4 Commits

Include renames

Author	SHA1	Message	Date
adminandClaude Fable 5.1	cc6db5f350	Three stores, three importers and three pickers that differ only in their type are one of each (#85) ... Refactoring round one: HttpFontStore/HttpSoundStore/HttpTextureStore collapse into a generic HttpFileStore over a FileStore contract; fontImport/soundImport/textureImport share assetImport with an ImportKind per sort; ImagePicker/FontPicker/SoundPicker sit on FilePicker and describe only their row. The group/comment drag logic leaves GraphCanvas for annotationDrag.ts (with its own unit test), declaredRows.ts replaces the duplicated row build in particles.ts and shaders.ts, and dead code in font/sound/compile/emission/transform/actions is gone. README: asset list, check count 958, the two-job CI, the Scenes wiki page. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>	2026-09-16 17:38:54 +02:00
rootandClaude Opus 5	d5e67cd80c	A shader costs a source nothing, because a field was already one draw call (#88) ... A particle source can wear a shader asset the way a sprite does, and it is the cheapest place in the editor to wear one: a shaded sprite gives up its batch and costs a draw call of its own, while `ParticleContainerPipe` breaks the batch for every field regardless. Five hundred shaded particles are the one call five hundred plain ones were. The fragment the compiler emits needed no change, only a second vertex stage: Pixi's particle vertex is GLSL ES 1.00 and ours is ES 3.00, so it is translated rather than invented. What it sees is a particle's picture, not its life — the buffer carries five fixed attributes and takes no sixth, so age, index and seed stay on the graph, which is what the graph is for. Two things the issue did not ask about: `Shader.resources` is an accessor over the names its constructor declared, so a slot that is not declared becomes a loose JS property and is never bound. The adaptor writes `uTranslationMatrix` into whichever shader the container names — without an empty block for it the vertex gets no matrix and the particles land elsewhere. `schemaOf` in serialize.ts hung on `schema !== schemaFor('sprite')`, an identity check a source never satisfies, so its shader values would have been dropped on save. A source is the one asset drawing rows from two graphs, so a name can arrive twice. Its own row wins and the uniform keeps its default; nothing is rejected on the shader's side, because the collision exists only at this one wearer. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>	2026-09-14 11:06:45 +02:00
rootandClaude Opus 5	8ae20d044f	Time is an argument, so speed scales it and a prewarm offsets it (#87) ... Speed and Prewarm are not state and not a second loop: both are arithmetic on the clock argument the field already passes to every particle formula. fieldTime = elapsed * timeScale + prewarmOffset, a pure rule next to liveRange, called at the one place ParticleField.sync forms its time, so preview and world inherit it together. Prewarm offsets by one lifetime because that is the smallest offset which fills a continuous emitter; a burst is dead at time >= lifetime, so the same offset would skip it entirely and it is suppressed there. The tab's slider counts wall-clock seconds, so its span divides by the scale. A missing row falls back to 1 / off, so old project files need no migration. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>	2026-09-14 07:31:07 +02:00
root	0f9e1611d1	A particle is a function of its age, not a thing that is stepped (#73) ... The asset type `particles` is placeable and carries a graph at the same time — `behavior` was the blueprint. Because a continuous emitter gives birth to particle n at time n/rate, the whole field is a function of the clock rather than of the frames: no ring-buffer bookkeeping, exact time jumps, reproducibility as a measurable property. compile.ts follows compileShader's frame but emits JS closures into a Float64Array instead of GLSL, and without eval; a node with two consumers is evaluated once. Drawing is one Pixi ParticleContainer per source, so one draw call. Collision, wind and braking in flight are out for good — that is the price of the formula model, not an omission. The tab count goes from nine to ten, which is exactly the case #84 was built for one day earlier: at 1280 px the row no longer holds every tab, so guard 14b now asks for hidden.length > 0 instead of === 0.	2026-09-14 05:44:25 +02:00