preference (or the detected manager executable)
and fall back to `conda` when necessary.
The preference is passed via `ZED_CONDA_MANAGER` from the terminal
settings, and the activation command is built accordingly (with proper
quoting for paths). Changes span
`zed/crates/terminal/src/terminal_settings.rs` (new `CondaManager` and
setting), `zed/crates/project/src/terminals.rs` (inject env var),
`zed/crates/languages/src/python.rs` (activation logic), and
`zed/assets/settings/default.json` (document the setting). Default
behavior remains unchanged for most users while enabling explicit
selection of `mamba` or `micromamba`.
Release Notes:
- Added: terminal.detect_venv.on.conda_manager setting to choose the
Conda manager (auto, conda, mamba, micromamba). Default: auto.
- Changed: Python Conda environment activation now respects the
configured manager, otherwise uses the detected environment manager
executable, and falls back to conda.
- Reliability: Activation commands quote manager paths to handle spaces
across platforms.
- Compatibility: No breaking changes; non-Conda environments are
unaffected; remote terminals are supported.
---------
Co-authored-by: Lukas Wirth <me@lukaswirth.dev>
Co-authored-by: Lukas Wirth <lukas@zed.dev>
Closes#40632
**TL;DR:** The `wrap selections in tag` action was unavailable in ERB
files, even when the cursor was positioned in HTML content (outside of
Ruby code blocks). This happened because `syntax_layer_at()` incorrectly
returned the Ruby language for positions that were actually in HTML.
**NOTE:** I am not familiar with that part of Zed so it could be that
the fix here is completely incorrect.
Previously, `syntax_layer_at` incorrectly reported injected languages
(e.g., Ruby in ERB files) even when the cursor was in the base language
content (HTML). This broke actions like `wrap selections in tag` that
depend on language-specific configuration.
The issue had two parts:
1. Missing start boundary check: The filter only checked if a layer's
end was after the cursor (`end_byte() > offset`), not if it started
before, causing layers outside the cursor position to be included. See
the `BEFORE` video: when I click on the HTML part it reports `Ruby`
language instead of `HTML`.
2. Wrong boundary reference for injections: For injected layers with
`included_sub_ranges` (like Ruby code blocks in ERB), checking the root
node boundaries returned the entire file range instead of the actual
injection ranges.
This fix:
- Adds the containment check using half-open range semantics [start,
end) for root node boundaries. That ensures proper reporting of the
detected language when a cursor (`|`) is located right after the
injection:
```
<body>
<%= yield %>|
</body>
```
- Checks `included_sub_ranges` for injected layers to determine if the
cursor is actually within an injection
- Falls back to root node boundaries for base layers without sub-ranges.
This is the original behavior.
Fixes ERB language support where actions should be available based on
the cursor's actual language context. I think that also applies to some
other template languages like HEEX (Phoenix) and `*.pug`. On short
videos below you can see how I navigate through the ERB template and the
terminal on the right outputs the detected language if you apply the
following patch:
```diff
diff --git i/crates/editor/src/editor.rs w/crates/editor/src/editor.rs
index 15af61f5d2..54a8e0ae37 100644
--- i/crates/editor/src/editor.rs
+++ w/crates/editor/src/editor.rs
@@ -10671,6 +10671,7 @@ impl Editor {
for selection in self.selections.disjoint_anchors_arc().iter() {
if snapshot
.language_at(selection.start)
+ .inspect(|language| println!("Detected language: {:?}", language))
.and_then(|lang| lang.config().wrap_characters.as_ref())
.is_some()
{
```
**Before:**
https://github.com/user-attachments/assets/3f8358f4-d343-462e-b6b1-3f1f2e8c533d
**After:**
https://github.com/user-attachments/assets/c1b9f065-1b44-45a2-8a24-76b7d812130d
Here is the ERB template:
```
<!DOCTYPE html>
<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
<style>
/* Email styles need to be inline */
</style>
</head>
<body>
<%= yield %>
</body>
</html>
```
Release Notes:
- N/A
When doing a project wide search in zed on windows for `hang`, zed
starts to freeze for a couple seconds ultimately starting to error with
`Not enough quota is available to process this command.` when
dispatching windows messages. The cause for this is that we simply
overload the windows message pump due to the sheer amount of foreground
tasks we spawn when we populate the project search.
