76fbbc89e9
Adds a line primitive back to geom::Shape since I just couldn't tolerate drawPoints mallocing a path for every line. When Shape provides a path like iterator, it will look equivalent to an SkPath that is a line, so there won't be any reason to analyze an incoming path for a line, but if we know we want a line, there's no reason to wrap it in a path. DrawList and DrawContext now take Shapes for their path rendering funcs, but are documented to use path rendering. It won't be like GrSDC that tries to choose the algorithm under the hood from you. Device has been heavily updated to funnel all the draw calls into a new drawShape(), so the primitive functions do not malloc skpaths any longer, even if they still end up using path renderering. Additionally, instead of making new paints that have updated styles or removed components, drawShape() takes an SkStrokeRec and ignores what's on the paint, and has additional flags that tell it to ignore aspects of the paint. This is used when it recurses, or for cases like drawPaint where path effects and mask filters should be ignored (and eventually other things like drawImage will ignore path effects and force a fill style). Bug: skia:12466 Change-Id: I4700c895ce3fefe2e437f3b4d329fd381593e037 Reviewed-on: https://skia-review.googlesource.com/c/skia/+/457398 Commit-Queue: Michael Ludwig <michaelludwig@google.com> Reviewed-by: Robert Phillips <robertphillips@google.com> |
||
---|---|---|
.. | ||
c-api-example | ||
docs | ||
documentation | ||
ffmpeg | ||
graphite | ||
lowp-basic | ||
minimal_ios_mtl_skia_app | ||
pvg | ||
skottiekit | ||
skrive | ||
sktext | ||
sorttoy | ||
tools | ||
tskit | ||
wasm-skp-debugger | ||
webgpu-bazel | ||
xform | ||
xps_to_png |