When the `List` element's state is `ListState::reset()`, it eagerly
trashes it's cached element heights in anticipation of a prompt render.
But, due to the recent `display_layer` changes, that re-render is not
always forthcoming. This is a problem for `ListState::scroll()`, which
depends on these cached elements to correctly calculate the new logical
scroll offset.
Solutions we attempted:
- Cache the element heights and continue the scroll calculation
- This was conceptually incorrect, reset should only be called when the
underlying data has been changed, making any calculation with the old
results meaningless.
- Lazily re-compute the element heights in scroll
- Beyond being a non-trivial refactor, this would probably also cause us
to double-render the list in a single frame, which is bad.
- Cache the scroll offset and only calculate it in paint
- This solution felt awkward to implement and meant we can't supply
synchronous list scroll events.
- Delay resetting until paint
- This means that all of the other APIs that `ListState` supplies would
give temporarily incorrect results, worsening the problem
Given these issues, we settled on the solution with the least
compromises: drop scroll events if the state has been `reset()` between
`paint()` and `scroll()`. This shifts the responsibility for the problem
out of the List element and into consumers of `List`, if you want
perfectly smooth scrolling then you need to use `reset()` judiciously
and prefer `splice()`.
That said, I tested this by aggressively scrolling the Collab panel, and
it seems to work as well as it did before.
This PR also includes some initial testing infrastructure for working
with input from the platform and rendered elements.
Release Notes:
- N/A
I essentially went through the publicly exported items and marked these
that are e.g. leaky reexports as pub(crate). I expect that'd be done on
Tuesday anyways.
Release Notes:
- N/A
This removes one of the path dependencies in gpui that's only really
needed by `workspace` (which can work around lack of these
implementations by itself). In theory it should also improve build
scheduling (as gpui doesn't have to wait for main dependency of sqlez -
libsqlite3 - to finish it's 25 seconds-long build in release), though in
practice I didn't notice a substantial improvement.
Moreover `sqlez` was unused by `settings` too, so that's removed
as well.
Release Notes:
- N/A
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>
When the help menu is open, the help menu's search field is the key window, and this was causing menu item action resolution to fail
co-authored-by: Max <max@zed.dev>
This is done by supplying the screen in the `WindowOptions` struct.
Note that it's optional, and we will let the operating system choose
which screen to show the window on when `screen` is not provided, as
we did before this change.
Mikayla
Piotr Osiewicz
Antonio Scandurra
Max BrunsfeldandMikayla
JuliaandNathan Sobo
Mikayla MakiandMax
Petros AmoiridisandAntonio Scandurra
Kay Simmons
