Closes#42303
Ollama added tool call identifiers
(https://github.com/ollama/ollama/pull/12956) in its latest version
[v0.12.10](https://github.com/ollama/ollama/releases/tag/v0.12.10). This
broke our json schema and made all tool calls fail.
This PR fixes the schema and uses the Ollama provided tool call
identifier when available. We remain backwards compatible and still use
our own identifier with older versions of Ollama. I added a `TODO` to
remove the `Option` around the new field when most users have updated
their installations to v0.12.10 or above.
Note to reviewer: The fix to this issue should likely get cherry-picked
into the next release, since Ollama becomes unusable as an agent without
it.
Release Notes:
- Fixed tool calling when using the latest version of Ollama
Release Notes:
- N/A
According to
[microsoft/vscode#254689](https://github.com/microsoft/vscode/issues/254689),
the json.schemastore.org domain has been deprecated and should now use
www.schemastore.org (or schemastore.org) instead.
This PR updates all occurrences of the old domain within the Zed
codebase,
including code, documentation, and configuration files.
The auto update helper already removes these when successful, so these
will always fail in the common case.
Additional replaces a mutable const with a static as otherwise we'll
rebuild the job list on every access
Release Notes:
- N/A *or* Added/Fixed/Improved ...
Lychee is a [Rust based](https://lychee.cli.rs) async parallel link
checker.
I ran it against the codebase to suss out stale links and fixed those
up.
There's currently 2 remaining cases that I don't know how to resolve:
1. https://flathub.org/apps/dev.zed.Zed - nginx is giving a 502 bad
gateway
2.
https://github.com/zed-industries/zed/actions/workflows/ci.yml/badge.svg
- I don't want to mess with the CI pipeline in this PR.
Once again, I'll punt to the Docs Czar to see if this gets incorporated
into CI later.
---
## Running `lychee` locally:
```
cargo binstall -y lychee
lychee .
```
---
Release Notes:
- N/A
Signed-off-by: mrg <miguelraz@ciencias.unam.mx>
Closes#36376
Problem:
Multi-cursor edits/selections in multi-buffers view were jumping to
incorrect locations after toggling buffer folds. When users created
multiple selections across different buffers in a multi-buffer view
(like project search results) and then folded one of the buffers,
subsequent text insertion would either:
1. Insert text at wrong locations (like at the top of the first unfolded
buffer)
2. Replace the entire content in some buffers instead of inserting at
the intended cursor positions
3. Create orphaned selections that caused corruption in the editing
experience
The issue seems to happen because when a buffer gets folded in a
multi-buffer view, the existing selections associated with that buffer
become invalid anchor points.
Solution:
1. Selection Cleanup on Buffer Folding
- Added `remove_selections_from_buffer()` method that filters out all
selections from a buffer when it gets folded
- This prevents invalid selections from corrupting subsequent editing
operations
- Includes edge case handling: if all selections are removed (all
buffers folded), it creates a default selection at the start of the
first buffer to prevent panics
2. Unfolding buffers before editing
- Added `unfold_buffers_with_selections()` call in `handle_input()`
ensures buffers with active selections are automatically unfolded before
editing
- This helps in fixing an edge case (covered in the tests) where, if you
fold all buffers in a multi-buffer view, and try to insert text in a
selection, it gets unfolded before the edit happens. Without this, the
inserted text would override the entire buffer content.
- If we don't care about this edge case, we could remove this method. I
find it ok to add since we already trigger buffer unfolding after edits
with `Event::ExcerptsEdited`.
Release Notes:
- Fixed multi-cursor edits jumping to incorrect locations after toggling
buffer folds in multi-buffer views (e.g, project search)
- Multi-cursor selections now properly handle buffer folding/unfolding
operations
- Text insertion no longer occurs at the wrong positions when buffers
are folded during multi-cursor editing
- Eliminated content replacement bugs where entire buffer contents were
incorrectly overwritten
- Added safe fallback behavior when all buffers in a multi-buffer view
are folded
---------
Co-authored-by: Smit Barmase <heysmitbarmase@gmail.com>
Update how the editor's `select_*` methods work in order to respect the
`search.case_sensitive` setting, or to be overriden by the
`BufferSearchBar` search options.
