@UlrikeFischer \undefined is only reported for directly input undefined control sequences, if it's inside a macro or the undefined sequence already appeared in another context then the error is correct. It's still annoying though.

@dedded Yes it's a rather typical LaTeX interface (at least it's supposed to be). The setup depends a bit on the platform, for Linux and Windows you can just follow the README and then it should work, for other platforms things get complicated since things need to be compiled.

@dedded github.com/zauguin/luametalatex is the repository with a README.

@TeXnician Thanks, I'll try to play with that tonight.

@barbarabeeton As usual, your wish is my command! (Hyphenation exceptions)

@egreg -- Thanks. I'm planning an update for early next year.

Still true.

@AlanMunn ooh pirates

@TeXnician we could issue a visa for Marcel :)

@PauloCereda And buy tickets with a famous airline?

@mickep ooh a secret dodo plan

@mickep If you choose the locations of 𝜓 in fontforge one at a time, the leftmost number in the line below the menu (i.e., the number below “File”) changes. I don't know what this number means.

@AlbertNash Thanks for looking at it. I think that is just some "running number" in the font.

@AlbertNash Regarding your recent comment in the second Termes question, I think we can conclude that the left sidebearing is working pretty well most of the times, right?

@mickep Except the original problem, i.sstatic.net/a8Ndb.png . Though you might say that this example could be viewed as a bit artificial, I really see no reason why anyone wouldn't start two lines below each other with 𝑊 or 𝑌 in one line and 𝐿, 𝑋 or 𝑍 in the next one or vice versa. It might even happen automatically during line breaking.

@mickep (Originally, my own motivation was to distinguish the widths of various symbols, which per se has nothing to do with 𝑊 or 𝑌 being moved to the right. However, if one does correct 𝑊 or 𝑌, one might also wish to set their widths properly.)

@mickep Having said this, you are right: most of the times, the result is good (except right subscripts, which one can readily improve locally). What really surprises me is WHY the right shift of 𝑌 is not visible in “𝑋𝑌𝑋” and in my real-usage examples except “𝑋∪𝑌 begins”.

@AlbertNash Look at this:

And perhaps also at:

@AlbertNash As mentioned by @barbarabeeton yesterday, also computer modern italic also has a large sidebearing to the left. (And not too many people seem to have noticed or thought about that during the years.)

@dedded I saw that @TeXnician already sent you the link. If you make it work please send me the binaries so that I can upload them too. :) Otherwise I'll try to build a OSX version, but building for OSX is always a bit complicated...

@MarcelKrüger I cannot even afford a Mac. :)

@PauloCereda I don't have one either, that's what makes building for it so hard...

@MarcelKrüger ooh we could try Kotlin MPP :)

@mickep For a reason unclear to me, in your diagram of the bounded boxes of 𝑋𝑌𝑋, the distance between the two bounding boxes in 𝑋𝑌 is smaller than the distance between the two bounding boxes in 𝑌𝑋. This makes up for 𝑌 being moved to the right with respect to its bounding box. Perhaps, there is some nonzero pairwise kerning… I don't know. Anyhow, this is a strange kludge.

@dedded I uploaded a installer for a prebuild OSX version under the link in the repo, but I have absolutely no idea if it works. If you try it please let me know what happens...