@JosephWright I'm wondering if \color_model_new:nnn should give an error or at least a warning, if the pdfmanagement is not found?

@UlrikeFischer But that would rule out using it with dvips, where it doesn't need the PDF management

@JosephWright well the problem is that if you compile this with pdflatex and view in adobe reader, you get a completly white page:

\documentclass{article}

\begin{document}
some text

\ExplSyntaxOn
\color_model_new:nnn{BarTone}{Separation}
  {
    name = BarTone~555~GN ,
    alternative-model = cmyk ,
    alternative-values = { 0.1 , 0.2 , 0.3 , 0.4 }
  }

\color_select:nn{BarTone}{0.5} BarTone
\ExplSyntaxOff

\end{document}

So imho there should be at least a warning in the documentation, that the pdfmanagement is required for the other engines.

@DavidCarlisle do you have an idea why I get sometimes an output like this from your test links? It works on my laptop but not on my PC (learnlatex and the various sites are fine, it seems only to fails with the github.io site).

@UlrikeFischer er no, which browser? (do you have any ad blocker running)

@DavidCarlisle opera. But I'm using it on both computers. And yes there is an (in-built) ad blocker.

@UlrikeFischer is it just that multifile one with the hidden datafile or all of them?

@DavidCarlisle I have seen it also with other examples. But at first I didn't thought much about, only that it is a special layout ;-). Here eg from davidcarlisle.github.io/latexcgi/test2-make4ht

Can some cookie do harm here? Perhaps I played around while you implemented the editor settings ...

@UlrikeFischer that one would be affected when I changed the make4ht settings but a hard refresh to force it to re-load everyting should clear that. I'll download opera and try it. it's a while since I did an opera install.

@DavidCarlisle oh, the editor just reappeared after I changed the zoom on the page.

@DavidCarlisle it doesn't like a zoom of 125%, if I use that and reload the page the editor disappear. When I then change the zoom again it works (also on 125%) but disappears again as soon as I reload.

On learnlatex that zoom is not a problem.

@UlrikeFischer windows opera zoomed 125%

@UlrikeFischer Ah but refresh then fails. OK I can reporoduce:-)

@DavidCarlisle ;-)

quack

<span style="display:none">Breakfast</span>

@PauloCereda will never see my secret plan

@DavidCarlisle oh no

\pdffield_checkbox:n {name=breakfeast,appearance=checkbox/duck,setannotflags=Hidden}

@DavidCarlisle ^^^pdf variant.

@UlrikeFischer oh no

@UlrikeFischer not-my-fault

@DavidCarlisle the opera thing?

@UlrikeFischer yes

@DavidCarlisle any idea who I can blame?

@DavidCarlisle ooh

@DavidCarlisle on golatex and texwelt the editors disappear too at 125% (@StefanKottwitz), only learnlatex resists ... So what it is doing differently?

@UlrikeFischer no idea, but I refer you to my previous answer

@PauloCereda counter-ooh!

@Skillmon hi Mr. Rabbit!

@Skillmon theguardian.com/uk-news/2021/apr/13/… a relative?

@Rmano Someone forgot to add scope to a \HUGE somewhere!

@PauloCereda lol. family dinner! :p

@Plergux ooh

@Plergux that guy would need tons of olives... italymagazine.com/recipe/rabbit-ligurian-style

@Rmano lol, I'd probably just eat the olives :p (mind you, I've never had rabbit so maybe I'd eat it. But from what I gather it's not very "exciting" :p).

oh no

@Plergux it depends. I like it, if correctly cooked. I'm Ligurian, and over there is quite appreciated, and I cook it sometimes at home. It's a very lean meat so it's important not to overcook it and it's not good at all if undercooked, so it's a bit tricky.

@Rmano I grew up eating meat that's some times almost 50% fat and either best a little undercooked or slow roasted to a mush, so accuracy is not required so much. Maybe that's why I can't cook XD

@UlrikeFischer try davidcarlisle.github.io/latexcgi/test-multifile

@Skillmon your guy is at a new level of "do it for me"-ing ... do it form me exactly as I like ;-P tex.stackexchange.com/q/592836/38080

@Plergux ^^ siunitx v3 progress from yesterday

@Rmano I don't know which boggles the mind more, the unashamed brass of it or the complete inability to make an effort.

@PauloCereda lol.

@PauloCereda good, no more bug-reporting magpies

@Rmano ooh pie

@Rmano Sometimes I ask myself why people do this. What kind of homework requires you to simply reproduce a schematic?

@Skillmon perhaps the question was to design the crcuit and a picture of the answer can be found on the internet that you therefore need to re-create?

@PauloCereda rabbit and duck?

@DavidCarlisle that works fine. What did you change?

$ git diff
diff --git a/docs/runlatex.js b/docs/runlatex.js
index 257a91f..adec129 100644
--- a/docs/runlatex.js
+++ b/docs/runlatex.js
@@ -164,6 +164,7 @@ function llexamples() {
                }
            }
            p[i].textContent=pretext.replace(/\s+$/,'');
+           p[i].style.height="1em"; // force redisplay in Opera zoom
            ace.config.set('basePath', 'https://cdnjs.cloudflare.com/ajax/libs/ace/1.4.12') ;
            editor = ace.edit(p[i]);
            editor.setTheme(rlacetheme);

@UlrikeFischer I lied so forced Opera to correct me.