This PR is an attempt at reducing this.
Release Notes:
- Reduced hangs and stutters in large project file searches
```json
"edit_predictions": {
"codestral": {
"api_url": "https://codestral.mistral.ai",
"model": "codestral-latest",
"max_tokens": 150
}
},
```
Release Notes:
- Added support for changing the Codestral endpoint. This was discussed
at #34371.
---------
Co-authored-by: Bennet Bo Fenner <bennetbo@gmx.de>
Closes#40581
Release Notes:
- git: No longer save clean files when staging (to avoid triggering
unnecessary rebuilds in external file watchers like vite)
Closes https://github.com/zed-industries/zed/issues/40047
Closes https://github.com/zed-industries/zed/issues/24798
Closes https://github.com/zed-industries/zed/issues/24788
Before, each editor, even if it's the same buffer split in 2, was
querying for inlay hints separately, and storing the whole inlay hint
twice, in `Editor`'s `display_map` and its `inlay_hint_cache` fields.
Now, instead of `inlay_hint_cache`, each editor maintains a minimal set
of metadata (which area was queried by what task) instead, and all LSP
inlay hint data had been moved into `LspStore`, both local and remote
flavors store the data.
This allows Zed, as long as a buffer is open, to reuse the inlay hint
data similar to how document colors and code lens are now stored and
reused.
Unlike other reused LSP data, inlay hints data is the first one that's
possible to query by document ranges and previous version had issue with
caching and invalidating such ranges already queried for.
The new version re-approaches this by chunking the file into row ranges,
which are queried based on the editors' visible area.
Among the corresponding refactoring, one notable difference in inlays
display are multi buffers: buffers in them are not
[registered](https://microsoft.github.io/language-server-protocol/specifications/lsp/3.17/specification/#textDocument_didOpen)
in the language server until a caret/selection is placed inside their
excerpts inside the multi buffer.
New inlays code does not query language servers for unregistered
buffers, as servers usually respond with empty responses or errors in
such cases.
Release Notes:
- Reworked inlay hints to be less error-prone
---------
Co-authored-by: Lukas Wirth <lukas@zed.dev>
Co-authored-by: dino <dinojoaocosta@gmail.com>
Co-authored-by: Lukas Wirth <me@lukaswirth.dev>
This simplifies some code and is also more correct in some others (I
believe some of these might've overflowed causing panics in sentry)
Release Notes:
- N/A *or* Added/Fixed/Improved ...
This means that existence of activation scripts for venv/virtualenv will
be checked locally either on the host if editing locally, or the remote
by the remote proxy if editing a remote project.
Closes https://github.com/zed-industries/zed/issues/40263
Release Notes:
- N/A
- Notable change is the use of a newtype for `ReplicaId`
- Fixes `WorktreeStore::create_remote_worktree` creating a remote
worktree with the local replica id, though this is not currently used
- Fixes observing the `Agent` (that is following the agent) causing
global clocks to allocate 65535 elements
- Shrinks the size of `Global` a bit. In a local or non-collab remote
session it won't ever allocate still.
Release Notes:
- N/A *or* Added/Fixed/Improved ...
We've been considering removing workspace-hack for a couple reasons:
- Lukas ran into a situation where its build script seemed to be causing
spurious rebuilds. This seems more likely to be a cargo bug than an
issue with workspace-hack itself (given that it has an empty build
script), but we don't necessarily want to take the time to hunt that
down right now.
