Closes #ISSUE
The model often uses the wrong closing tag, or has spaces around the
closing tag name. This PR makes it so that opening tags are treated as
authoritative and any closing tag with the name `new_text` `old_text` or
`edits` is accepted based on depth. This has the additional benefit that
the parsing is more robust with contents that contain `new_text`
`old_text` or `edits. I.e. the following test passes
```rust
#[test]
fn test_extract_xml_edits_with_conflicting_content() {
let input = indoc! {r#"
<edits path="component.tsx">
<old_text>
<new_text></new_text>
</old_text>
<new_text>
<old_text></old_text>
</new_text>
</edits>
"#};
let result = extract_xml_replacements(input).unwrap();
assert_eq!(result.file_path, "component.tsx");
assert_eq!(result.replacements.len(), 1);
assert_eq!(result.replacements[0].0, "<new_text></new_text>");
assert_eq!(result.replacements[0].1, "<old_text></old_text>");
}
```
Release Notes:
- N/A *or* Added/Fixed/Improved ...
### What this solves
This PR adds support for HTTP and SSE (Server-Sent Events) transports to
Zed's context server implementation, enabling communication with remote
MCP servers. Currently, Zed only supports local MCP servers via stdio
transport. This limitation prevents users from:
- Connecting to cloud-hosted MCP servers
- Using MCP servers running in containers or on remote machines
- Leveraging MCP servers that are designed to work over HTTP/SSE
### Why it's important
The MCP (Model Context Protocol) specification includes HTTP/SSE as
standard transport options, and many MCP server implementations are
being built with these transports in mind. Without this support, Zed
users are limited to a subset of the MCP ecosystem. This is particularly
important for:
- Enterprise users who need to connect to centralized MCP services
- Developers working with MCP servers that require network isolation
- Users wanting to leverage cloud-based context providers (e.g.,
knowledge bases, API integrations)
### Implementation approach
The implementation follows Zed's existing architectural patterns:
- **Transports**: Added `HttpTransport` and `SseTransport` to the
`context_server` crate, built on top of the existing `http_client` crate
- **Async handling**: Uses `gpui::spawn` for network operations instead
of introducing a new Tokio runtime
- **Settings**: Extended `ContextServerSettings` enum with a `Remote`
variant to support URL-based configuration
- **UI**: Updated the agent configuration UI with an "Add Remote Server"
option and dedicated modal for remote server management
### Changes included
- [x] HTTP transport implementation with request/response handling
- [x] SSE transport for server-sent events streaming
- [x] `build_transport` function to construct appropriate transport
based on URL scheme
- [x] Settings system updates to support remote server configuration
- [x] UI updates for adding/editing remote servers
- [x] Unit tests using `FakeHttpClient` for both transports
- [x] Integration tests (WIP)
- [x] Documentation updates (WIP)
### Testing
- Unit tests for both `HttpTransport` and `SseTransport` using mocked
HTTP client
- Manual testing with example MCP servers over HTTP/SSE
- Settings validation and UI interaction testing
### Screenshots/Recordings
[TODO: Add screenshots of the new "Add Remote Server" UI and
configuration modal]
### Example configuration
Users can now configure remote MCP servers in their `settings.json`:
```json
{
"context_servers": {
"my-remote-server": {
"enabled": true,
"url": "http://localhost:3000/mcp"
}
}
}
```
### AI assistance disclosure
I used AI to help with:
- Understanding the MCP protocol specification and how HTTP/SSE
transports should work
- Reviewing Zed's existing patterns for async operations and suggesting
consistent approaches
- Generating boilerplate for test cases
- Debugging SSE streaming issues
All code has been manually reviewed, tested, and adapted to fit Zed's
architecture. The core logic, architectural decisions, and integration
with Zed's systems were done with human understanding of the codebase.
AI was primarily used as a reference tool and for getting unstuck on
specific technical issues.
Release notes:
* You can now configure MCP Servers that connect over HTTP in your
settings file. These are not yet available in the extensions API.
```
{
"context_servers": {
"my-remote-server": {
"enabled": true,
"url": "http://localhost:3000/mcp"
}
}
}
```
---------
Co-authored-by: Conrad Irwin <conrad.irwin@gmail.com>
Fixes#42945
## Problem
When opening a single file via command line (e.g., `zed
~/Downloads/file.txt`), the terminal panel was opening in the root
directory (/) instead of the file's directory.
## Root Cause
The code only checked for active project directory, which returns None
when a single file is opened. Additionally, file worktrees weren't
handling parent directory lookup.
## Solution
Added fallback logic to use the first project directory when there's no
active entry, and made file worktrees return their parent directory
instead of None.