- Update both the `SearchableItem` and `SearchableItemHandle` traits
with a new `set_search_is_case_sensitive` method that allows callers
to set the case sensitivity of the search
- Update the `BufferSearchBar` to leverage
`SearchableItemHandle.set_search_is_case_sensitive` in order to sync
its case sensitivity options with the searchable item
- Update the implementation of the `SearchableItem` trait for `Editor`
so as to store the argument provided to the
`set_search_is_case_sensitive` method
- Update the way search queries are built by `Editor` so as to rely on
`SearchableItem.set_search_is_case_sensitive` argument, if not `None`,
or default to the editor's `search.case_sensitive` settings
Closes#41070
Release Notes:
- Improved the "Select Next Occurrence", "Select Previous Occurrence"
and "Select All Occurrences" actions in order to respect the case
sensitivity search settings
---------
Co-authored-by: Conrad Irwin <conrad.irwin@gmail.com>
- Introduce a `default_range` field to `VimCommand`, to be optionally
used when no range is specified for the command
- Update `VimCommand.parse` to take into consideration the
`default_range`
- Introduce `CommandRange::buffer` to obtain the `CommandRange` which
corresponds to the whole buffer
- Update the `VimCommand` definitions for both `sort` and `sort i` to
default to the whole buffer when no range is specified
Closes#41750
Release Notes:
- Improved vim's `:sort` command to sort the buffer's content when no
selection is used
Adds a simple notification when cloning a repo using the integrated git
clone on Zed. Before this, the user had no feedback after starting the
cloning action.
Demo:
https://github.com/user-attachments/assets/72fcdf1b-fc99-4fe5-8db2-7c30b170f12f
Not sure about that icon I'm using for the animation, but that can be
easily changed.
Release Notes:
- Added notification when cloning a repo from zed
We've had several outages with a proximate cause of "vercel is
complicated",
and auto-update is considered a critical feature; so lets not use vercel
for
that.
Release Notes:
- Auto Updates (and remote server binaries) are now downloaded via
https://cloud.zed.dev instead of https://zed.dev. As before, these URLs
redirect to the GitHub release for actual downloads.
Similar to #42246, we'd like to avoid having Vercel on the critical
path.
https://zed.dev/install.sh is served from Cloudflare by intercepting a
route on that page, so this makes the shell-based install flow vercel independent.
Release Notes:
- `./script/install.sh` will now fetch assets via
`https://cloud.zed.dev/`
instead of `https://zed.dev`. As before it will redirect to GitHub
releases
to complete the download.
Closes https://github.com/zed-industries/zed/issues/42312
The issue here was that we assumed that context servers provided by
extensions would always need a config in the settings to be present when
actually the opposite was the case - context servers provided by
extensions are the only context servers that do not need a config to be
in place in order to be available in the UI.
Release Notes:
- Fixed an issue where context servers provided by extensions could not
be uninstalled if they were previously unconfigured.
This PR drops the `channel_messages` table and its
dependents—`channel_message_mentions` and `observed_channel_messages`—as
they are no longer used.
Release Notes:
- N/A
Applications can select a policy for when the app quits using the new
function `Application::with_quit_mode`:
- Only on explicit calls to `App::quit`
- When the last window is closed
- Platform default (former on macOS, latter everywhere else)
Release Notes:
- N/A
There was a TODO in `crates/settings/src/settings_content/theme.rs` to
make this rename.
This PR is just splitting off this change from
https://github.com/zed-industries/zed/pull/40035 to make reviewing that
one a bit easier since that PR is a bit more involved than expected.
Release Notes:
- N/A
Signed-off-by: Connor Tsui <connor.tsui20@gmail.com>
Document "any-typed" (`serde_json::Value`) "lsp" keys to include them in
json-language-server completions.
The vscode-json-languageserver seems to skip generically typed keys when
offering completion.
For this schema
```
"LspSettings": {
"type": "object",
"properties": {
...
"initialization_options": true,
...
}
}
```
"initialization_options" is not offered in the completion.
