Cherry-picked gpui: Fix slow scrolling in lists (#33608)

matches editor element's behavior



https://github.com/user-attachments/assets/f70912e1-5adb-403b-a98c-63e2e89929ac


- in first version editor scrolls like 1.5 pages, but agent panel only
scrolls half a page.
- in second version, agent panel also scrolls like 1.5 pages.

Release Notes:

- Fixed skipping of some scroll events in the non-uniform list UI
element, which fixes slow scrolling of the agent panel.

Co-authored-by: maan2003 <49202620+maan2003@users.noreply.github.com>

Cherry-picked Revert "languages: Bump ESLint LSP server to version
3.0.10 (#32717)" (#33659)

This reverts commit 1edaeebae5b0f1b67cdad1a994b88a3c11f75705.

Based on an elevated number of ESLint-related issues, reverting the
upgrade.
Many people upvoted the issues and did not share any repro details, so
cannot be certain what's more broken: seems relatively generic as
related to *.ts ESLint configs.

Checked the revert on 2 projects from the issues below:

Closes https://github.com/zed-industries/zed/issues/33425

With https://github.com/adamhl8/zed-33425 as an example repo: there,
both eslint configurations worked for me when I stopped Zed and opened a
project.
Somehow, switching various Zed's with different vscode-eslint package
versions, eventually I get
`Error: Cannot find module

'~/.local/share/zed/languages/eslint/vscode-eslint-3.0.10/vscode-eslint/server/out/eslintServer.js'`-ish
error.

Not very related to issues with newer vscode-eslint integration, but
worth mentioning as is related to the package updates.


Closes https://github.com/zed-industries/zed/issues/33648

With a good example of
https://github.com/florian-lackner365/zed-eslint-bug monorepo project.
The monorepo part seems not to be related, but somehow,
`eslint.config.js` is involved as the newer vscode-eslint fails to find
a config.
Works well with the older vscode-eslint.

Release Notes:

- Downgraded to vscode-eslint-2.4.4 as a ESLint language server

Co-authored-by: Kirill Bulatov <kirill@zed.dev>

Follow-up of https://github.com/zed-industries/zed/pull/33605

Release Notes:

- N/A

Use a div-based rendering code instead of using a text

Closes https://github.com/zed-industries/zed/discussions/33507

Before:
<img width="410" alt="before_dark"
src="https://github.com/user-attachments/assets/66ad63ae-7836-4dc7-8176-a2ff5a38bcd4"
/>
After:
<img width="407" alt="after_dark"
src="https://github.com/user-attachments/assets/0b627da8-461b-4f19-b236-4a69bf5952a0"
/>


Before:
<img width="409" alt="before_light"
src="https://github.com/user-attachments/assets/ebcfabec-fcda-4b63-aee6-c702888f0db4"
/>
After:
<img width="410" alt="after_light"
src="https://github.com/user-attachments/assets/c0da42a1-d6b3-4e08-a56c-9966c07e442d"
/>

The border is not that contrast as in VSCode examples in the issue, but
I'm supposed to use the right thing in

https://github.com/zed-industries/zed/blob/1e11de48eeba01dc12761ec1274f8b0963de5514/crates/editor/src/display_map/inlay_map.rs#L357

based on 


https://github.com/zed-industries/zed/blob/41583fb066629d1e54d600e930be068a68984c5c/crates/theme/src/styles/colors.rs#L16-L17

Another oddity is that the border starts to shrink on `cmd-=`
(`zed::IncreaseBufferFontSize`):

<img width="1244" alt="image"
src="https://github.com/user-attachments/assets/f424edc0-ca0c-4b02-96d4-6da7bf70449a"
/>

but that needs a different part of code to be adjusted hence skipped.

Tailwind CSS example:

<img width="1108" alt="image"
src="https://github.com/user-attachments/assets/10ada4dc-ea8c-46d3-b285-d895bbd6a619"
/>


Release Notes:

- Reworked color indicators visual representation

Closes https://github.com/zed-industries/zed/issues/33575

* Fixes inlay colors spoiled after document color displayed
* Optimizes the query pattern for large multi buffers

Release Notes:

- Fixed document colors issues with other inlays and multi buffers

Cherry-picked Don't panic on vintage files (#33543)

Release Notes:

- remoting: Fix a crash on the remote side when encountering files from
before 1970.

