Add support for character-level and glyph-level mirroring of RTL operators #277
Conversation
spec.html
Outdated
| <ul> | ||
| <li> | ||
| If <code>c</code> can be mirrored, then return | ||
| the glyph corresponding to the result of mirroring <code>c</code>, and exit with success. [[BIDI]] |
There was a problem hiding this comment.
Can you please give us some references to the spec definition of "can be mirrored" and "mirroring"? That will help review this patch.
There was a problem hiding this comment.
Sure -- in particular, it's this section of UAX9:
https://www.unicode.org/reports/tr9/#Mirroring
OpenType rtlm falls under HL6, I believe, so no check on Bidi_Mirrored is necessary:
https://www.unicode.org/reports/tr9/#HL6
The "corresponding codepoints" mentioned are given in Data9:
spec.html
Outdated
| the glyph corresponding to the result of mirroring <code>c</code>, and exit with success. [[BIDI]] | ||
| </li> | ||
| <li> | ||
| Otherwise, if there exists an OpenType rtlm variant of the input glyph, return that and exit with success. [[OPEN-FONT-FORMAT]] |
There was a problem hiding this comment.
spec.html
Outdated
| Otherwise, if there exists an OpenType rtlm variant of the input glyph, return that and exit with success. [[OPEN-FONT-FORMAT]] | ||
| </li> | ||
| <li> | ||
| Otherwise, return the input glyph and exit with success. |
There was a problem hiding this comment.
The algorithm is always successful... so I guess we don't need to specify "with success" here and elsewhere?
I think we indeed don't need to fail when a glyph is not mirrorable in RTL, we should just stretch the non-mirrored glyph.
There was a problem hiding this comment.
Yes, I agree, I've added this extra fallback step. I initially explicitly specified "with success" since I thought that this could fail if it wasn't possible to get any glyph for a codepoint (i.e. if the font didn't map one), but comparing this to the old version of the spec this should be fine(?).
icesfont
force-pushed
the
feat/rtl
branch
3 times, most recently
from
7ed76b0 to
349f89b
Compare
|
I forgot to reference it, this is related to this issue: #67 |
Depends on #276.
Please see #67 and https://people.igalia.com/fwang/mathml-operator-mirroring-explainer.html.
This augments the various steps that get the glyph for the character in the "layout of operators" algorithm to additionally mirror the character/glyph in RTL writing modes, using Unicode's BIDI or OpenType's rtlm respectively.