This fixes two things:
- Go-to-def to absolute paths (i.e. opening stdlib files) multiple times
(opening, dropping, and re-opening worktrees)
- Re-opening abs paths from the file picker history that were added
there by go-to-def
Release Notes:
- N/A
---------
Co-authored-by: Bennet <bennet@zed.dev>
* `Open Excerpts` command always opens the locations in the base buffer
* LSP features like document-highlights, go-to-def, and inlay hints work
correctly in branch buffers
* Other LSP features like completions, code actions, and rename are
disabled in branch buffers
Release Notes:
- N/A
Reworks the way tasks are stored, accessed and synchronized in the
`project`.
Now both collab and ssh remote projects use the same TaskStorage kind to
get the task context from the remote host, and worktree task templates
are synchronized along with other worktree settings.
Release Notes:
- Adds ssh support to tasks, improves collab-remote projects' tasks sync
As @maan2003 noted in #18473, we should warn the user if direnv call
fails
Release Notes:
- Show a notice in the activity indicator if an error occurs while
loading the shell environment
Closes#18261
This makes sure that we find external formatters in the project
environment.
TODO:
- [x] Use a different type for the triplet of `(buffer_handle,
buffer_path, buffer_env)`. Something like `FormattableBuffer`.
- [x] Test this!!
Release Notes:
- Fixed external formatters not being found, even when they were
available in the `$PATH` of a project.
---------
Co-authored-by: Bennet <bennet@zed.dev>
Closes#11991
Release Notes:
- Added support for color swatches for language server completions.
<img width="502" alt="Screenshot 2024-10-02 at 19 02 22"
src="https://github.com/user-attachments/assets/57e85492-3760-461a-9b17-a846dc40576b">
<img width="534" alt="Screenshot 2024-10-02 at 19 02 48"
src="https://github.com/user-attachments/assets/713ac41c-16f0-4ad3-9103-d2c9b3fa8b2e">
This implementation is mostly a port of the VSCode version of the
ColorExtractor. It seems reasonable the we should support _at least_
what VSCode does for detecting color swatches from LSP completions.
This implementation could definitely be better perf-wise by writing a
dedicated color parser. I also think it would be neat if, in the future,
Zed handled _more_ color formats — especially wide-gamut colors.
There are a few differences to the regexes in the VSCode implementation
but mainly so simplify the implementation :
- The hex vs rgb/hsl regexes were split into two parts
- The rgb/hsl regexes allow 3 or 4 color components whether hsla/rgba or
not and the parsing implementation accepts/rejects colors as needed
---------
Co-authored-by: Marshall Bowers <elliott.codes@gmail.com>
The indicator turns yellow when ssh client is trying to reconnect. Note
that the state tracking is probably not ideal (we'll see how it pans out
once we start dog-fooding), but at the very least "green=good" should be
a decent mental model for now.
Release Notes:
- N/A
This PR does not change how things work for settings, but lays the
ground work for the future functionality.
After this change, Zed is prepared to sync more than just
`settings.json` files from local worktree and user config.
* ssh tasks
Part of this work is to streamline the task sync mechanism.
Instead of having an extra set of requests to fetch the task contents
from the server (as remote-via-collab does now and does not cover all
sync cases), we want to reuse the existing mechanism for synchronizing
user and local settings.
* editorconfig
Part of the task is to sync .editorconfig file changes to everyone which
involves sending and storing those configs.
Both ssh (and remove-over-collab) .zed/tasks.json and .editorconfig
files behave similar to .zed/settings.json local files: they belong to a
certain path in a certain worktree; may update over time, changing Zed's
functionality; can be merged hierarchically.
Settings sync follows the same "config file changed -> send to watchers
-> parse and merge locally and on watchers" path that's needed for both
new kinds of files, ergo the messaging layer is extended to send more
types of settings for future watch & parse and merge impls to follow.
Release Notes:
- N/A
This restructures the code in `remote` so that it's easier to replace
the current SSH connection with a new one in case of
disconnects/reconnects.
Right now, it successfully reconnects, BUT we're still missing the big
piece on the server-side: keeping the server process alive and
reconnecting to the same process that keeps the project-state.
Release Notes:
- N/A
---------
Co-authored-by: Bennet <bennet@zed.dev>
This is a follow-up to #18530 thanks to this comment here:
https://github.com/zed-industries/zed/pull/18530#issuecomment-2382870564
In short: it fixes the `blinking` setting and the `cursor_shape` setting
as it relates to blinking.