@DavidCarlisle ;-).

@DavidCarlisle oh no

@Skillmon it's really puzzling. But we had similar ones in the past with pgfplots, it seems that the teacher changed package fancies...

@samcarter_looks_forward_TUG'21 @UlrikeFischer I don't think I want to join another site just to comment but on texlive.net questions: Joulev asks about asy that could be added just as makeindex/biber/bibtex are (it is already installed on the server, which has a full texlive 2021) running examples with local files is possible (eg there is a picture of shakespere and some tex4ht configurations)...

... but we (I) don't have the bandwith to handle and vet uploads so any files need to be submitted with the form or already in texlive. The mult-file support is easy enough if you have full control over the html or markdown as at learnlatex, less so in forum sites where the end user only controls individual divs, but interface suggestions welcome. (issue at github.com/davidcarlisle/latexcgi/issues is best)

@DavidCarlisle Thanks a lot for your comment! I'll pass the message on

Good afternoon everybody....

Is there someone that could help this user, please? Thank you very much.

@Rmano now he asks me to move stuff. The answer is very simple, but I'm not sure whether I should give it to him...

@Skillmon yeah... the mental laziness is really astounding. It's clear that "teach 'em fishing" is not going with them.

@Rmano at least they should now know all possible ways how to add a code block to stackexchange, because I can't think of any other reason for this edit tex.stackexchange.com/posts/592836/revisions :)

@Rmano give them a picture mode answer

@DavidCarlisle will the picture mode show a fish?

@DavidCarlisle is there a circuit package for picture?

@Skillmon tikz needs ever more wrapper packages to hide the underlying syntax, picture mode's elegant simplicity needs no such wrappings.

@samcarter_looks_forward_TUG'21 we have ducks donkeys and pigs already on site, a fish could be added if needed

@DavidCarlisle ooh

@samcarter_looks_forward_TUG'21 we sure they're not bots? Somebody measuring our response? ;-P

@DavidCarlisle :D

@Rmano Let's hope they are, that would be rather impressive work :)

@Rmano I thought my description of what he should do was relatively clear... (tex.stackexchange.com/questions/592781/…)

And now I told him anyway.

@Skillmon aaaaaand... another one (with an interesting answer) tex.stackexchange.com/q/592874/38080

@Rmano it's funny how every now and then we're flooded by this.

@Skillmon first time one of them ask for private connection, though. Let's see if they're able to find my email and not ending in the spam folder... tex.stackexchange.com/questions/592836/…

@Rmano It would be a great service to humanity if someone could tell this user that the OP is automatically notified and it is not necessary to announce every single answer with a comment...

@samcarter_looks_forward_TUG'21 :-) ah, you also noticed it? I've been tempted to flag them...

Spelling poll: if you have used TikZ [overlay] to place a picture, is the picture "overlayed" or "overlaid"? I've used both in the same answer so clearly I don't have a clue.

@AlanMunn y (I think)

Looking forward to getting a feedback from MathJax experts: chat.stackexchange.com/transcript/message/57650810#57650810

@DavidCarlisle Yes, I think that's my intuition too, and my prescriptive self got the better of me with the 'i' version.

@raf looks OK to me

@AlanMunn Shouldn't it follow the same rule as "lay" -> "laid"? (as in "lay down" -> "laid down", for example)

@PhelypeOleinik Ah, well there's the rub. @DavidCarlisle and I have a different intuition, and there's likely a good linguistic explanation for why, but I'll wait to see what other native speakers say before revealing it.

@AlanMunn Oh, a good suspense :)

@AlanMunn ooh a secret

@AlanMunn I would say, ;) , the overlayed picture is on top of, the overlaid one below something else.

And the [overlay] TikZpicture is on the top.

@AlexG Hmm. That's a lot of semantic work for a spelling difference. :)

@DavidCarlisle ok. Thank you. Though I am having a little issue with the output of minus signs : chat.stackexchange.com/transcript/message/57651126#57651126

@raf you would see same in tex $---$ is spaced like $x-y if you want the middle one not to be spaced as a binary op use $-{-}-$

Is this expected behaviour for resume in enumitem?

\documentclass{article}
\usepackage{enumitem}
\newenvironment{foo}
{}{}
\begin{document}
\begin{enumerate}
\item Item one
\item Item two
\end{enumerate}

This is some more text.

\begin{foo}

\begin{enumerate}[resume]
\item This is item three
\item This is item four
\end{enumerate}
\end{foo}

\begin{enumerate}[resume]
\item This should be item 5
\end{enumerate}

\end{document}

I think it has come up before that it uses local settings, I should run your example I guess...

@DavidCarlisle Inside the environment, the resume works, but afterwards not.

