I am wondering can someone help with this question: tex.stackexchange.com/questions/615560/…

@Node.JS do it for me drawing questions aren't that popular on this site, you'd get a better response if you showed a sample document attempting that and asked a tex question about the error you got

@DavidCarlisle I was wrong about the xdef, it explodes already earlier in \pdfstringdef. But I don't see how to avoid there that robust commands are expanded, apart from switching to Joseph's code instead:

   \begingroup
      \let\GenericError\@gobblefour
      \let\GenericWarning\@gobbletwo
      \let\GenericInfo\@gobbletwo
      \ifx\nofrenchguillemets\@undefined
      \else
        \nofrenchguillemets
      \fi
      \let\Hy@temp\xdef
      \let\def\HyPsd@DefCommand
      \let\gdef\HyPsd@DefCommand
      \let\edef\HyPsd@DefCommand
      \let\xdef\HyPsd@DefCommand
      \let\futurelet\HyPsd@LetCommand
      \let\let\HyPsd@LetCommand
      \Hy@temp#1{#2}%
    \endgroup

@UlrikeFischer and what do you have against @JosephWright? Or more seriously, if we think we might do that in the medium term we perhaps shouldn't risk making slighly scary changes to \fbox and friends in the format and do a more temporary fix with \protected in xcolor then see how things are needed with expl3 color and string functions

@DavidCarlisle I have nothing against Joseph ;-). Actually it is the plan to use his code, but this won't happen directly, and we still need a concept how people then should define command they want to expand in bookmarks. But regarding the fbox: I have a tendency to say that the user should protect it like other exploding, unusual heading commands. If we start with \fbox we get the next problem with the bookmarks with a \fcolorbox:

\documentclass{report}
\usepackage{xcolor}
\usepackage{hyperref}
\usepackage{etoolbox}
\makeatletter
\robustify\@frameb@x
\begin{document}
\chapter{\colorbox{red}{abc}}
\chapter{\fcolorbox{red}{green}{abc}}
\end{document}

@UlrikeFischer yes the issue certainly isn't just fbox more more or less all box commands. So for now we recommend something like:

\documentclass{report}
\usepackage{xcolor}
\usepackage{hyperref}
\usepackage{etoolbox}
\pdfstringdefDisableCommands{%
\def\colorbox#1#2{#2}%
\def\fcolorbox#1#2#3{[#3]}%
}

\begin{document}
\chapter{\colorbox{red}{abc}}
\chapter{\fcolorbox{red}{green}{abc}}
\end{document}

@PauloCereda most likely they are all siblings.

@UlrikeFischer ?

@UlrikeFischer or without etoolbox and with the original fbox example, suggest something like this in the issue with a note to say we don't at this time want to do this for all commands that might possibly be used in sections so the user needs to fix as required and choose how they want the commands to appear in bookmarks eg I used [] for f(color)box here.

\documentclass{report}
\usepackage{xcolor}
\usepackage{hyperref}

\pdfstringdefDisableCommands{%
\def\fbox#1{[#1]}%
\def\colorbox#1#2{#2}%
\def\fcolorbox#1#2#3{[#3]}%
}

\begin{document}
\chapter{\fbox{abc}}
\chapter{\colorbox{red}{abc}}
\chapter{\fcolorbox{red}{green}{abc}}
\end{document}

@DavidCarlisle yes, I will add something.

@Skillmon :)

Yaaaaaaaaaaay twitter.com/dante_ev/status/1438397908202762244

@StefanKottwitz ooh

In other happy news, IoT has a GitLab gold membership! Yay!

Just got a call from an unknown ungooglable phone number from Cyprus. Had a very short ring from a hidden number earlier today. No, I'm not gonna answer these...

@DavidCarlisle is excessive top spacing when the AMS math display environments (gather, align, multline appear in vertical mode a known issue? Compare foo\par\begin{equation} to foo\par\begin{align} if you're not sure what I'm talking about.

@DonHosek well it's essentially user error so I think sort of ugly by design. equation does have some code to try to hide the issue but align just does $$ and you get a spurious empty paragraph, scary to contemplate changing this as people for example set mathdisplayskip to 0 then rely on the bad spacing to get space above the display, so if you fix that you'd get really tight spacing

Please, is this MWE running for you in TeX Live 2021 LuaLaTeX or XeLaTeX?? It fails for us at Overleaf when you uncomment the three lines. Seems to be some sort of lipsum+polyglossia issue...