## Testing
- All existing tests pass
- Added test coverage for file worktree scenarios
- Manually tested with `zed ~/Downloads/file.txt` - terminal now opens
in correct directory
This improves the user experience for users who frequently open single
files from the command line.
## Release Notes
- Fixed terminal opening in root directory when editing single files
from the command line
The idea is that we only store running (`!self.finished`) or finished
(`self.finished`) pending ops, while everything else (skipped, errored)
jobs are pruned out immediately. We don't really need them in the grand
scheme of things anyway.
Release Notes:
- N/A
Closes#42922
Release Notes:
- Fixed Correctly highlighting the 'for' keyword in Rust as
keyword.control only in for loops.
- Fixed Highlighting the 'await' keyword in Rust as keyword.control
This PR's goal is to improve discoverability of how Zed "remembers" the
currently selected agent when hitting `cmd-n` (or `ctrl-n`). Hitting
that binding starts a new thread with whatever agent is currently
selected.
In the example below, I am in a Claude Code thread and if I hit `cmd-n`,
a new, fresh CC thread will be started:
<img width="500" height="822" alt="Screenshot 2025-11-18 at 1 13@2x"
src="https://github.com/user-attachments/assets/d3acd1aa-459d-4078-9b62-bbac3b8c1600"
/>
Release Notes:
- agent: Improved discoverability of the `cmd-n` keybinding to create a
new thread with the currently selected agent.
This should make it easier to correlate Sentry reports with user reports
and
github issues (for users who have diagnostics enabled)
Release Notes:
- N/A
As discussed in the first responders meeting. We have collected a lot of
backtraces from these, but it's not quite clear yet what causes this.
Removing these should ideally make things a bit more stable even if we
may run into panics later one when the faulty anchor is used still.
Release Notes:
- N/A *or* Added/Fixed/Improved ...
Closes#40906Closes#39729
SFTP uploads weren't quoting the install directory which was causing
extension syncing to fail. We were also only running `install_extension`
once per remote-connection instead of once per project (thx @feeiyu for
pointing this out) so extension weren't being loaded in subsequently
opened remote projects.
Release Notes:
- N/A
---------
Co-authored-by: Conrad Irwin <conrad.irwin@gmail.com>
This fixes various issues where rustfmt failed to format code due to too
long strings, most of which I stumbled across over the last week and
some additonal ones I searched for whilst fixing the others.
Release Notes:
- N/A
The `Vim.exit_temporary_normal` method had been updated
(https://github.com/zed-industries/zed/pull/42742) to expect and
`Option<&Motion>` that would then be used to determine whether to move
the cursor right in case the motion was `Some(EndOfLine { ..})`.
Unfortunately this meant that all callers now had to provide this
argument, even if just `None`.
After merging those changes I remember that we could probably play
around with `clip_at_line_ends` so this commit removes those intial
changes in favor of updating the `vim::normal::Vim.move_cursor` method
so that, if vim is in temporary mode and `EndOfLine` is used, it
disables clipping at line ends so that the newline character can be
selected.
Closes [#42278](https://github.com/zed-industries/zed/issues/42278)
Release Notes:
- N/A
Closes #ISSUE
Makes it so that a file named `bucketed_analysis.md` is written to the
runs directory after an eval is ran with > 1 repetitions. This file
buckets the predictions made by the model by comparing the edits made so
that seeing how many times different failure modes were encountered
becomes much easier.
Release Notes:
- N/A *or* Added/Fixed/Improved ...
Returning the favor from @rgbkrk in
https://github.com/posit-dev/air/pull/445
I noticed the R docs around Air are a bit incorrect / out of date. I'll
make a few more comments inline. Feel free to take over for any other
edits.
Release Notes:
- Improved R language support documentation
This prompt is for a fine-tuned model. It has the following changes,
compared to `minimal`:
- No instructions at all, except for one sentence at the beginning of
the prompt.
- Output is a simplified unified diff -- hunk headers have no line
counts (e.g., `@@ -20 +20 @@`)
- Qwen's FIM tokens are used where possible (`<|file_sep|>`,
`<|fim_prefix|>`, `<|fim_suffix|>`, etc.)
To evaluate this model:
```
ZED_ZETA2_MODEL=zeta2-exp [usual zeta-cli eval params ...] --prompt-format minimal-qwen
```
This will point to the most recent Baseten deployment of zeta2-exp
(which may change in the future, so the prompt-format may get out of
sync).
Release Notes:
- N/A
Closes https://github.com/zed-industries/zed/issues/32701
Release Notes:
- agent: Fixed a bug where hitting the `NewThread` keybinding when
focused inside a terminal within the agent panel would create a new
terminal tab instead of a new thread.
