Conversation
whereswaldon
left a comment
There was a problem hiding this comment.
Thanks for putting this together @benoitkugler! I think it looks reasonable, though I lack your deep expertise in the internals of opentype. My only real question is whether we need so many implementations of binary search? Not a huge deal though. Great work!
I'm not a big fan of the sort.Search API, but I guess it is OK as a temporary solution until we can use go 1.21 and |
andydotxyz
left a comment
There was a problem hiding this comment.
Sorry it took so long. I agree the search could be cleaner - but also that having it this way until base version increases to 1.21 makes sense.
I'm hoping to bump Fyne version to there soon (months though, not days)...
| } | ||
|
|
||
| func TestLoadColor(t *testing.T) { | ||
| ld := readFontFile(t, "color/NotoColorEmoji-Regular.ttf") |
There was a problem hiding this comment.
I can't see those files in this PR or the repo before the PR - apologies if I am missing something obvious.
But given that I cannot understand how the tests are passing.
There was a problem hiding this comment.
We are loading the test files from the typesetting-utils module, so that typesetting remains lightweight.
This is a first step towards support for the COLR/CPAL tables.
We expose a new
GlyphColortype returned byFace.GlyphData. With this API, a renderer may now load the layers defining a colored glyph and combine them. This could be implemented for instance in go-text/render.This PR is probably not enough to support variable fonts, but starting with non variable ones should already be pretty nice.