Crafter.Graphics/implementations/Crafter.Graphics-UI-Shared.cpp
catbot 65c19ea84a perf(ui): gate per-pixel clip compares behind a flags bit (#50)
uiResolveScreenPixel ran four float compares against hdr.clipRectPx for
every pixel of every UI dispatch, even though the clip rect is the
full-surface default {0,0,1e9,1e9} in the overwhelmingly common case.

Reserve the high bit of the header flags word (UI_FLAG_CLIP) and have
FillHeader set it only when the clip rect is narrower than the surface
(extracted into the unit-testable UIRenderer::ClipFlags). The shader
gates the compares on the bit; the branch is push-constant-uniform so it
never diverges. When the rect covers the surface the skipped compares
could never reject an in-surface pixel, so output is pixel-identical.

Mirrored on the WebGPU/WGSL path (dom-webgpu.js) for parity. Adds the
UIClipFlag CPU test covering the gate decision.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-16 16:57:07 +00:00

141 lines
5.4 KiB
C++

/*
Crafter®.Graphics
Copyright (C) 2026 Catcrafts®
catcrafts.net
*/
// Backend-agnostic UIRenderer methods. FillHeader and ShapeText do not
// touch any GPU API; they only read window dimensions and (for ShapeText)
// the CPU-side font atlas. Split out so both Vulkan and WebGPU impls
// share the same source.
module Crafter.Graphics:UI_shared_impl;
import :UI;
import :Window;
import :Font;
import :FontAtlas;
import :GraphicsTypes;
import std;
using namespace Crafter;
UIDispatchHeader UIRenderer::FillHeader(std::uint32_t itemBufferSlot,
std::uint32_t itemCount,
std::array<float,4> clipRectPx,
std::uint32_t flags) const noexcept {
UIDispatchHeader h{};
h.outImage = outImageSlot_;
h.itemBuffer = itemBufferSlot;
h.surfaceWidth = window_->width;
h.surfaceHeight = window_->height;
h.clipX = clipRectPx[0];
h.clipY = clipRectPx[1];
h.clipW = clipRectPx[2];
h.clipH = clipRectPx[3];
h.itemCount = itemCount;
#ifndef CRAFTER_GRAPHICS_WINDOW_DOM
h.frameIdx = window_->currentBuffer;
#else
h.frameIdx = 0;
#endif
h.flags = ClipFlags(clipRectPx, window_->width, window_->height, flags);
h._pad = 0;
return h;
}
std::uint32_t UIRenderer::ClipFlags(std::array<float,4> clipRectPx,
std::uint32_t surfaceWidth,
std::uint32_t surfaceHeight,
std::uint32_t flags) noexcept {
// The reserved kUIFlagClip bit signals "clip rect is narrower than the
// surface". When the rect already covers the whole surface (the common
// {0,0,1e9,1e9} default), the four per-pixel clip compares in
// uiResolveScreenPixel can never reject an in-surface pixel, so we leave
// the bit clear and the shader skips them — pixel-identical output.
const bool clipCoversSurface =
clipRectPx[0] <= 0.0f && clipRectPx[1] <= 0.0f
&& clipRectPx[0] + clipRectPx[2] >= static_cast<float>(surfaceWidth)
&& clipRectPx[1] + clipRectPx[3] >= static_cast<float>(surfaceHeight);
return clipCoversSurface ? (flags & ~kUIFlagClip) : (flags | kUIFlagClip);
}
std::uint32_t UIRenderer::ShapeText(Font& font, float pxSize,
float x, float baselineY,
std::string_view utf8,
std::array<float,4> color,
GlyphItem* out, std::uint32_t outCapacity,
float* outAdvance,
TextAlign align) {
if (fontAtlas == nullptr) {
std::println("UIRenderer::ShapeText: no FontAtlas (set fontAtlas before Initialize)");
std::abort();
}
// Look up (or build) the origin-relative shaped run. It is stored with
// the pen starting at x=0, baseline=0 so a cache hit only needs the
// translate below. The whole string is shaped (no outCapacity limit) so
// a later call with a larger buffer still gets the full run.
ShapedRunKey key{&font, pxSize, color, std::string(utf8)};
auto it = shapedRuns_.find(key);
if (it == shapedRuns_.end()) {
// Miss: shape from scratch. Ensure() each glyph so brand-new ones
// rasterise into the atlas and mark it dirty for the next upload —
// a hit skips this, which is safe because atlas entries are never
// evicted once placed.
const float scale = pxSize / FontAtlas::kBaseSize;
ShapedRun run;
float cursor = 0.0f;
std::size_t i = 0;
while (i < utf8.size()) {
std::uint32_t cp = DecodeUtf8(utf8, i);
if (cp == 0) break;
if (cp == '\n') { continue; }
const Glyph* g = fontAtlas->EnsureAndGet(font, cp);
if (g == nullptr) continue;
if (g->w > 0 && g->h > 0) {
GlyphItem gi;
gi.x = cursor + g->xoff * scale;
gi.y = g->yoff * scale;
gi.w = g->w * scale;
gi.h = g->h * scale;
gi.u0 = g->u0; gi.v0 = g->v0;
gi.u1 = g->u1; gi.v1 = g->v1;
gi.r = color[0]; gi.g = color[1]; gi.b = color[2]; gi.a = color[3];
run.glyphs.push_back(gi);
}
cursor += g->advance * scale;
}
run.advance = cursor;
if (shapedRuns_.size() >= kMaxShapedRuns) shapedRuns_.clear();
it = shapedRuns_.emplace(std::move(key), std::move(run)).first;
}
const ShapedRun& run = it->second;
if (outAdvance) *outAdvance = run.advance;
// Fold the alignment shift into the translate so callers don't need a
// second pass over the glyphs.
const float alignShift = (align == TextAlign::Center) ? -run.advance * 0.5f
: (align == TextAlign::Right) ? -run.advance
: 0.0f;
const float tx = x + alignShift;
std::uint32_t written = 0;
for (const GlyphItem& src : run.glyphs) {
if (written >= outCapacity) break;
GlyphItem& gi = out[written++];
gi = src;
gi.x += tx;
gi.y += baselineY;
}
return written;
}
void UIRenderer::InvalidateFont(const Font& font) noexcept {
std::erase_if(shapedRuns_, [&font](const auto& kv) {
return kv.first.font == &font;
});
}