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
Previously, we were using `size_of` but passing the wrong type in
(MonochromeSprite instead of PathSprite). This caused us to read outside
of the `sprites` smallvec and triggered the segfault.
We're occasionally seeing a crash in MetalRenderer::draw.
Looking at the backtrace, it seems almost certainly to be happening in
the call to `ptr::copy_nonoverlapping` on line 604 (see `#Don't Panic!`
channel notes)
As we already have added bounds checking to the destination, it seems
most
likely (however improbable) that somehow we're getting an invalid Ptr
and
length from the SmallVec.
To try and make progress on this, let's try a Vec for a bit lest there
is a subtle issue in SmallVec (though I couldn't spot one).
Release Notes:
- (maybe) Fixes SEGFAULT in MetalRenderer::draw
According to Chromium source, `NSScreen::screens` should always get us
one display.
We made this change because we ran into panics caused by the previous
`unwrap()` when `CGGetActiveDisplayList` might return an empty list.
Previously we called CGGetActiveDisplayList twice: once to get the
number of displays and then to get the displays.
We saw a panic due to no displays being returned here. As a first
attempt to fix the panic, we're reducing the amount of calls to
CGGetActiveDisplayList and just do one with the trade-off being that we
pre-allocate 32 pointers in a Vec.
During layout of EditorElement we use 'm' character from current font to calculate sizes, panicking with fonts that do not have that character (e.g. Arabic fonts). It's not really EditorElement's fault, as it assumes that the font it's dealing with is gonna have that character available.
To prevent a crash, I added validation while loading a family that a given font contains the glyphs we're gonna use down the line.
This commit mostly fixes invalid URLs in docstrings. It also
encapsulates crates we reexport (serde stuff + linkme) into a public
module named "private" in order to reduce the API surfaced through docs.
Moreover, I fixed up a bunch of crates that were pulling serde_json in
through gpui explicitly instead of using Cargo manifest.
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
* Fix an incorrect use of `Any::type_id` that prevented the disabling of
key bindings
* Restructured the representation of disabled key bindings so that they
handle context predicates correctly. Previously, to disable key binding,
you needed to supply the exact same context predicate (e.g. `Editor &&
mode == "full"`) as the binding that you are trying to disable. Now, the
context predicates of disabled key bindings are evaluated just like any
other context predicate (with the current context) to see if they apply.
Release Notes:
- Fixed an issue where disabling key bindings didn't work. To disable a
key binding, set it to `null` in your keymap.
Adding guest roles led us down a rabbit hole where we'd have liked to
rely on a
side-effect of activating a window in tests; but the test window didn't
implement that.
Looking into that, I realized our TestWindow wasn't doing a great job of
emulating the MacWindow, so this makes the two more similar.
After implementing it a while ago, our previous interpolation scheme
didn't really make sense to me and was causing borders to be rendered
incorrectly. We don't really draw backgrounds and borders as part of
the same draw call anymore, but it seemed reasonable to have a correct
implementation in the shader anyway.
This commit uses Porter-Duff compositing (i.e., `over`) to produce
a color that is the result of superimposing the border on top of the
background. Then, we linearly interpolate towards the background color
as we slide out of the border and into the background.
Piotr Osiewicz
Mikayla Maki
Antonio Scandurra
Conrad Irwin
Thorsten Ball
Kirill Bulatov
Marshall Bowers
Max Brunsfeld
Joseph T. Lyons