Co-authored-by: Conrad Irwin <conrad.irwin@gmail.com>

Closes https://github.com/zed-industries/zed/issues/33522

Turns out a bunch of Zed requests were not checking their capabilities
correctly, due to odd copy-paste and due to default that assumed that
the capabilities are met.

Adjust the code, which includes the document colors, add the test on the
colors case.

Release Notes:

- Fixed excessive document colors requests for unrelated files

Cherry-picked Fix blend alpha colors with editor background in inline
preview (#33513)

Closes #33505

## Before

<img width="434" alt="Screenshot 2025-06-27 at 12 22 57"

src="https://github.com/user-attachments/assets/ac215a39-b3fe-4c9e-bd7d-0d7568d5fd1f"
/>

## After

<img width="441" alt="Screenshot 2025-06-27 at 12 22 47"

src="https://github.com/user-attachments/assets/28218ed6-c1aa-4d3f-a268-def2fa9f0340"
/>

Release Notes:

- Fixed inline color previews not correctly blending alpha/transparency
values with the editor background

Co-authored-by: ddoemonn <109994179+ddoemonn@users.noreply.github.com>

Cherry-picked debugger: Fix treatment of node-terminal scenarios
(#33432)

- Normalize `node-terminal` to `pwa-node` before sending to DAP
- Split `command` into `program` and `args`
- Run in external console

Release Notes:

- debugger: Fixed debugging JavaScript tasks that used `"type":
"node-terminal"`.

Co-authored-by: Cole Miller <cole@zed.dev>

## Summary

This PR improves the workaround introduced in #33335 that handles cases
where the pinned tab count exceeds the actual tab count during workspace
deserialization.

## Problem

The original workaround in #33335 successfully prevented the panic but
had two issues:
1. **Console spam**: The warning message was logged repeatedly because
`self.pinned_tab_count` wasn't updated to match the actual tab count
2. **Auto-pinning behavior**: New tabs up until you exceed the old safe
tab count were automatically pinned after the workaround was triggered.

## Solution

Updates the defensive code to set `self.pinned_tab_count = tab_count`
when the mismatch is detected, ensuring:
- The warning is only logged once when encountered.
- New tabs behave normally (aren't auto-pinned)
- The workspace remains in a consistent state

This is an immediate fix for the workaround. I'll attempt to open up a
follow-up PR when i get the chance that will address the root cause by
implementing serialization for empty untitled tabs, as discussed in
#33342.

Release Notes:

- N/A

Fixes a regression introduced in
https://github.com/zed-industries/zed/pull/33289

Release Notes:

- Fixed a bug where some text thread actions were accidentally removed.

Closes #32959

Release Notes:

- Fixed the issue where accepting variable completion in the Debugger
would append the entire variable name instead of the remaining part.

Closes #33210 

Release Notes:

- Fixed an issue where a trailing single quote was included when opening
a link from the terminal.

Closes https://github.com/zed-industries/zed/issues/33221

Bedrock has similar to anthropic caching api, if we want to cache
messages up to a certain point, we should add a special block into that
message.

Additionally, we can cache tools definition by adding cache point block
after tools spec.

See: [Bedrock User Guide: Prompt
Caching](https://docs.aws.amazon.com/bedrock/latest/userguide/prompt-caching.html#prompt-caching-models)

Release Notes:

- bedrock: Added prompt caching support

---------

Co-authored-by: Oleksiy Syvokon <oleksiy@zed.dev>

Follow up to previous PRs:
- Return `true` in `supports_images` - v0 supports images already
- Rename model id to match the exact version of the model `v0-1.5-md`
(For now we do not expose `sm`/`lg` variants since they seem not to be
available via the API)
- Provide autocompletion in settings for using `vercel` as a `provider`

Release Notes:

- N/A

Follow up to #33292

Since Vercel's API is OpenAI compatible, we can reuse a bunch of code.

Release Notes:

- N/A

## Description of Feature or Change

Zed currently lacks a built-in way to convert a file’s indentation style
on the fly. While it's possible to change indentation behavior via
global or language-specific settings, these changes are persistent and
broad in scope as they apply to all files or all files of a given
language. We believe this could be improved for quick one-off
adjustments to specific files.

This PR introduces two new editor commands:
`Editor::convert_indentation_to_spaces` and
`Editor::convert_indentation_to_tabs`. These commands allow users to
convert the indentation of either the entire buffer or a selection of
lines, to spaces or tabs. Indentation levels are preserved, and any
mixed whitespace lines are properly normalized.

