iris-ai
8223a55cfb
`try_reuse`'s pure-translation case now writes the widget's move slot instead of remapping every primitive in its subtree. Counted on a span of 20 rows, each five primitives deep, when the row above them changes height: before 100 primitive region writes now 0, and 20 slot writes -- one per row the span re-placed `window_region` walks the same chain on the CPU, so hit testing and anyone asking in window pixels see a widget where the shader draws it. `Moves::resolve` stops at `CHAIN_LIMIT` like the shader, and asserts in debug that it got to the end rather than running out. Two things fall out of it. Rewriting a region is the one thing a slot cannot express, so `mov` zeroes the slots of everything it rewrites: a region is what its slot was a delta from. And `try_reuse` loses its `old == region` shortcut, which was wrong once a slot exists -- a widget offered exactly the box it drew against has to have its delta cleared, not skipped. `Moves` lives on `UiRenderState` rather than `UiData`, because the draw is what produces it and `window_region` should not need the ui's resources to answer where something is. The renderer already takes both. A slot is retired in `remove_rec`, after the descendants whose slots name it as their parent. Either order is correct here -- nothing can claim a freed index while a subtree is coming down, since `on_undraw` cannot reach the slots -- but this way `remove`'s `undraw` flag only notifies rather than also deciding slot lifetime, and the retirement sits beside the recursion it follows. Checked: fmt, clippy and 41 tests. `tabs` (with the image replay), `view`, `minimal` and `text` all still render byte-identical, and the live sway resize round trip -- which re-places most of the tree at the same size, so it is the slot path throughout -- matches a cold start at each size.
iris
My experimental attempt at a rust ui library (also my first ui library).
It's currently designed around using retained data structures (widgets), rather than diffing generated trees from data like xilem or iced. This is an experiment and I'm not sure if it's a good idea or not.
Examples are in examples, eg. cargo run --example tabs.
Goals, in general order:
- does what I want it to (text, images, video, animations)
- very easy to use ignoring ergonomic ref counting
- reasonably fast / efficient (a lot faster than electron, save battery life, try to beat iced and xilem)
dev details
not targeting web rn cause wanna use actual nice gpu features & entire point of this is to make desktop apps / not need a web browser
general ideas trynna use rn / experiment with:
- retained mode
- specifically designed around wgpu so there's no translation
- postfix functions for most things to prevent unreadable indentation (going very well)
- events can be done directly where you draw the widgets
- almost no macros in user code & actual LSP typechecking (variadic generics if you can hear me please save us)
- relative anchor + absolute offset coord system (+ "rest" / leftover during widget layout)
- single threaded ui & pass context around to make non async usage straightforward (pretty unsure about this)
- widgets store outside of the actual rendering so they can be moved around and swapped easily (unsure about this but seems to work good for now)
under heavy initial development so not gonna try to explain status, maybe check TODO for that; sizable chance it gets a rewrite once I know everything I need and what seems to work best
it's called iris because it's the structure around what you actually want to display and colorful