This PR brings back https://github.com/zed-industries/zed/pull/30969 and
adds some initial testing.
https://github.com/zed-industries/zed/pull/30969 did indeed allow Zed to
continue doing downloads after downloading one, but it introduced a bug
where Zed would download a new binary every time it polled, even if the
version was the same as the running instance.
This code could use a refactor to allow more / better testing, but this
is a start.
Release Notes:
- N/A
TODO:
- [x] Release a new version of `zed_llm_client`
Release Notes:
- N/A
---------
Co-authored-by: Mikayla Maki <mikayla.c.maki@gmail.com>
Co-authored-by: Antonio Scandurra <me@as-cii.com>
Co-authored-by: Ben Brandt <benjamin.j.brandt@gmail.com>
Co-authored-by: Marshall Bowers <git@maxdeviant.com>
Issues: #30994
I've implemented an important optimisation in response to GitHub
Copilot's recent rate limit on concurrent Vision API calls. Previously,
our system was defaulting to vision header: true for all API calls. To
prevent unnecessary calls and adhere to the new limits, I've updated our
logic: the vision header is now only sent if the current message is a
vision message, specifically when the preceding message includes an
image.
Prompt used to reproduce and verify the fix: `Give me a context for my
agent crate about. Browse my repo.`
Release Notes:
- copilot: Set Copilot-Vision-Request header based on message content
Closes#5255, #1046, #28322, #15728
This PR makes `AddSelectionBelow` and `AddSelectionAbove` not skip lines
that are shorter than the current cursor column. This follows the same
behavior as VSCode and Sublime.
This change is only applicable in the case of an empty selection; if
there is a non-empty selection, it continues to skip empty and shorter
lines to create a Vim-like column selection, which is the better default
for that case.
- [x] Tests
The empty selection no longer skips shorter lines:
https://github.com/user-attachments/assets/4bde2357-20b6-44f2-a9d9-b595c12d3939
Non-empty selection continues to skip shorter lines.
https://github.com/user-attachments/assets/4cd47c9f-b698-40fc-ad50-f2bf64f5519b
Release Notes:
- Improved `AddSelectionBelow` and `AddSelectionAbove` to no longer skip
shorter lines when the selection is empty, aligning with VSCode and
Sublime behavior.
The panic occurred when querying a second search in the project search
multibuffer while there were dirty buffers.
The panic only happened in Nightly so there's no release notes
Release Notes:
- N/A
1. Add system prompt: this is how it's called from threads. Previously,
we were sending
2. Fix an issue with writing agent thought into a newly created empty
file.
Release Notes:
- N/A
---------
Co-authored-by: Ben Brandt <benjamin.j.brandt@gmail.com>
Co-authored-by: Antonio Scandurra <me@as-cii.com>
Currently, `search::ReplaceNext` works only first time it is executed
because Zed switches the focus to the editor. It seems
`self.editor_focus` call is unnecessary.
Closes#17466
Release Notes:
- Fixed `Replace Next Match` command. Previously it worked once, then
Zed incorrectly switched the focus to the editor
https://github.com/user-attachments/assets/66ef61d6-1efe-43ca-8d8c-6b40540a9930
In accordance with #30327, I saw no reason for included files to get
special treatment, and I actually get use out of prefilling excluded
files because I like not to search symlinked files which, in my
workflow, use a naming convention.
This is simply implementing the same exact changes, but for excluded. It
was tested with `"space /": ["pane::DeploySearch", { "excluded_files":
"**/_*.tf" }]` and works just fine.
Release Notes:
- Added `excluded_files` to `pane::DeploySearch`.
Builds on top of https://github.com/zed-industries/zed/pull/30942
This turns on incremental compilation and decreases extension
compilation times by up to another 41%
Putting us at roughly 92% improved extension load times from what is in
the app today.
Because we only have a static engine, I can't reset the cache between
every run. So technically the benchmarks are always running with a
warmed cache. So the first extension we load will take the 8.8ms, and
then any subsequent extensions will be closer to the measured time in
this benchmark.
This is also measuring the entire load process, not just the
compilation. However, since this is the loading we likely think of when
thinking about extensions, I felt it was likely more helpful to see the
impact on the overall time.
This works because our extensions are largely the same Wasm bytecode
(SDK code + std lib functions etc) with minor changes in the trait impl.