\documentclass{report}
\usepackage{lipsum}
% \usepackage{polyglossia}
% \setdefaultlanguage{english}
% \setotherlanguages{latin}  %% In TL2021 if polyglossia is loaded it insists on loading latin
\begin{document}
blah \lipsum[1]
\clearpage
\lipsum[1]
\chapter{Foo}
\lipsum[1]
\end{document}

@yo' fails for me on both xelatex and lualatex (using up to date TL2021)

@yo' I can't test now but there was a discussion with Phelype a few weeks ago about polyglossia problems. Tell the user to use babel.

@UlrikeFischer it's internal testing at the moment

@UlrikeFischer The more entertaining solution is to tell the users to use duckuments instead of lipsum :)

@Rmano Probably not a problem, but it slows down the startup, of course if you use it nothing can be done there, anyway some optimization can be done, for example udisks2.service is closely related to fstab and its configuration, man-db.service is not useful at all (it only updates man pages and can be configured otherwise) and NetworkManager-wait-online.service does not make any sense on a home computer.

@Rmano Anyway, improving the boot time depends on what you use and what you use the computer for, in my case I have a boot as minimal as possible for the use I give it, of course as I've been using fedora for many years I think I already know the list by heart of what I should disable without affecting the boot time of the computer.

@samcarter :)

@DavidCarlisle I bumped into this in my example code where I'm setting source and output side by side. Since I have a handy environment which lets me not repeat the two, it's a bit of a pain. (My namelet question is around simplifying the repetitive task of "fixing" all the AMS math displays for the book).

@UlrikeFischer Thanks for the extra info. Good to know that Phelype is on it, but I'm not sure how much it's gonna help us, we'd like to deploy TL21 soon and this is a potential blocker as some templates use polyglossia+lipsum so they can't be updated to TL21... Anyway, still much appreciated to know what's happenning!

@daleif thanks!

@samcarter dafuq?

@samcarter Having `\` in filenames is a nice technique to stop Windows users from messing with your files :P

@Plergux no idea. I mean spaces and such are already annoying, but complete tex equations? That's a whole new level

@MarcelKrüger :) Interessting interpretation!

@samcarter It boggles the mind! It's probably one of those people who all their files on the desktop. :p

@Plergux This!

@Plergux :D

@MarcelKrüger wouldn't a : suffice? Or have things changed in the Windows world? I remember that all my nice measure-YYYY-MM-DD-HH:MM:SS.dat were disappearing when shared via Dropbox to a Windows-using friend...

@samcarter new Switch game: twitter.com/NintendoAmerica/status/1438533205901127683

@Rmano "colon - sometimes works, but is actually NTFS Alternate Data Streams" ooh

@Rmano I actually don't know, but I wouldn't be surprised.

@PauloCereda LPT1: ;)

@Rmano best is to use " then the file can't be read by TeX on any platform.

@DavidCarlisle :)

@JosephWright ooh

@DavidCarlisle You can make it work as long as you use a reasonably modern engine.

@PauloCereda ohh, birds.

@DavidCarlisle oh wow: ! Unbalanced quotes in argument li"lla.tex indeed

@DonHosek -- Late to the party, I know, but there's a little more explanatory information in the answer to this question: Spurious space above align environment at top of page/minipage

@MarcelKrüger ah brace group versions hmm

@PabloGonzálezL Fedora <3

@Rmano kpathsea assumes all quotes come in pairs to quote filenames with spaces and provides no way to make a literal quote in the filename

@DavidCarlisle yes, I see --- I tried several combinations to no avail. I didn't know... anyway, with the full Unicode available, one can avoid "...

...and it works even with illegal UTF8 sequences, nice: cp ../lilla.tex $(echo -ne '\xC0') can be compiled happily...

@PauloCereda That's 9th circle of hell level stuff for me. I can just about handle the recycle bin in the top left corner, that's it. :þ

@Plergux :)

@JosephWright @DavidCarlisle there is here a bit discussion about the best syntax to set a color as model/value github.com/lvjr/tabularray/issues/106. Do you think that it is reasonable to promote [rgb]{1,0.5,5} as syntax?

@UlrikeFischer yes saw that, I think [] would be better but didnt understand the comment on using {rgb}{1,0,1}

@DavidCarlisle whose comment? From the package author? I think they use braces everywhere and so don't want suddenly brackets.

@UlrikeFischer yes

@UlrikeFischer would color-model=rgb, color=1,1,0 work ie two keys rather than one with a composite value?

Oh... RIP Mr. Sinclair. :-( amp-theguardian-com.cdn.ampproject.org/v/s/amp.theguardian.com/…

@DavidCarlisle Well it would probably work (but the color values would need braces, but I doubt that users would like such an long interface to set e.g. a linkcolor.