Switch all the way to expl3 int_compare_p instead?

@Mico: No need to delete you answer. It works as you stated. Just because I would prefer a local version, it does not mean to imply that your answer was not good. Also, I think it is a mistake to only provide answers that solve the issue exactly as the OP requests. Your answer would be useful for others looking for such a thing.

@PeterGrill I don't think he's been here recently enough to see that ping (the name isn't auto-completing) you need to comment on the question (whichever that was)

@CarLaTeX in case you haven't seen it: you got mail.

@Skillmon I didn't see it, replied :) Thanks!

@UlrikeFischer most likely we should fix this in the latex code if the engine can't be changed to make compatible byte code, presumably in \everyjob we could do effectively if (something) unpack bytecode else execute the original lua definition where (something) could be a safe test or try to unpack and trap errors or ...

@DavidCarlisle perhaps. Do you know how that impacts the performance? (on the other side: at least on my system the 32 bits and 64 weren't quite often not really in sync regarding the versions, so I needed different formats anyway.). Life is curious ;-) Just when I write that it is probably not a problem in the real world a real world problem pops.

@UlrikeFischer yes I thought same about real world timing. performance impact would be minimal in the usual case, although you would have to access the original lua presumably off the file system in the other case but still better than failing.

@UlrikeFischer Ping

@moewe ah thanks. I want to ask if you have some bib files to test and show utf8/unicode examples. E.g. with greek or russian etc. I found caspervector.bib and one in biblatex-gost but the more the better.

@UlrikeFischer Off the top of my head I don't know any. I usually only use biblatex-examples.bib which uses ASCII escape for BibTeX compatibility.

@moewe pity ;-). I think it could be useful to get more examples in this area. Even if one ignores the additional problems of multiscripts, it would be good to know if and how well other scripts can be handled.

@UlrikeFischer I just remembered one thing: 93-nameparts.tex in doc/examples. It has a few examples of non-Western names some of which use different scripts.

@moewe ah yes. That is nice too.

@UlrikeFischer Petr Olšák in texmf-dist/tex/optex/demo/op-biblist.bib ?

@UlrikeFischer bibtex/bib/testidx/testidx-glossaries-diglyphs-utf8.bib (although that's not really a bib file for a bibliography)

@DavidCarlisle how do you search for them?

grep -l '[^a-zA-Z0-9= .;,:?@#/{}\\"'"'"'()%-]' `find /usr/local/texlive/2021/texmf-dist/bibtex/ -name \*.bib`

@UlrikeFischer ^^ more or less then checked a couple by eye

@DavidCarlisle helpful, I found a few more in the doc folder with it ;-)

@UlrikeFischer yes I pruned out doc as it's not in the default input path so harder to use in examples

Oh, so now there is a Concrete Math otf font (ctan.org/pkg/concmath-otf). Nice! :)

@mickep I wonder if @PauloCereda wrote section 3.6 of the manual

@DavidCarlisle Oh, you are reading the manual? I thought you wrote them...

@mickep don't believe everything you read on the internet (especially this bit of the internet)

@DavidCarlisle Oh, the albatross...

@DavidCarlisle I think @PauloCereda goes free this time. If he wrote it there would be an "oh" before the "no" in Missing character: There is no ⥤ (U+2964) in font ErewhonMath

@UlrikeFischer @AlexG github.com/latex3/graphics-def/commit/…

@DavidCarlisle Ah good. I just wanted to ask about it.

@UlrikeFischer anything else pending or should I push to ctan?

@DavidCarlisle I don't think that there is something else.

@UlrikeFischer done

@DavidCarlisle ooh

:60350137: Yeah was afraid of that. Thanks for letting me know. And congratulations on finaly being called out for all the rubbish you post in the answers. :-)

@PeterGrill I'm most shocked by @Skillmon's scurrilous claim that my postings are subject to typos

What is the trick, to make a proper standalone-class-output? Is this allways something with vbox / hbox?

@DavidCarlisle he probably meant "If submitter was David Carlisle, answer is factually right, might contain features"

drive.google.com/uc?id=1eL2ElvZ-qbowBCKS-WLOvvDwmT6i20ly

Can't upload the picture here...

\documentclass[a4paper]{article}
\usepackage{standalone}
%\documentclass[margin=5mm, crop=true, preview=false]{standalone}
\newenvironment{mysaenv}{}{}%{\hbox{\vbox{}{}}}
%\standaloneenv{mysaenv}

\def\showdata{0}
\def\showspoiler{0}

\usepackage[margin=5mm, paperwidth=300mm, paperheight=90mm]{geometry}


\usepackage{colortbl}
\usepackage{pgfplotstable}
\pgfplotsset{compat=newest}
\pgfplotstableset{string type, col sep=comma, header=false}
\pgfmathsetlengthmacro\MyRule{0.9pt}

\usepackage{multicol}

(freeze frame)

@cis tex doesn't "crash" what error do you get show the log, I just ran it twice with no error getting 103 pages output

@DavidCarlisle Wow, you can compile it correctly?

@cis yes with lualatex, ran without any error

@cis so it ran without error for you and made a 103 page pdf file:

xmf-dist/fonts/type1/public/amsfonts/cm/cmsy10.pfb>
Output written on KoenigstauschRaetsel-NewritePgfplotstableChessboard.pdf (103 p
ages, 35750402 bytes).

@cis Just out of curiosity, are you making these chess problem thing just for fun or is it for some chess club or something?

@mickep This is not really a chess puzzle. But it makes sense to represent the puzzle on a chess board.