The more different that extensions implementation is, there will be less
benefit, however, there will always be a large part of every extension
that is always the same across extensions, so this should be a speedup
regardless.
I used `moka` to provide a bound to the cache. We could use a bare
`DashMap`, however if there was some issue this could lead to a memory
leak. `moka` has some slight overhead, but makes sure that we don't go
over 32mb while using an LRU-style mechanism for deciding which
compilation artifacts to keep.
I measured our current extensions to take roughly 512kb in the cache.
Which means with a cap of 32mb, we can keep roughly 64 *completely
novel* extensions with no overlap. Since our extensions will have more
overlap than this though, we can actually keep much more in the cache
without having to worry about it.
#### Before:
```
load/1 time: [8.8301 ms 8.8616 ms 8.8931 ms]
change: [-0.1880% +0.3221% +0.8679%] (p = 0.23 > 0.05)
No change in performance detected.
```
#### After:
```
load/1 time: [5.1575 ms 5.1726 ms 5.1876 ms]
change: [-41.894% -41.628% -41.350%] (p = 0.00 < 0.05)
Performance has improved.
```
Release Notes:
- N/A
This PR adds a new `GET /users/look_up` endpoint for retrieving users by
various identifiers.
This endpoint can look up users by the following identifiers:
- Zed user ID
- Stripe Customer ID
- Stripe Subscription ID
- Email address
- GitHub login
Release Notes:
- N/A
When building for the `x86_64-unknown-linux-musl` target, the default
`openssl-dev` is compiled for the GNU toolchain, which causes a build
error due to missing OpenSSL. This PR fixes the issue by avoiding the
use of OpenSSL on non-macOS and non-Windows platforms.
Release Notes:
- N/A
This PR changes it so we only create a snapshot and get the syntax tree
for a buffer if we didn't detect that auto_close is enabled.
<img width="1205" alt="Screenshot 2025-05-16 at 21 10 28"
src="https://github.com/user-attachments/assets/1ada445f-77bc-4c7c-bffe-953f34ee5384"
/>
Release Notes:
- Improved project search performance
https://github.com/zed-industries/zed/issues/30972 brought up another
case where our context is not enough to track the actual source of the
issue: we get a general top-level error without inner error.
The reason for this was `.ok_or_else(|| anyhow!("failed to read HEAD
SHA"))?; ` on the top level.
The PR finally reworks the way we use anyhow to reduce such issues (or
at least make it simpler to bubble them up later in a fix).
On top of that, uses a few more anyhow methods for better readability.
* `.ok_or_else(|| anyhow!("..."))`, `map_err` and other similar error
conversion/option reporting cases are replaced with `context` and
`with_context` calls
* in addition to that, various `anyhow!("failed to do ...")` are
stripped with `.context("Doing ...")` messages instead to remove the
parasitic `failed to` text
* `anyhow::ensure!` is used instead of `if ... { return Err(...); }`
calls
* `anyhow::bail!` is used instead of `return Err(anyhow!(...));`
Release Notes:
- N/A
This PR hides hover info/diagnostic popovers when code action menu is
shown. We already hide hover info/diagnostic popover on code completion
menu trigger (handled on input).
Note: It is still possible to see hover popover if code completion or
code action menu is already open. This is intended behavior.
- [x] Test hover popover hides when code action is triggered
Release Notes:
- Fixed issue where info and diagnostic hover popovers were still
visible when code action menu is triggered.
Some providers sometimes send `{ "type": "text", "text": ... }` instead
of just the text as a string. Now we accept those instead of erroring.
Release Notes:
- N/A
Closes: https://github.com/zed-industries/zed/issues/29535
Broken in: https://github.com/zed-industries/zed/pull/28559/files
Removes `editor::FindNextMatch` and `editor::FindPreviousMatch` from the
default sublime mappings. If you would like to use this, you will have
to add them to your user keymap. Reverts the previous behavior where
cmd-g / cmd-shift-g relies on the base keymap.
