Z-2357
I've found a crate that handles both comments and trailing commas in
JSON. It is a fork of `serde_json`, but it is maintained & up-to-date.
Sadly RawValue seems to not play nicely with it; I've ran into
deserialisation issues around use of RawValue. For this PR I've migrated
to `Value` API.
Obviously this is just a point of discussion, not something I'd merge
straight away. There may be better solutions to this particular problem.
I've also noticed that `serde_json_lenient` does not handle trailing
commas after bindings array. I'm not sure how big of an issue that is.
Release Notes:
- Improved handling of trailing commas in settings files.
[#1322](https://github.com/zed-industries/community/issues/1322)
Rather than assuming a specific family exists, try a set of specific
names and if they fail, just grab any old font that the system reports
as existing
Co-Authored-By: Nathan Sobo <nathan@zed.dev>
This adds basic ability to interact with OpenAI inside a buffer.
Release Notes:
* If`OPENAI_API_KEY` is defined in your environment, you can now run the
`ai: assist` command to pass the text of your current buffer to GPT-4.
If you're editing a file with a `.zmd` extension, you can also invoke
the model with `cmd-enter.` (preview-only)
Drop dependency on tokio introduced by async-openai and do it ourselves.
The approach I'm taking of replacing instead of appending is causing issues. Need to just append.
Previously, this was used because we didn't have access to the current
view and `EventContext` was an element-only abstraction. Now that the
`EventContext` wraps the current view's `ViewContext` we can simply check
for the view's focus and avoid querying ancestors.
During `layout`, we now pass a mutable reference to the element's
parent view. This is a recursive process that causes the view to
be removed from `AppContext` and then re-inserted back into it once
the layout is complete.
As such, querying parent views during `layout` does not work as such
views will have been removed from `AppContext` and not yet re-inserted
into it. This caused a bug in `KeystrokeLabel`, which used the `keystrokes_for_action`
method to query its ancestors to determine their type id and keymap context.
Now, any time a view notifies, we will cache its keymap context so that
we don't need to query the parent view during `layout`.
This is achieved by moving `available_actions` into `AsyncAppContext` (where
we know no view/window is on the stack) and `keystrokes_for_action` into `LayoutContext`
where we'll fetch the previous frame's ancestors and notify the current view if those
change after we perform a layout.
Closes https://github.com/zed-industries/community/issues/1095
by forcing the non-Cancel button to get a focus.
Due to the way macOs handles buttons on modals, the focus gain had to be
achieved via certain button addition order, rather than conventional
"setFocus"-ish API, see the related comment for details.
Co-authored-by: Antonio Scandurra <antonio@zed.dev>
Previously, `debug_elements` was available on `WindowContext`. If that
method was called while having a borrow out to a view, it would panic because
the view would already have been borrowed.
By moving it to an `AsyncAppContext` we ensure the method can't be called while
a view is being used.
This avoids an invitation to hold strong view handles across async await
points, which is a common source of leaks.
Co-Authored-By: Nathan Sobo <nathan@zed.dev>
Previously, the workspace view was on the stack when we were computing the
available actions, which excluded it. This is a stopgap. We should find a
better solution ASAP.
Multi-element are now generic over any drawable child, which can be converted
into an element.
Co-Authored-By: Nathan Sobo <nathan@zed.dev>
Co-Authored-By: Max Brunsfeld <max@zed.dev>