This feature is inspired by VS Code’s "Convert Indentation to
Tabs/Spaces" commands, but offers faster execution and supports
selection-based conversion, making it more flexible for quick formatting
changes.

## Implementation Details

To enable selection-based indentation conversion, we initially
considered reusing the existing `Editor::manipulate_lines` function,
which handles selections for line-based manipulations. However, this
method was designed specifically for operations like sorting or
reversing lines, and does not allow modifications to the line contents
themselves.

To address this limitation, we refactored the method into a more
flexible version: `Editor::manipulate_generic_lines`. This new method
passes a reference to the selected text directly into a callback, giving
the callback full control over how to process and construct the
resulting lines. The callback returns a `String` containing the modified
text, as well as the number of lines before and after the
transformation. These counts are computed using `.len()` on the line
vectors during manipulation, which is more efficient than calculating
them after the fact.


```rust
fn manipulate_generic_lines<M>(
  &mut self,
  window: &mut Window,
  cx: &mut Context<Self>,
  mut manipulate: M,
) where
   M: FnMut(&str) -> (String, usize, usize),
 {
   // ... Get text from buffer.text_for_range() ...
   let (new_text, lines_before, lines_after) = manipulate(&text);
   // ...
``` 

We now introduce two specialized methods:
`Editor::manipulate_mutable_lines` and
`Editor::manipulate_immutable_lines`. Each editor command selects the
appropriate method based on whether it needs to modify line contents or
simply reorder them. This distinction is important for performance: when
line contents remain unchanged, working with an immutable reference as
`&mut Vec<&str>` is both faster and more memory-efficient than using an
owned `&mut Vec<String>`.

## Demonstration


https://github.com/user-attachments/assets/e50b37ea-a128-4c2a-b252-46c3c4530d97



Release Notes:

- Added `editor::ConvertIndentationToSpaces` and
`editor::ConvertIndentationToTabs` actions to change editor indents

---------

Co-authored-by: Pedro Silveira <pedroruanosilveira@tecnico.ulisboa.pt>

Closes #32341

Release Notes:

- Fixed DeepSeek R1 errors for reasoning blocks being sent back to the model.

Bedrock [do not accept][1] `null` as a JSON value input for the tool
call when called back.

Instead of passing null, we will pass back an empty object, which is
accepted by API

Closes #33204

Release Notes:

- Fixed project diagnostic tool call for bedrock

[1]:
https://docs.aws.amazon.com/bedrock/latest/APIReference/API_runtime_ToolUseBlock.html

Release Notes:

- agent: Fix an issue where MCP servers that were provided by extensions
would sometimes not start up

Closes #ISSUE

Release Notes:

- N/A

Closes #30714

Bedrock converse api expect to see tool options if at least one tool was
used in conversation in the past messages.

Right now if `LanguageModelToolChoice::None` isn't supported edit agent
[remove][1] tools from request. That point breaks Converse API of
Bedrock. As was proposed in [the issue][2] we won't drop tool choose but
instead will deny any of them if model will respond with a tool choose.

[1]:
https://github.com/x-qdo/zed/blob/fceba6c79540677c2504d2c22191963b6170591a/crates/assistant_tools/src/edit_agent.rs#L703
[2]:
https://github.com/zed-industries/zed/issues/30714#issuecomment-2886422716

Release Notes:

- Fixed bedrock tool calls in edit mode

This reduces code complexity and avoids unnecessary roundtripping
through `DisplayPoint`. Hopefully this doesn't cause behavior changes,
but has one known behavior improvement:

`clip_at_line_ends` logic caused `is_inside_word` to return false when
on a word at the end of the line. In vim mode, this caused
`select_all_matches` to not select words at the end of lines, and in
some cases crashes due to not finding any selections.

Closes #29823

Release Notes:

- N/A

- Closes: #30962 

Nushell does not support mkdir -p
So invoke sh -c "mkdir -p" instead which will also work under nushell.

Release Notes:

- Fixed ssh remotes running Nushell (and possibly other non
posix-compliant shells)

---------

Co-authored-by: Conrad Irwin <conrad.irwin@gmail.com>

This helps correlate sessions in the picker with entries in the debug
adapter logs view.

