@cis I would not use superfluous to mean no longer important. I would read "XYZ became superfluous" as "XYZ is no longer necessary or needed". Although it might help to know what XYZ is.
@PauloCereda I saved this one for you to answer, as I know you would love to mention the editor with the original and greatest read-eval-print-loop tex.stackexchange.com/q/602487
@JosephWright not an interactive lisp session built in to your editor?
;; This buffer is for text that is not saved, and for Lisp evaluation.
;; To create a file, visit it with C-x C-f and enter text in its buffer.
(+ 1 2 3)
6
(cos pi)
-1.0
@DavidCarlisle For non unicode-math fonts the design already takes this into account and just needs appropriate lookup tables. The same could be done for unicode-math, but first unicode-math has to decide what to do with these.
@DavidCarlisle Do you know of any test fonts which actually use these?
@MarcelKrüger is it expected that the fonts actually implement this internally or do you have to see teh variation selector and then dig out the alternative glyph?
@DavidCarlisle In general, OpenType contains special cmap tables for variation selectors and these should be used inside of the font. For TeX in particular, we still need to see the variation selector and dig out the alternative glyph based on this since the default support for variation selectors does not apply in math mode. I have some code for this somewhere (TM).
@MarcelKrüger ah OK thanks I think that sounds vaguely familiar so I may have known that once:-) So in theory you could support this in luatex even if the fonts don't change (if they have the two styles as stylistic variants for example so the glyphs are "somewhere") ?
@JosephWright unless we get braver and rip out the whole of color.sty and make it a wrapper for l3color
@DavidCarlisle Right, but I would expect this to be mostly a output concern, I would expect I to be pretty unlikely that people start using 𝒜︀ and 𝒜︁ as LaTeX input.
@MarcelKrüger sure but you may get that in conversions from Word to LaTex, and also if it worked then logically \mathscr{A} and \mathcal{A} would be defined to produce the variant selector pairs wouldn't they?
@MarcelKrüger remind me to come back and look at the transcript here in a years time and see if those A look different:-)
@DavidCarlisle Mapping \mathscr{A} to the glyph described by the variation selector will probably work by a different mechanic than mapping literal input of the variation selector. (Maybe a separate family for changing the default style? Or mathcodes representing internal glyph codes?) Something for Will to think about...
@JosephWright the patches I use in pdfmanagement for xcolor seems to work, until now I hadn't much problems. But probably we really need to get in contact with Uwe Kern to sort that out properly.
@DavidCarlisle I just tried to convince Herbert to take over some of the seminar patches in hyperref, and the side-effect is that we discovered that seminar is broken in a new latex as it redefines \document ;-(.
@DavidCarlisle my former me seems to have known it ;-). I wonder, does hyperref have to contain support for a class that is broken in a current latex? Perhaps I could move them into a "old-hyperref-patches" package and people could load or copy them if they need them.
@DavidCarlisle well the question is if he wants to fix. Last year when I notified him about the document problem, he wrote that he no longer wants to touch seminar. And I guess the guy in the question wrote him too and nothing happened.
@DavidCarlisle -- I can understand the "reserved" locations in the math alphanumerica (those letters are in the "letterlike symbols" block), but can't figure what the ones in this block are reserved for. (Unicode isn't always scrutable.)
@DavidCarlisle -- If variation selectors are to be used for calligraphic letters, the least that should be done is to document it in TR#25. (Bug Murray.)
@DavidCarlisle -- But the official Unicode variants aren't documented yet that I can find. So even if Cambria and stix provide the glyphs, there's no Unicode backing.
@JosephWright recently, I saw that \__quark_if_empty_if:o still uses \ifx internally, instead of \if (was among the bunch of minor speed ups in l3tl). Is there any particular reason why \@@_if_empty_if:o is needed twice (ones inside l3quark once inside l3tl)? Couldn't we just do a \cs_set_eq:NN in whichever module is loaded later, or just use the same internal macro (maybe moving it to the \__kernel name space).
@DavidCarlisle -- Yes, you already posted that. It only says "these are similar". The text at the top of that section is nice, but there need to be at least two distinct codings. Among the examples I provided, there were at least two that showed distinctly different math usages in the same document for script vs. calligraphic. That requirement hasn't yet been satisfied.
@barbarabeeton It has hasn't it? the unmodified U+1D49C can be an A in either style but there are now two variant selectors to force caligraphic/chancery or script/roundhand
@DavidCarlisle -- Yes, you're correct. (I just had my annual eye exam, and my eyes are still fuzzy.) Okay, thanks. I will ask Murray, politely, to add this to TR#25. (And at the same time mention bold math.)
@DavidCarlisle -- Do you remember seeing any evidence that the two variation selectors were added to the script letters in block 2100 (letterlike symbols)?
@barbarabeeton unicode.org/L2/L2021/21009.htm, B.1 — 25: "Consensus: The UTC accepts 52 variation sequences to distinguish roundhand and chancery style mathematical script alphabetic characters, for Unicode version 14.0. (Reference: L2/20-275R)"
@MarcelKrüger -- Yippee! Well, that takes care of the lightface. I should still check with Murray about the boldface and TR#25. Thanks. (Who's your mole on the UTC?)