Turns out our `blinking` setting was always the wrong value when using
`terminal_controlled` and the terminal _would_ control the blinking.
Example script to test with:
```bash
echo -e "0 normal \x1b[\x30 q"; sleep 2
echo -e "1 blink block \x1b[\x31 q"; sleep 2
echo -e "2 solid block \x1b[\x32 q"; sleep 2
echo -e "3 blink under \x1b[\x33 q"; sleep 2
echo -e "4 solid under \x1b[\x34 q"; sleep 2
echo -e "5 blink vert \x1b[\x35 q"; sleep 2
echo -e "6 solid vert \x1b[\x36 q"; sleep 2
echo -e "0 normal \x1b[\x30 q"; sleep 2
echo -e "color \x1b]12;#00ff00\x1b\\"; sleep 2
echo -e "reset \x1b]112\x1b\\ \x1b[\x30 q"
```
Before the changes in here, this script would set the cursor shape and
the blinking, but the blinking boolean would always be wrong.
This change here makes sure that it works consistently:
- `terminal.cursor_shape` only controls the *default* shape of the
terminal, not the blinking.
- `terminal.blinking = on` means that it's *always* blinking, regardless
of what terminal programs want
- `terminal.blinking = off` means that it's *never* blinking, regardless
of what terminal programs want
- `terminal.blinking = terminal_controlled (default)` means that it's
blinking depending on what terminal programs want. when a terminal
program resets the cursor to default, it sets it back to
`terminal.cursor_shape` if that is set.
Release Notes:
- Fixed the behavior of `{"terminal": {"blinking":
"[on|off|terminal_controlled]"}` to work correctly and to work correctly
when custom `cursor_shape` is set.
- `terminal.cursor_shape` only controls the *default* shape of the
terminal, not the blinking.
- `terminal.blinking = on` means that it's *always* blinking, regardless
of what terminal programs want
- `terminal.blinking = off` means that it's *never* blinking, regardless
of what terminal programs want
- `terminal.blinking = terminal_controlled (default)` means that it's
blinking depending on what terminal programs want. when a terminal
program resets the cursor to default, it sets it back to
`terminal.cursor_shape` if that is set.
Demo:
https://github.com/user-attachments/assets/b3fbeafd-ad58-41c8-9c07-1f03bc31771f
Co-authored-by: Bennet <bennet@zed.dev>
I've been running with direct direnv loading for a while now and haven't
experienced any significant issues other than #18473. Making it default
would make direnv integration more reliable and consistent. I've also
updated the docs a bit to ensure that they represent current status of
direnv integration
Release Notes:
- Made direnv integration use direct (`direnv export json`) mode by
default instead of relying on a shell hook, improving consistency and
reliability of direnv detection
This builds on top of @Yevgen's #15840 and combines it with the settings
names introduced in #17572.
Closes#4731.
Release Notes:
- Added a setting for the terminal's default cursor shape. The setting
is `{"terminal": {"cursor_shape": "block"}}``. Possible values: `block`,
`bar`, `hollow`, `underline`.
Demo:
https://github.com/user-attachments/assets/96ed28c2-c222-436b-80cb-7cd63eeb47dd
Before this we we would stop loading the environment if the call to
direnv failed, which is not necessary in any way
cc @mrnugget
Release Notes:
- Fixed the environment not loading if `direnv` mode is set to `direct`
and `.envrc` is not allowed
If specifying a formatter in the settings like this:
"languages": {
"foo": {
"formatter": {
"external": {
"command": "/path/to/foo-formatter"
}
}
}
}
Zed will show an error like this:
Invalid user settings file
data did not match any variant of untagged enum SingleOrVec
This is because the arguments are not optional. The error is hard to
understand, so let's make the arguments actually optional, which makes
the above settings snippet valid.
Release Notes:
- Make external formatter arguments optional
This is not an ideal solution to
https://github.com/fasterthanlime/zed-diags-readme, but current status
quo is not great either; we were just going through all of the language
servers and notifying them, whereas we should ideally do it based on a
glob.
/cc @fasterthanlime
Release Notes:
- N/A
Fixes an issue where evals were hitting "too many open files" errors
because we were adding (and detaching) new directory watches for each
project. Now we add those watches globally/at the worktree level, and we
store the tasks so they stop watching on drop.