Release Notes:

- N/A

Release Notes:

- Debugger: Add support for completion triggers in debug console

Release Notes:

- N/A

Closes #33061


https://github.com/user-attachments/assets/3b3e146e-7c12-412e-b4dd-c70411891b9e

Release Notes:

- Fixed cursor unexpectedly moving when entering/exiting insert mode in
Helix mode, making the behavior consistent with the Helix editor.

Rounding broke (among other things, probably) pixel-perfect image
rendering with non-power-of-two scaling factor.

An example which reproduces the problem can be found
[here](https://github.com/WaffleLapkin/gpui_taffy_rounding_whyyyyy).

How it looks with `gpui` from `main`:
![2025-05-31
11:34:25+CEST](https://github.com/user-attachments/assets/2cb19312-6ba6-4e80-8072-f89ddedff77b)

How it looks with this patch:
![2025-05-31
11:35:28+CEST](https://github.com/user-attachments/assets/114b52a9-58c0-4600-871c-a20eceb7179e)

Both screenshots are made on kde+wayland with magnification using kde's
built-in magnification (`Meta`+`+`, `Meta`+`-`). Note that screenshot
apps have a high chance of lying 🙃

The image itself is 400 by 300 pixels of red/green checkerboard pattern
made specifically to exaggerate scaling issues.

Release Notes:

- N/A

After much investigation, I have not been able to track down what is
causing [this
panic](https://github.com/zed-industries/zed/issues/33342). I'm clamping
the value for now, because a bug is better than a crash. Hopefully
someone finds reproduction steps, and I will implement a proper fix.

Release Notes:

- N/A

This is an easy way to shave some microseconds off the critical path for
frame rendering. On my machine this reduces typical frame rendering
latency by ~100 microseconds, probably quite a bit more on slower
machines.

Here is how long it typically takes to drop elements from the arena,
from a fairly brief run:


![image](https://github.com/user-attachments/assets/65cfd911-eccf-4393-887d-8cad2cd27148)

Release Notes:

- N/A

Release Notes:

- N/A

Closes #33303

Release Notes:

- debugger: Swap step in/out icon positions in debug panel to be
consistent with other editors

## Context

To support inline values a language will have to implement their own
provider trait that walks through tree sitter nodes. This is overly
complicated, hard to accurately implement for each language, and lacks
proper extension support.

This PR switches to a singular inline provider that uses a language's
`debugger.scm` query field to capture variables and scopes. The inline
provider is able to use this information to generate inlays that take
scope into account and work with any language that defines a debugger
query file.

### Todos
- [x] Implement a utility test function to easily test inline values
- [x] Generate inline values based on captures
- [x] Reimplement Python, Rust, and Go support
- [x] Take scope into account when iterating through variable captures
- [x] Add tests for Go inline values
- [x] Remove old inline provider code and trait implementations

Release Notes:

- debugger: Generate inline values based on a language debugger.scm file

Closes: https://github.com/zed-industries/zed/issues/33256

Move some Editor keymap entries into `Editor && mode == full`

Release Notes:

- N/A

In helix the `f`, `F`, `t`, `T`, left and right motions wrap lines. I
added that by default.

Release Notes:

- vim: The `use_multiline_find` setting is replaced by binding to the
correct action in the keymap:
    ```
"f": ["vim::PushFindForward", { "before": false, "multiline": true }],
"t": ["vim::PushFindForward", { "before": true, "multiline": true }],
"shift-f": ["vim::PushFindBackward", { "after": false, "multiline": true
}],
"shift-t": ["vim::PushFindBackward", { "after": true, "multiline": true
}],
    ```
- helix: `f`/`t`/`shift-f`/`shift-t`/`h`/`l`/`left`/`right` are now
multiline by default (like helix)

cc @osyvokon 

We were seeing a bunch of errors in our backend when people were using
Claude models with thinking enabled.

In the logs we would see
> an error occurred while interacting with the Anthropic API:
invalid_request_error: messages.x.content.0.type: Expected `thinking` or
`redacted_thinking`, but found `text`. When `thinking` is enabled, a
final `assistant` message must start with a thinking block (preceeding
the lastmost set of `tool_use` and `tool_result` blocks). We recommend
you include thinking blocks from previous turns. To avoid this
requirement, disable `thinking`. Please consult our documentation at
https://docs.anthropic.com/en/docs/build-with-claude/extended-thinking

However, this issue did not occur frequently and was not easily
reproducible. Turns out it was triggered by us not correctly handling
[Redacted Thinking
Blocks](https://docs.anthropic.com/en/docs/build-with-claude/extended-thinking#thinking-redaction).

