This PR expands our automatic retry behavior for certain classes of
completion errors (e.g., rate limit errors).
Previously this was only available when using burn mode.
We now auto-retry when:
- Using the Zed provider while on a token-based plan
- Using the Zed provider while on a legacy plan with burn mode enabled
- Using a non-Zed provider
Release Notes:
- Expanded automatic retry behavior for errors in the Agent. Errors
classified as "retryable" (such as rate limit errors) will now
automatically be retried when:
- Using the Zed provider while on a token-based plan
- Using the Zed provider while on a legacy plan with burn mode enabled
- Using a non-Zed provider
---------
Co-authored-by: David Kleingeld <davidsk@zed.dev>
These changes update the way the file finder decides wether to only look
for an absolute path or for a relative path too.
When the provided query started with a slash (`/`) the file finder would
assume this to be an absolute path so would always try to find an
absolute path and return no matches if none was found. This is meant to
support situtations where, for example, a CLI tool might output the
absolute path of a file and the user can copy and paste that in the file
finder.
However, it's should be possible to use slash (`/`) at the start of the
query to specify that only relative files inside a folder should be
matched, which would not work in this scenario.
With these changes, the file finder will first check if the path is
absolute and, if it is and no absolute matches were found, it'll still
try to find relative matches, otherwise it'll simply look for relative
matches.
Closes#39350
Release Notes:
- Fixed project files matches when using slash (`/`) at the start in
order to consider relative paths
---------
Co-authored-by: Piotr Osiewicz <piotr@zed.dev>
Co-authored-by: Piotr Osiewicz <24362066+osiewicz@users.noreply.github.com>
We don't get an ExitStatus from a remote terminal, so this check was
failing.
Ideally we move all of this to just needing an exit code, but we will
have to revisit that later.
Release Notes:
- N/A
Update Vim's `%` motion to first attempt finding the exact matching
bracket/tag under the cursor, then fall back to the previous
nearest-enclosing logic if none is found. This prevents accidentally
jumping to nested pairs in languages like TSX and Svelte where `<>`,
`</>`, and `/>` are also treated as brackets.
Closes#39368
Release Notes:
- Fixed an edge case with the `%` motion in vim, where the cursor could
end up in a closing HTML tag instead of the matching bracket
This commit fixes an issue where opening zed using `--user-data-dir`
with an empty directory would cause the first run to display a "Failed
to migrate settings" error.
This was caused by the migrator attempting to migrate an empty string,
so if that's the case, we'll simply return `Ok(None)` and avoid
attempting to migrate anything at all.
Relates to #39400
Release Notes:
- N/A
Co-authored-by: Smit Barmase <heysmitbarmase@gmail.com>
Use path! macro for platform-specific path formatting in test
assertions, fixing hardcoded Unix-style paths that failed on Windows.
Release Notes:
- N/A
Signed-off-by: Xiaobo Liu <cppcoffee@gmail.com>
This PR adds support for HTML block quotes, that also allows you to have
nested variant of it.
<img width="1441" height="804" alt="Screenshot 2025-10-08 at 10 25 57"
src="https://github.com/user-attachments/assets/4e1da766-fb54-4e87-8654-1ea14330bc97"
/>
Code example used in screenshot:
```html
<blockquote>
<p>
Words can be like X-rays, if you use them properly—they’ll go through
anything. You read and you’re pierced.
</p>
<blockquote>
<p>
lorem ipsum dolor sit amet, consectetur adipiscing elit. Sed non risus. Suspendisse lectus tortor, dignissim sit amet, adipiscing nec, ultricies sed, dolor.
</p>
</blockquote>
</blockquote>
```
Release Notes:
- Markdown: Added support for `HTML` block quotes
Follow up to https://github.com/zed-industries/zed/pull/39468.
Unlike `agent_ui_font_size`, the `agent_buffer_font_size` setting does
have a default value, which means it does not fall back to the regular
UI font size, but rather to its default value.
Release Notes:
- N/A
Closes#38538
Release Notes:
- Fixed an issue where opening a remote terminal failed on systems like
BusyBox, Alpine, Amazon Linux 2, some CentOS images, etc., due to an
invalid option 'C'.
This PR makes selecting a sub-entry in the settings UI nav bar scroll to
that section in the settings page. It also updates the selected
sub-entry when scrolling through a settings page to match what a user is
viewing on the page.
I also added a new helper method to `ScrollHandle` type called
`scroll_to_top_of_item` that scrolls until an item is the top element
visible.
Release Notes:
- N/A
Closes#38919
Now, when unfocusing the filename editor while creating a file or
directory in the project panel, it will create it by default unless the
name is empty or already exists.
Release Notes:
- Improved behavior where unfocusing while creating a new file or
directory in the project panel now creates it instead of discarding it.
The previous modifier detection treated `AltGr` presses as `Ctrl+Alt`,
which broke entering characters produced by AltGr. For example, on a
Hungarian layout `{` is typed with `AltGr+B`; our code saw that as
`Ctrl+Alt+B` and the keybind took precedence, so the character couldn’t
be entered.
