* reorganize GlyphCache::get_or_insert() to use the HashMap entry API
* change DrawGlyphs::hint() parameter name from `yes` to `hint` for clarity
* change Layout::path_data() to return bytes instead of [f32; 2]. Path segments could be encoded as i16, making this incorrect
The test scenes can now supply their own optional base (background)
color. In the with_winit example, we also allow the user to provide a
base color as a CLI option. The precedence is as follows:
1. Use the color from the CLI options, if any.
2. Otherwise use the scene provided base color, if any.
3. Otherwise default to black.
The texture and surface render API now takes render-time parameters
(such as clear color, target width/height) as a RenderParams struct.
The examples have been updated to demonstrate this. The with_winit
example now accepts a clear color as a command line option.
* make SimpleText::add_run accept glyph_transform and style parameters so it doesn't unconditionally do oblique strokes
* replace fill/stroke methods on DrawGlyphs with a single draw method that accepts either fill or stroke styles
* update peniko rev to access the new style types used above
* for now, change glyph cache to only cache non-zero fills. Prior to this, style was ignored in the key which could lead to incorrect rendering.
This fixes an incorrect application of the inverse transform for radial gradients in fine.
Also fixes an edge case in `SceneBuilder` where a brush transform is identical to the path transform leading to a corrupt encoding.
Added two scenes that demonstrate conflation artifacts as described in
https://github.com/linebender/vello/issues/49. The first scene
demonstrates adjacent triangles and rects that belong to the same path
and use opposite winding.
The second scene demonstrates strokes with overlapping square caps
(these strokes are currently expressed as rects painted with the NonZero
fill rule).
Just load the atomic bump counter directly instead of piping it through a shared variable, when workgroupUniformLoad is not available. The value is in fact dynamically uniform, but that depends on the stage not setting its own failure flag, a fairly subtle invariant.
I think there was a write-after-read hazard for the reuse of sh_part_count[0]. However, doing the experiment of just changing that doesn't fix the problem on mac. It's possible there's a shader compilation problem (possibly the same one as provoking the storageBarrier workaround in tile_alloc), or also possibly a logic error I'm not understanding.
In any case, this change does appear to fix the hangs on mac.
Fixes#267
