I was unable to run the collab tests locally because I would run out of
file descriptors.

From some digging it turned out that tokio allocates a new file
descriptor to do work on the CurrentThread using KQUEUE.

We create a new tokio Runtime with each database connection, and these
database connections were being retained by the Client, which is
retained by the Context.

Cleaning up our leaked contexts (and an unrelated retain cycle in the
UserStore) fixes the problem (though does make me
wonder if a different approach might be preferrable).

Release Notes:

- N/A

rust analyzer has a tendency to return markdown of the form:

```rust
 // <-- note the leading space
blah blah blah
```

This is clearly defectuous, so we used to .trim() the output.

Unfortunately we trim after applying syntax highlighting, so that causes
the output to look goofy.

Fix this by updating the highlighting when we trim.

[[PR Description]]

Release Notes:

- Fixed a bug where syntax highlighting was off in hover tooltips

* Remove ability to act on people defined in parent channels
* Show promote buttons on guests

Release Notes:

- Improved channel membership manangement

Before this change once the UserStore had been created, the Client would
never be dropped.

I was unable to run the collab tests locally because I would run out of
file descriptors.

From some digging it turned out that tokio allocates a new file
descriptor to do work on the CurrentThread using KQUEUE.

We create a new tokio Runtime with each database connection, and these
database connections were being retained by the Client, which is
retained by the Context.

Cleaning up our leaked contexts fixes the problem (though does make me
wonder if a different approach might be preferrable).

TODO:
- [x] Clippy pass
- [x] Element / WindowContext refactor
- [x] `geometry `
- [x] `text_system`
- [x] `styled`
- [x] `element_cx`

Release Notes:

- N/A

Make gpui pass clippy

Don't allow granting guests write access in a call where the channel
or one of its ancestors requires the zed CLA, until that guest has
signed the Zed CLA.

Co-authored-by: Marshall <marshall@zed.dev>

Release Notes:

- Fixed an issue where mouse wheel events could be ignored while
selecting text near the top or bottom of the editor with the mouse.

* Do not add fonts' `font_family` into the font completion list, if
there's no corresponding attribute in the font data
* Remove `font_style_name` and `font_name` attribute from the completion
options, since those bring `-Italic`, `-Bold` variants that are not
family names (which we use to load the fonts, so they are useless)

Release Notes:

- Improved font family names completion in the settings

co-authored-by: Piotr <piotr@zed.dev>

rust analyzer has a tendency to return markdown of the form:

```rust
 // <-- note the leading space
blah blah blah
```

This is clearly defectuous, so we used to .trim() the output.

Unfortunately we trim after applying syntax highlighting, so that causes
the output to look goofy.

Fix this by updating the highlighting when we trim.

This will be used by CLA Bot.

Co-authored-by: Marshall <marshall@zed.dev>

* Remove ability to act on people defined in parent channels
* Show promote buttons on guests

Release Notes:

- Fixed a bug where editor hover state could appear while menu open
([#2384](https://github.com/zed-industries/community/issues/2384)).

Co-authored-by: Marshall <marshall@zed.dev>

Co-authored-by: Marshall <marshall@zed.dev>

Release builds are currently failing:
https://github.com/zed-industries/zed/actions/runs/7616903969/job/20744857084

It is unclear to me why debug builds are not.

Release Notes:

- (Added|Fixed|Improved) ...
([#<public_issue_number_if_exists>](https://github.com/zed-industries/community/issues/<public_issue_number_if_exists>)).

In the old world, panel loading happened strictly before workspace
deserialization. Now it's inverted.

Fix this by storing on the dock the serialized state so they can restore
state as panels are loaded.

[[PR Description]]

Release Notes:

- Fixed Zed forgetting which panels were open on boot.
([#2406](https://github.com/zed-industries/community/issues/2406)).

In the old world, panel loading happened strictly before workspace
deserialization. Now it's inverted.

Fix this by storing on the dock the serialized state so they can restore
state as panels are loaded.

Co-authored-by: Mikayla <mikayla@zed.dev>

- it seems like layer.drawable_size() is 0,0 in metal_renderer.rs:189 
- we set this in two places:
    - in response to a CALayerDelegate setFrameSize: event
- in response to a CALayerDelegate viewDidChangeBackingProperties:
event.
- it looks like if we don't set it in either of these cases we get a
different failure mode: the view is zoomed just wrong.
- That said, I can reproduce the screenshot if .scale_factor() returns
0.
- This might happen if [nativeWindow screen] is nil, which happens when
the window is off screen.
    - possible that zed started with offscreen window?
- I would expect that viewDidChangeBackingProperties would fire when
that changed.
    - potential fix: default to 2.0

Release Notes:

- Fixed Zed occasionally rendering blank on laungh
([#2422](https://github.com/zed-industries/community/issues/2422)).

Release Notes:

- Fixes a bug where the bottom of hover popovers would be cut off
([#2434](https://github.com/zed-industries/community/issues/2434)).

Add support for mapping `jk` to escape in vim mode.

This changes the behaviour of the keymatches when there are pending
matches.

Before: Even if there was a pending match, any complete matches would be
triggered and the pending state lost.

After: If there is a pending match, any complete matches are delayed by
1s, or until more keys are typed.

Release Notes:

- Added support for mapping `jk` in vim mode
([#2378](https://github.com/zed-industries/community/issues/2378)),
([#176](https://github.com/zed-industries/community/issues/176))

Part of `was_top_layer` was checking if a opaque level starts with the
current stacking order, but now that each entry in the order has an id,
those comparison checks were failing as every entry generated is now
unique

I'll admit I'm still a little fuzzy on the specific reasons for this
function to be defined as it is so I'd appreciate another set of eyes

Release Notes:

- N/A