On Windows, AltGr isn’t a first-class modifier. It’s emulated as a
combination of `Right Alt (VK_RMENU)` plus a synthetic `Left Ctrl
(VK_LCONTROL)` press. When users press AltGr, `GetKeyState` reports both
Ctrl and Alt as down, which makes AltGr indistinguishable from a real
`Ctrl+Alt` chord if we only look at aggregate modifier state.
Fix: detect the AltGr pattern by checking `VK_RMENU && VK_LCONTROL`.
When that pattern is present, treat it as text-entry intent and suppress
`control` and `alt` in `current_modifiers()`. This prevents
AltGr-produced characters from colliding with `Ctrl+Alt` keybinds while
keeping other modifiers intact.
Limitation: there is no Windows API to tell whether the active layout
actually has AltGr. As a result, on non-AltGr layouts (e.g. US),
pressing `Right Alt + Left Ctrl` will be interpreted as AltGr and will
not trigger `Ctrl+Alt` keybinds. This is an acceptable trade-off to
ensure AltGr layouts can reliably enter characters; users can still
invoke `Ctrl+Alt` keybinds using `Left Alt` or by choosing bindings that
avoid common AltGr pairs.
I based this on https://github.com/zed-industries/zed/pull/36115 after
trying other different approaches, but this one is a bit more specific.
Does this approach make sense, or is slightly breaking US input in favor
of fixing international input a no-go? I think the benefit - being able
to type certain characters _at all_ - outweighs the shortcomings.
Otherwise, there's a way to detect if the keyboard layout uses AltGr or
not, but it's quite hacky, and involves reading the registry to find the
current layout dll's name, opening that dll, manually declaring struct
layouts that it uses, then parsing out the AltGr flag from a function
call result. I don't think that's worth it, but if needed, I can give
that a shot, let me know.
Release Notes:
- windows: Fixed handling of AltGr to avoid keybinds preventing
character input
Reverts zed-industries/zed#39581
This has done its job uncovering incorrect constructions of the
highlight ranges pretty fast. Reverting this to prevent this from
spilling into preview until I can fix the call sites next week
Closes#39007
Strings should be escaped with backticks in PowerShell, so the following
```
\"pwsh.exe -C pytest -m \\\"some_test\\\"\"
```
becomes
```
\"pwsh.exe -C pytest -m `\"some_test`\"\"
```
Otherwise PowerShell will misinterpret the invocation resulting in
weirdness all-around such as the issue linked above.
Release Notes:
- N/A
Follow-up to #39649, possible fix for #39669
This implements an alternate strategy for showing/hiding the agent panel
in response to `disable_ai`. We don't load the panel at all if AI is
disabled at startup, and when the value of `disable_ai` changes, we load
the panel or destroy it as needed.
Release Notes:
- N/A
Closes#39028
Fixed empty lines appearing when collapsing files with diagnostic
messages in the diagnostics panel.
Added a flag to track when processing a `FoldedBuffer` and skip
`Near/Below` blocks (diagnostic messages) that immediately follow it.
This prevents diagnostics from rendering as empty lines when their file
is collapsed.
Before:
<img width="1489" height="429" alt="before"
src="https://github.com/user-attachments/assets/5e233290-1f6e-403c-a6b3-a65107586d01"
/>
After:
<img width="981" height="270" alt="after"
src="https://github.com/user-attachments/assets/a877b651-6b7f-4441-805c-38ea41e73a18"
/>
Release Notes:
- Fixed empty lines when collapsing files with diagnostics in the
diagnostics panel
This PR fixes an issue where scrollbars would overagressively capture
some events, which could lead to clicks being lost in the process. Also
improves how hovering of the parent is detected to lead to less false
positives.
Release Notes:
- Fixed a rare issue where scrollbars would react to and capture events
they should not react to.
Fixes ZED-1NX
This panic could occur when an `bulk_staging` was set to `Some(...)` and
`sort_by_path` was set to `true`.
When setting `sort_by_path: true`, we call `update_visible_entries(...)`
which then checks if `bulk_staging ` is `Some(...)` and calls
`entry_by_path`. That function accesses `entries`, which still consists
of both headers and entries. But the code
(`entry.status_entry().unwrap()`) assumes that there are no headers in
the entry list if `sort_by_path: true`.
```rust
if GitPanelSettings::get_global(cx).sort_by_path {
return self
.entries
.binary_search_by(|entry| entry.status_entry().unwrap().repo_path.cmp(path)) //This unwrap() would panic
.ok();
}
```
This has now been fixed by clearing all the entries when `sort_by_path`
changes, as this is the only case where our assumptions are invalid. I
also added a test which 1) actually tests the sort_by_path logic 2)
ensures that we do not re-introduce this panic in the future.
Release Notes:
- Fixed a panic that could occur when using `sort_by_path: true` in the
git panel
- **paths: Cache away results of static construction of RelPath**
- **agent: Cache away results of converting rules file names into
relpaths**
This PR fixed a regression from relpath PR where we've started doing
more work when working with static (Rel-)Paths.
Release Notes:
- N/A
Prior we only logged the crate in `log_err`, which is not too helpful.
We now assemble the module path from the file system path.
Release Notes:
- N/A *or* Added/Fixed/Improved ...
