Diggory Hardy

My 2c (just as someone who uses winit) is that winit not supporting theme mode detection on Linux is fine, provided that it's clearly documented which platforms the theme mode...

In terms of characters, I would advocate that the following should be allowed: - Everything under the "Arrows" blocks, including supplements (see https://www.unicode.org/charts/) - Everything under the "Mathematical Operators" blocks...

For your info, I've just stumbled into this again with #2349: synthetic `Pressed` events were handled but clearly shouldn't have been in this example. For what it's worth, I'd like...

Well, uniform behaviour across platforms would be a reason: see [above](https://github.com/rust-windowing/winit/issues/1272#issuecomment-609384601). But in that case, we shouldn't generate synthetic key-press events, and should not emit (non-synthetic) key-release events if no...

> Also, requires updating MSRV to 1.89 for stable AVX-512 intrinsics support. I would prefer we make a stable release for `rand_core` v0.10 before merging this.

Would you prefer to merge changes piecemeal (just this commit) or shall I push more IME changes into this PR? The plan is [roughly outlined here](https://github.com/rust-windowing/winit/issues/4412#issuecomment-3551431472).

As you like. I already made further changes (not pushed yet): move `winit::window::Ime*` types to new `winit_core::ime` module, drop `Ime` prefix for these types.

Pushed. My further plans are less concrete: - Split enable / update into separate requests. For the interface I think this makes sense but I'm hesistant because these ops share...

FYI this will clash with #4324. I suggest waiting until that is merged since it was already approved.

Note: currently my code recalculates the "cursor area" every frame simply because there is no easy way to know whether any widget translation (scroll position) has changed, so though it...