- Marshall mentioned hakari interacts poorly with automated crate
updates (in our case provided by rennovate) because you'd need to have
`cargo hakari generate && cargo hakari manage-deps` after their changes
and we prefer to not have actions that make commits.
Currently removing workspace-hack causes our workspace to grow from
~1700 to ~2000 crates being built (depending on platform), which is
mainly a problem when you're building the whole workspace or running
tests across the the normal and remote binaries (which is where
feature-unification nets us the most sharing). It doesn't impact
incremental times noticeably when you're just iterating on `-p zed`, and
we'll hopefully get these savings back in the future when
rust-lang/cargo#14774 (which re-implements the functionality of hakari)
is finished.
Release Notes:
- N/A
A small follow-up to the settings refactor of a few weeks ago to move
all the VSCode settings imports
to one place.
This should make it easier to spot missing imports, and easier to test
the importer.
Release Notes:
- N/A
Closes#40334
This reverts the change made in #39983, and includes a replacement
migration that will transform formatter settings values consisting of
only `code_action` format steps into the previously deprecated
`code_actions_on_format` in an attempt to restore the behavior to what
it was before the migration that deprecated `code_actions_on_format`.
This PR will result in a modified order in the `code_actions_on_format`
setting if it existed, however the decision was made to explicitly
ignore this for now, as this PR is primarily targeting users who have
already had the deprecation migration run, and no longer have the
`code_actions_on_format` key
Release Notes:
- Fixed an issue with a settings migration that deprecated the
`code_actions_on_format` setting. The `code_actions_on_format` setting
has been un-deprecated, and affected users will have the bad migration
rolled back with an updated migration
---------
Co-authored-by: Cole Miller <cole@zed.dev>
Co-authored-by: Mikayla Maki <mikayla.c.maki@gmail.com>
Co-authored-by: HactarCE <6060305+HactarCE@users.noreply.github.com>
Follow up for: #39983 and
https://github.com/zed-industries/zed/pull/40040#issuecomment-3393902691
Previously it was possible to have formatting done using prettier or
language server using `"formatter": "auto"` and specify code actions to
apply on format using the `"code_actions_on_format"` setting. However,
post #39983 this is no longer possible due to the removal of the
`"code_actions_on_format"` setting. To rectify this regression, this PR
makes it so that the `"auto"` and `"language_server"` strings that were
previously only allowed as top level values on the `"formatter"` key,
are now allowed as format steps like so:
```json
{
"formatter": ["auto", "language_server"]
}
```
Therefore to replicate the previous behavior using `"auto"` and
`"code_actions_on_format"` you can use the following configuration:
```json
{
"formatter": [{"code_action": ...}, "auto"]
}
```
Release Notes:
- N/A *or* Added/Fixed/Improved ...
Closes #ISSUE
Release Notes:
- settings: Deprecated `code_actions_on_format` in favor of specifying
code actions to run on format inline in the `formatter` array.
Previously, you would configure code actions to run on format like this:
```json
{
"code_actions_on_format": {
"source.organizeImports": true,
"source.fixAll.eslint": true
}
}
```
This has been migrated to the new format:
```json
{
"formatter": [
{
"code_action": "source.organizeImports"
},
{
"code_action": "source.fixAll.eslint"
}
]
}
```
This change will be automatically migrated for you. If you had an
existing `formatter` setting, the code actions are prepended to your
formatter array (matching the existing behavior). This migration applies
to both global settings and language-specific settings
Release Notes:
- Added Codestral edit predictions provider which can be enabled by adding an API key in the Mistral section of agent settings.
## Config
Get API key from https://console.mistral.ai/codestral and add it in the Mistral section of the agent settings.
```
"features": {
"edit_prediction_provider": "codestral"
},
"edit_predictions": {
"codestral": {
"model": "codestral-latest",
"max_tokens": 150
}
},
```
---------
Co-authored-by: Michael Sloan <michael@zed.dev>