@AlanMunn well it works but it resumes from a locally stored restart value which has been reset at the end foo

@AlanMunn I would classify "overlayed" between "overlaying" (fully active) and "overlaid" (fully passive). While an "overlayed" picture is a passive object (placed by the document author), it still has a somewhat active role by itself overlaying something else. But this is only my personal and perhaps wrong feeling.

@DavidCarlisle That's a shame. I have a supposedly drop-in replacement for gb4e using enumitem lists, but this breaks it badly if it can't keep a global counter independent of environments.

@AlanMunn could patch in a forced global but the code seems to have some support for global settings in some code paths \enit@setresumekeys{series@\enit@series}\global\global but there are lots of code oaths and not sure I want to look now:-)

@DavidCarlisle Maybe I'll post an issue on Javier's repo. Although there's been no response to my last issue on titleps, so it might take a while.

@AlanMunn you should setup a series:

\documentclass{article}
\usepackage{enumitem}
\newenvironment{foo}
{}{}
\begin{document}
\begin{enumerate}[series=blub]
\item Item one
\item Item two
\end{enumerate}

This is some more text.

\begin{foo}

\begin{enumerate}[resume*=blub]
\item This is item three
\item This is item four
\end{enumerate}
\end{foo}

\begin{enumerate}[resume*=blub]
\item This should be item 5
\end{enumerate}

\end{document}

@UlrikeFischer Perfect! Thanks a lot.

@AlanMunn the * is not needed (it resets settings too)

@UlrikeFischer Yes, thanks.

@DavidCarlisle looking in the docu is a bit faster here ;-)

@UlrikeFischer :) I was actually aware of the series use, but hadn't realized that it was required for this sort of global use. But the behaviour of [resume] in my example is kind of weird really. I could understand if the counter reset inside an environment, but it doesn't but then reverts back outside it.

@UlrikeFischer there was no documentation in the file I looked at

@UlrikeFischer I pushed that height=1em change to learnlatex.org I did a quick check and I don't think I broke anything but can you keep a lookout

Is, by chance, someone running Arch Linux here?

@AlexG Yes.

@TeXnician OP is running Arch. I you have some spare time, would you please test the example?

@AlexG Hm, if I run the document and then click on one of the links in the example my Okular exits with a SIGSEV.

@TeXnician This is what OP reports too :\ But I am running the same (stable) package versions on my Gentoo box and I don't have problems with this example.

@UlrikeFischer Do you know if there's a way to get the effect of \counterwithin for a series counter?

@TeXnician But thanks, ;)

@UlrikeFischer I guess only using \restartlist and patching the relevant document division.

@AlanMunn imho yes, I don't think that counters are used that can be reset.

@AlanMunn -- You didn't include "overlain". "Overlaid" would be proper (US) English, but I think @AlexG has got the right idea. A genuine conundrum.

@AlexG it works in Evince in my Arch...

@UlrikeFischer Unfortunately it seems that \restartlist doesn't work with series:

\documentclass{article}
\usepackage{enumitem}
\newlist{exx}{enumerate}{1}
\setlist*[exx]{resume=exx,label=(\arabic*)}
\begin{document}
\section{A section}

\begin{exx}
\item An item
\item An item
\end{exx}

\section{A section}\restartlist{exx}

\begin{exx}
\item An item
\item An item
\end{exx}


\end{document}

@barbarabeeton So here's what's behind my intuition, but I will admit it's a bit of a stretch. Assumptions: 1) irregular forms are instances of root allomorphy, and as such can only affect the root. 2) irregular spelling indicates an irregular form of the verb. 3) Words formed with that root should not show irregular morphology if the morpheme attaches to the derived word rather than the root itself.

@barbarabeeton Examples: the past tense of the verb 'grandstand' is 'grandstanded' not 'grandstood'. The plural of 'leadfoot' is 'leadfoots' not 'leadfeet'. etc. So 'overlay' is similar for me: it's a compound formed as 'over+lay' but then the participle or tense morphology must be regular, because it's attaching to the derived word and not the root, hence the regular spelling.

@barbarabeeton The stretch part is assumption 2) which I actually think is right, but is a kind of strange interaction between morphology and spelling not mediated by pronunciation. Can you think of any other cases like that?

@AlanMunn seems reasonable but as far as I can see with free access searches of cambridge and oxford dictionaries, overlaid is a word and overlayed isn't which I found a bit surprising I thought it would be at least listed as a variant spelling

@DavidCarlisle Yes, I suspected that might be so. Because although our intuitions on 'leadfeet' etc. are clear, it's much more subtle to have that extend to just a spelling distinction alone. So it's not too surprising that only the 'i' variant is listed.

@AlanMunn you can do \makeatletter\gdef\enit@resume@series@exx{\c@exxi 0}\makeatother but I don't see a key for it.

@UlrikeFischer Wow, thanks for that. It would have taken me a lot of time to get to that I think. It seems like a useful feature request for Javier then.

@UlrikeFischer I assume you found that by reading the documentation too?

@DavidCarlisle sure, isn't that the file with all the @?

\@@par