unicode-math
unicode-math copied to clipboard
latex3
Fall-back characters
See discussion in http://tex.stackexchange.com/questions/12877
There \perp maps to U+27C2 which is not in Cambria. Unicode says U+27C2 is perpendicular relation. MathML suggests U+22A5 for that. Therefore we should have \mathcode U+27C2 = U+27C2 \mathchardef \perp = U+27C2 if the font has U+27C2, and \mathchardef \perp = U+22A5 otherwise; even \mathcode U+27C2 = U+22A5 might make sens in the latter case. This change can probably be done only after we have switched to the flexible Lua table.
@phst — do you remember any thing about "the flexible Lua table" we were discussing then? I seem to remember a few approaches that people had taken but I've lost track of them. From memory, someone had already taken the unicode-math table and produced a Lua table from it — do you recall?
@phst — ah, maybe i was thinking of this: http://milde.users.sourceforge.net/LUCR/Math/ (from #190). It's not a Lua table but it is comprehensive.
Has there been any progress? I'm trying to use a sans serif math font (Fira Math). A number of characters are missing, which MdSymbol has defined, but trying to use both simultaneously is proving quite difficult
@beojan — I strongly suspect that unicode-math will never support loading a non-Opentype Math font in parallel.
MdSymbol has an Opentype version, which is what I'm trying to use. It actually seems to work fairly well, and using the range option I've got it using MdSymbol as the main math font and Fira Math for Latin, Greek, and Numerals. At this point the problem is simple symbols like the comma which MdSymbol doesn't define but don't fall into the available ranges in unicode-math.
Oh, that’s good to hear. Sorry for the misunderstanding. In that case could you open a separate issue for the specific problem you’re facing?
#45 is a close match, and it was closed and merged here. Maybe you could reopen that?
Just leaving a comment here as I was looking for issues with range in an answer I just posted at tex.stackexchange
https://tex.stackexchange.com/a/622610/1090
In my answer and @eg9's we "filled the gaps" by setting \Umathcode for the affected slots (which is OK for the digits here although text shaping in general would be problematic) But was my expectation that range= would do this unrealistic?