The effect is easy to verify by triggering completion inside:
```
"lsp": {
"basedpyright": {
COMPLETE HERE
```
<img width="797" height="215" alt="image"
src="https://github.com/user-attachments/assets/d1d1391c-d02c-4028-9888-8869f4d18b0f"
/>
By adding a documentation string the keys are offered even if they are
generically typed:
<img width="809" height="238" alt="image"
src="https://github.com/user-attachments/assets/9a072da9-961b-4e15-9aec-3d56933cbe67"
/>
---
Note: I did some cursory research of whether it's possible to make
vscode-json-languageserver change behavior without success. IMO, not
offering completions here is a bug (or at minimal should be
configurable)
---
Release Notes:
- N/A
---------
Co-authored-by: Kirill Bulatov <mail4score@gmail.com>
The "Option" key only exists on a Mac. On other operating systems, it is
always expected that the Alt key generates escaped characters.
Fixes https://github.com/zed-industries/zed/issues/40583
Release Notes:
- N/A
Closes ZED-1SV
Closes ZED-TG
Closes ZED-22G
Closes ZED-22J
This seems to fix the reported error there, but ultimately, this might
benefit from a test to reproduce. Hence, marking as draft for now.
Release Notes:
- Fixed a rare panic whilst wrapping lines.
Closes#42191
This PR adds support for relative line numbers in deleted hunks. Note
that this only applies in cases where there is a form of relative
numbering.
It also adds some tests for this functionality as well as missing tests
for other cases in line layouting that was previously untested.
Release Notes:
- Line numbers will now be shown in deleted git hunks if relative line
numbering is enabled
This PR creates a dedicated function for the agent panel message
editor's placeholder text so that we can wait for the agent
initialization to capture whether they support slash commands or not. On
the one (nice) hand, this allow us to stop matching agents by name and
make this a bit more generic. On the other (bad) hand, the "/ for
commands" bit should take a little second to show up because we can only
know whether an agent supports it after it is initialized.
This is particularly relevant now that we have agents coming from
extensions and for them, we would obviously not be able to match by
name.
Release Notes:
- agent: Fixed agent panel message editor's placeholder text by making
it more accurate as to whether agents support slash commands,
particularly those coming from extensions.
This hopefully resolves the lingering test failures on linux,
but also adds some logging just in case this isn't the problem...
Release Notes:
- N/A
---------
Co-authored-by: Ben Kunkle <ben@zed.dev>
Closes#41214
Release Notes:
- Added support for relative paths in LSP and DAP binaries
---------
Co-authored-by: Cole Miller <cole@zed.dev>
Co-authored-by: Julia Ryan <juliaryan3.14@gmail.com>
Closes https://github.com/zed-industries/zed/issues/42356
This PR fixes the history view losing focus by simply always displaying
the search editor. I don't think it's too weird to not have it when it's
empty, and it also ends up matching how regular pickers work.
Release Notes:
- agent: Fixed a bug where navigating the agent panel with the keyboard
wouldn't work if you visited the history view and it was empty/had no
entries.
Update `GitStore.on_buffer_store_event` so that, when a
`BufferStoreEvent::BufferChangedFilePath` event is received, we check if
there's any diff state for the buffer and, if so, update it according to
the new file path, in case the file exists in the repository.
Closes#40499
Release Notes:
- Fixed issue with git diff tracking when updating a buffer's file from
an untracked to a tracked file
We'll now cache LLM responses at the request level (by hash of
URL+contents) for both context and prediction. This way we don't need to
worry about mistakenly using the cache when we change the prompt or its
components.
Release Notes:
- N/A
---------
Co-authored-by: Oleksiy Syvokon <oleksiy.syvokon@gmail.com>
New error variants were stabilized in 1.83, and this might've led to us
mis-handling not-a-directory errors.
Co-authored-by: Dino <dino@zed.dev>
Release Notes:
- N/A
Co-authored-by: Dino <dino@zed.dev>
Closes#42178
Now it's consistent with the DateAndTime path which already does
timezone conversion.
- **Future Work**
Happy to tackle the TODO in `time_format.rs` about implementing native
Windows APIs for proper localized formatting (similar to macOS's
`CFDateFormatter`) as a follow-up.
Release Notes:
- agent: Fixed the thread history item timestamp, which was being shown
in GMT instead of in the user's local timezone on Windows.
Bennet Bo Fenner