Linux:
```json
{
"context": "Editor && mode == full",
"bindings": {
"f3": "editor::FindNextMatch",
"shift-f3": "editor::FindPreviousMatch"
}
}
```
MacOS:
```json
{
"context": "Editor && mode == full",
"bindings": {
"cmd-g": "editor::FindNextMatch",
"cmd-shift-g": "editor::FindPreviousMatch"
}
},
```
Release Notes:
- Fixed a regression in Sublime Text keymap for find next/previous in
the search bar
Closes: #30730
It conflicts with the `>` key on the Czech keyboard layout
If you want the previous behavior, add `"alt-.": ["terminal::SendText",
"\u001b."]` to your keymap under the `Terminal` context.
Release Notes:
- Improved the default terminal keybind to not conflict on Czech
keyboards
Co-authored-by: Peter Tripp <peter@zed.dev>
Global settings were implemented in #30444, but `Settings`
implementations need to consider that source for it to be useful. This
PR does just that for `TelemetrySettings` so these can be controlled via
global settings.
Release Notes:
- N/A
This eval checks that Edit Agent can create an empty file without
writing its thoughts into it. This issue is not specific to empty files,
but it's easier to reproduce with them.
For some mysterious reason, I could easily reproduce this issue roughly
90% of the time in actual Zed. However, once I extract the exact LLM
request before the failure point and generate from that, the
reproduction rate drops to 2%!
Things I've tried to make sure it's not a fluke: disabling prompt
caching, capturing the LLM request via a proxy server, running the
prompt on Claude separately from evals. Every time it was mostly giving
good outcomes, which doesn't match my actual experience in Zed.
At some point I discovered that simply adding one insignificant space or
a newline to the prompt suddenly results in an outcome I tried to
reproduce almost perfectly.
This weirdness happens even outside the Zed code base and even when
using a different subscription. The result is the same: an extra newline
or space changes the model behavior significantly enough, so that the
pass rate drops from 99% to 0-3%
I have no explanation to this.
Release Notes:
- N/A
Reverts zed-industries/zed#31022
Sorry @mikayla-maki, I found that things are more complicated than I
thought.
The lines returned by shape_text must maintain the same length as all
the original characters, otherwise the subsequent offset needs to always
consider the difference of `\r\n` or `\n` to do the offset.
Before, we only needed to add +1 after each offset after the line, but
now we need to consider +1 or +2, which is much more complicated.
This PR makes it so we only create a Zed Free subscription if there is
no other active subscription, rather than just having another Zed Free
subscription.
Release Notes:
- N/A
Co-authored-by: Max Brunsfeld <maxbrunsfeld@gmail.com>
Release Notes:
- N/A
---
Today I check the shape_text result on Windows, I get:
<img width="409" alt="屏幕截图 2025-05-20 222908"
src="https://github.com/user-attachments/assets/3ee93911-3de1-4e01-9433-00c626fc2369"
/>
Here the `shape_text` split logic I think it should use `lines` method,
not `split('\n')`, the newline on Windows is `\r\n`.
Release Notes:
- N/A
---
When we use `window_handle` to draw WebView on Windows, this will crash
by:
This error caused by when used WebView2.
```
thread 'main' panicked at crates\gpui\src\app\async_context.rs:91:28:
already borrowed: BorrowMutError
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace
thread 'main' panicked at library\core\src\panicking.rs:221:5:
panic in a function that cannot unwind
```
Try this https://github.com/tauri-apps/wry/pull/1383 on Windows can
replay the crash.
In fact, we had done [a similar fix around August last
year](https://github.com/huacnlee/zed/pull/6), but we used the unsafe
method to avoid crashes in that version, we felt that it was not a good
change, so we do not make PR.
Today @sunli829 thought about it again and changed the method. Now using
`try_borrow_mut` is similar to the previous `borrow_mut`.
691de6b4b3/crates/gpui/src/app.rs (L70-L78)
I have tested to start Zed by those changes, it is looks no problem.
Co-authored-by: Sunli <scott_s829@163.com>
Extensions cannot look up available port themselves, hence the new API.
With this I'm able to port our Ruby implementation into an extension.
Release Notes:
- N/A
`ImageItem`'s `file` is returning `""` as its `path` for single-filed
worktrees like the ones are created for the images dropped from the OS.
`ImageItem::load_image_metadata` had used that `path` in FS operations
and the other method tried to use for icon resolving.
Rework the code to use a more specific, `worktree::File` instead and
always use the `abs_path` when dealing with paths from this `file`.
Release Notes:
- Fixed images not opening on drag and drop into the editor
