Epic: Unsupported MathML elements and attributes

[Omikhleia]

This is an "Epic" (in the sense of a parent issue for related sub-issues)

Maybe this is lost in some old discussion on some of the MathML implementation issue? Anyhow, here's a list (to be enriched possibly).

Some constructs SILE's MathML implementation does not support as of v0.15.5:

• multiscripts (f2) - and mprescripts
• mstyle (f6) - this one is legacy in MathML Core = layout as mrow (at least) = Math more fixes and features #2151; see below for styling
• mpadded (f22) - see Math more fixes and features #2151
• mtext (f18, f19, f22) - honestly this one is much needed and shouldn't be complex... see feat(math): Support MathML mtext and ms elements #2140
• mphantom (f22) - see Support MathML mphantom and TeX-like phantom, hphantom, vphantom #2143
• msqrt - see feat(math): Support square roots #2119
• mroot - see MathML mroot with degree not supported. #2120
• ms - see feat(math): Support MathML mtext and ms elements #2140
• maction - MathML core says it can be considered as an mrow where only the first child is rendered = see Math more fixes and features #2151
• menclose - not in MathML Core, only MathML 3-4, but in the absence of CSS it's useful. MathJax supports it. - see Math boxing and cancelling commands #2318

Some other unsupported things

• Most attributes on elements, ex.
  • "linethickness" on mfrac (f10), -- see Math more fixes and features #2151
  • "form" on mo (for mphantom esp.),
  • Consistent support of "displaystyle" and "scriptlevel" on several tags - interesting in several cases (e.g. TeX cfrac)
  • mathcolor, mathbackground - see Support more MathML attributes and other math fixes and enhancements #2340
  • etc. (but what else would we want?)
• Improper sizing of some parentheses (several cases) - Math more fixes and features #2151 addressed the stretchy attribute and I think that was it!
• Stretchy horizontal braces (f19; f22) - some naive solution proposed in Math more fixes and features #2151 (test-based, maybe not general)
• Very large stretchy operators (whether done by glyph assembly or faked, they must be improved)
  • Very large vertical stretchy parentheses (f23) - see Math more fixes and features #2151 (achieved with a hack/workaround)
  • Very large vertical stretchy determiner bar (f24) - same hack comment as previous

Other topics

• Styling ought to be discussed. MathML core makes many attributes and the mstyle element legacy, but uses CSS instead... A hornets' nest...

Notes

• For cross-checking see also https://w3c.github.io/mathml-core/#mathml-elements-and-attributes
• The (f-numbers) above are the tests in https://fred-wang.github.io/MathFonts/mozilla_mathml_test/ - Whatever the status of this (old) 30-scenario Firefox MathML test, having the numbered items above would imply it passes for us (all other unmentioned tests are already passing).

[Omikhleia]

Besides the above-mentioned old Firefox test, another interesting test to check: http://eyeasme.com/Joe/MathML/MathML_browser_test.html (EDIT: It has both the TeX sources and the MathML code = we could also use it for checking our TeX-like math...)

[Omikhleia]

And of course, another great series of examples and checks... is just the MathML test suite... I just started trying some of the "torture" tests, and it's... interesting ;)