I could constantly reproduce this issue by including this magic string:
`ANTHROPIC_MAGIC_STRING_TRIGGER_REDACTED_THINKING_46C9A13E193C177646C7398A98432ECCCE4C1253D5E2D82641AC0E52CC2876CB
` in the request, which forces `claude-3-7-sonnet` to emit redacted
thinking blocks (confusingly the magic string does not seem to be
working for `claude-sonnet-4`). As soon as we hit a tool call Anthropic
would return an error.

Thanks to @osyvokon for pointing me in the right direction 😄!


Release Notes:

- agent: Fixed an issue where Anthropic models would sometimes return an
error when thinking was enabled

Before, `space` was always causing a channel join.
Now it's less fluent, one has to press `ESC` to get the focus out of the
filter editor and then `space` starts joining the channel.

Release Notes:

- Fixed being unable to input a whitespace character in collab channels
filter

Closes #ISSUE

Adds a very simple API to track metadata about keybindings in GPUI,
namely the source of the binding. The motivation for this is displaying
the source of keybindings in the [keymap
UI](https://github.com/zed-industries/zed/pull/32436).

The API is designed to be as simple and flexible as possible, storing
only a `Option<u32>` on the bindings themselves to keep the struct
small. It is intended to be used as an index or key into a table/map
created and managed by the consumer of the API to map from indices to
arbitrary meta-data. I.e. the consumer is responsible for both
generating these indices and giving them meaning.

The current usage in Zed is stateless, just a mapping between constants
and User, Default, Base, and Vim keymap sources, however, this can be
extended in the future to also track _which_ base keymap is being used.

Release Notes:

- N/A *or* Added/Fixed/Improved ...

Summary

I've successfully implemented the required billing address collection
feature for Stripe Checkout sessions. Here's what was done:

### 1. **Added New Data Structures** (`stripe_client.rs`):
- Added `StripeBillingAddressCollection` enum with `Auto` and `Required`
variants
- Added `billing_address_collection` field to
`StripeCreateCheckoutSessionParams`

### 2. **Updated Stripe Client Implementation**
(`real_stripe_client.rs`):
- Added conversion from `StripeBillingAddressCollection` to Stripe's
`CheckoutSessionBillingAddressCollection`
- Updated the `TryFrom` implementation to map the billing address
collection field when creating checkout sessions
- Added the necessary import

### 3. **Updated Billing Service** (`stripe_billing.rs`):
- Set `billing_address_collection` to `Required` in both
`checkout_with_zed_pro()` and `checkout_with_zed_pro_trial()` methods
- Added the necessary import

### 4. **Updated Test Infrastructure** (`fake_stripe_client.rs`):
- Added `billing_address_collection` field to
`StripeCreateCheckoutSessionCall`
- Updated the `create_checkout_session` implementation to capture the
new field
- Added the necessary import

### 5. **Updated Tests** (`stripe_billing_tests.rs`):
- Added assertions to verify that `billing_address_collection` is set to
`Required` in all three test cases:
  - `test_checkout_with_zed_pro`
  - `test_checkout_with_zed_pro_trial` (regular trial)
  - `test_checkout_with_zed_pro_trial` (extended trial)
- Added the necessary import

The implementation follows the pattern established in the codebase and
ensures that whenever a Stripe Checkout session is created for Zed Pro
subscriptions (both regular and trial), the billing address will be
required from customers. This aligns with the Stripe documentation you
provided, which shows that setting `billing_address_collection=required`
will ensure the billing address is always collected during checkout.

Release Notes:

- N/A

Co-authored-by: Marshall Bowers <git@maxdeviant.com>

Vercel v0 is an OpenAI-compatible model, so this is mostly a dupe of the
OpenAI provider files with some adaptations for v0, including going
ahead and using the custom endpoint for the API URL field.

Release Notes:

- Added support for Vercel as a language model provider.

Add `can_save` and `save` methods to `DiffView`, enabling users to save
changes made within the diff view.

Release Notes:

- Allow saving changes in the `zed --diff` view

Release Notes:

- Made the horizontal outputs scrollable in REPL

This breaks a transitive dependency of `agent` on UI crates. I've also
found and eliminated some dead code in assistant_context_editor.

Release Notes:

- N/A