Brian Sweeney
Brian Sweeney
I'm going to keep this issue open so we can address the underlying bug for a forthcoming release.
There's potentially a few things going on here, but the biggest problem is that the line height calculation is off. I noticed that we weren't really doing a good job...
@gochinj nice, I'll definitely be using that bit of code since I, unfortunately, hadn't had a chance to do this. I didn't expect it to require too much code.
What were your thoughts regarding block rendering. I was thinking we render from the last child instead of the first child for containers where direction is "rtl". But it would...
I was looking at this and I have some ideas about 712. Let me collect my thoughts and get back to you.
@simonberger I went ahead and resolved a conflict in the Style class so this PR should be up-to-date with the changes I made.
I'll take another quick look, but I think we're probably good to merge.
Regarding subsetting: So as you can see the whole reason we do subsetting the way we do is because the subset is built at the time the font is selected....
With the most recent change to the code I'm seeing an odd character replacement with the sample I posted above using the Amiri font. Spaces are being replaced by exclamations...
@simonberger I think that text is _supposed_ to be rendered RTL due to strong directionality of the characters. This goes back to our discussion of enabling the bidi parsing at...