Release Notes:
- N/A
---------
Co-authored-by: Max <max@zed.dev>
Co-authored-by: Piotr <piotr@zed.dev>
Co-authored-by: Piotr Osiewicz <24362066+osiewicz@users.noreply.github.com>
Release Notes:
- Fixed overriding the path of a language server binary for all language
servers. `{"lsp":{"<lsp-name>":{"binary":{"path": "_"}}}}` will now work
for all language servers including those defined by extensions.
- (breaking change) To disable finding lsp adapters in your path, you
must now specify
`{"lsp":{"<lsp-name>":{"binary":{"ignore_system_version": true}}}}`.
Previously this was `{"lsp":{"<lsp-name>":{"binary":{"path_lookup":
false}}}}`. Note that this setting still does not apply to extensions.
- Removed automatic reinstallation of language servers. (It mostly
didn't work)
---------
Co-authored-by: Mikayla <mikayla@zed.dev>
This now shows an error message if you try open a project over SSH that
doesn't exist. If it's a possible file-path though, it acts like Zed's
`cli` and opens the file so that it can be created.
- Works: `cargo run ssh://127.0.0.1/~/folder-exists/file-does-not-exist`
— this will open `file-does-not-exist`
- Shows error: `cargo run
ssh://127.0.0.1/~/folder-does-not-exist/file-does-not-exist` — this will
show an error
Release Notes:
- N/A
Co-authored-by: Bennet <bennet@zed.dev>
Co-authored-by: Conrad <conrad@zed.dev>
This fixes the `SshSession` being leaked.
There were two leaks:
1. `Arc<SshSession>` itself got leaked into the `SettingsObserver` that
lives as long as the application. Fixed with a weak reference.
2. The two tasks spawned by an `SshSession` had a circular dependency
and didn't exit while the other one was running. Fixed by fixing (1)
and then attaching one of the tasks to the `SshSession`, which means
it gets dropped with the session itself, which leads the other task
to error and exit.
Co-authored-by: Bennet <bennet@zed.dev>
Release Notes:
- N/A
---------
Co-authored-by: Bennet <bennet@zed.dev>
Release Notes:
- (Potentially breaking change) Zed will now use the node installed on
your $PATH (if it is more recent than v18) instead of downloading its
own. You can disable the new behavior with `{"node":
{"disable_path_lookup": true}}` in your settings. We do not yet use
system/project-local node_modules.
---------
Co-authored-by: Mikayla <mikayla@zed.dev>
Related to https://github.com/rust-lang/rust-analyzer/pull/18167
* Declare more completion item fields in the client completion resolve
capabilities
* Do resolve completions even if their docs are present
* Instead, do not resolve completions that could not be resolved when
handling the remote client resolve requests
* Do replace the old lsp completion data with the resolved one
Release Notes:
- Improved completion resolve mechanism
This pushes the new LanguageServerName type to more places.
As both languages and language servers were identified by Arc<str>, it
was
sometimes hard to tell which was intended.
Release Notes:
- N/A
This PR introduces functionality for creating *branches* of buffers that
can be used to preview and edit change sets that haven't yet been
applied to the buffers themselves.
Release Notes:
- N/A
---------
Co-authored-by: Marshall Bowers <elliott.codes@gmail.com>
Co-authored-by: Marshall <marshall@zed.dev>
This fixes the bug with hover tooltips appearing multiple times.
Turns out everytime we receive the `CreateLanguageServer` message we'd
add a new adapter but only have a single server running for all of them.
And we send a `CreateLanguageServer` message everytime you open a
buffer.
What this does is to only add a new adapter if it hasn't already been
registered, which is also what we do locally.
Release Notes:
- N/A
Closes#16876
We only ever showed parsing errors, but not if something failed to
deserialize.
Basically, if you had a stray `,` somewhere, we'd show a notification
for user errors, but only squiggly lines if you had a `[]` instead of a
`{}`.
The squiggly lines would only show up when there were schema errors.
In the case of `formatter` settings, for example, if someone put in a
`{}` instead of `[]`, we'd never show anything.
With this change we always show a notification if parsing user or
project settings fails.
(Right now, the error message might still be bad, but that's a separate
change)
Release Notes:
- Added a notification to warn users if their user settings or
project-local settings failed to deserialize.
Demo:
https://github.com/user-attachments/assets/e5c48165-f2f7-4b5c-9c6d-6ea74f678683
This PR makes the `Buffer::apply_ops` method infallible for
`text::Buffer` and `language::Buffer`.
We discovered that `text::Buffer::apply_ops` was only fallible due to
`apply_undo`, which didn't actually need to be fallible.
Release Notes:
- N/A
