@AlanMunn There's a couple before I discovered forest. Since then, yes, pretty much ;).

I just find it annoying when people don't explain downvotes. Actually, I find it even more annoying on other people's posts than on my own.

@cfr Yes, I agree. I've had downvotes on answers many years after they were posted and sometimes when there's no competing answer to the question. But in the grand scheme of things, it's really irrelevant except for putting you off the palindromic rep track.

@AlanMunn I find it particularly awkward on other people's posts because, if I comment (especially if I make a criticism), the assumption will be I downvoted.

@DavidCarlisle Looking at the question again. How did you know? :)

@Skillmon square = @(x) x.^2; is the Matlab version. They’re useful but a bit annoying because you can only have one “line” in the expression (so no loops or conditionals, or intermediate calculations, say). There isn’t a lightweight multiline version like you find in Lua

@Skillmon Worth adding a new data type to try it out? :-) I agree if we could rewrite history that would likely be a good change to make

@WillRobertson yes, but recent enough Matlab versions support defining additional functions at the bottom of the script without adding more files...

@WillRobertson as @JosephWright put it, a thought experiment, but nothing that should be implememted now. But it would fit the idea of total control of expansion. One could even put a marker at the start of the variables throwing an error of misused variable if they somehow end up in the typeout routine, similar to the way it is currently done for sequences.

@JosephWright would it be easy to lose the surrounding environment when tagging the math (I haven't looked in detail) tex.stackexchange.com/questions/697295/… (or @UlrikeFischer if you are not really having a holiday:-)

@DavidCarlisle I will look.

@UlrikeFischer thanks I had a quick attempt to modify \grabaformulapartandstart from latex-lab-math(tagging).ltx but I think i got the timing wrong as nothing changed, and I had lunch instead:-)

@DavidCarlisle the command to change is \@kernel@math@begin, there the text for actualtext is built. But one should think a bit about a better interface.

@DavidCarlisle but more worrying than the backslash is that the ^ is not read, or do you hear it?

@UlrikeFischer is it reading the actual text or /EmbeddedFile ?

@DavidCarlisle the actual text (I tested with blub ;-)), acrobat can't handle pdf 2.0 features like associated files yet.

@UlrikeFischer it reads it as "caret" (@Skillmon would hear it as carrot)

@egreg deleted the comment under tex.stackexchange.com/a/697274/117050

@DavidCarlisle hm, tasty symbols!

@DavidCarlisle in german I hear nothing, it says "unterstrich t j"

@UlrikeFischer I'll fix both, update coming up..

@DavidCarlisle öh where?

@UlrikeFischer answer on site (just listening to acrobat before posting)

@JosephWright An introduction to LaTeX3 programming would great to have. In the short term, however, it might be useful to regorganize and rename the current documentation. With all of the functionality now built into the kernel, I think it would make sense to combine the interface3 and xparse documentation and give it a better name.

@AlanMunn I agree the interface3 name isn't great but all the necessary xparse doc is now in texdoc usrguide

@DavidCarlisle yes just saw it. probably one will have to split the commands for the actual text and the one for the embedded file and some time. And in any case this long text even for inline math isn't good anyway. But I'm not sure if we want lots of regex.

@UlrikeFischer performance won't be great with regex and you could use a custom delimited argument to replace _ by sub (or subscript or whatever) but it shows the idea...

@DavidCarlisle yes.

@UlrikeFischer anyway who cares what it sounds like, important thing is I get a badge for answering?

@DavidCarlisle Hmm. usrguide is about as generic a name as one could fine. Is that the successor to the old clsguide?

@DavidCarlisle a tagpdf badge?

@AlanMunn usrguide and clsguide have been available from the start of 2e, but we (Joseph) recently updated both of them to match recent additions

@UlrikeFischer great answer?

Hello everyone. I am writing my thesis in LaTex in VSCode on Windows 10 and I'm using Arara. Arara worked perfectly fine until yesterday when I've upgraded the package to version 7.1.1 and then I installed Open JDK 21 (after uninstalling Java RE). Now Arara doesn't work anymore. Please see my question for more info tex.stackexchange.com/questions/697292/…

I don't think I've ever even heard of usrguide. :) Which may say something about me, or something about the advertising campaign...

@Giuseppe everyone here will have seen that already. @PauloCereda is usually around, but he is allowed a weekend (sometimes)

or @TeXnician (but it's more traditional to blame @PauloCereda)

@Giuseppe gitlab.com/islandoftex/arara/-/issues/119, arara on Windows is broken. And the arara core devs have no Windows development setup to reproduce these things easily. arara 7.1.1 unbreaks things on Linux and mac but breaks on Windows it seems. Unfortunately, we cannot tell MikTeX/TL to roll back for one platform only, so we can only suggest to use 7.1.0 for now until we get someone with a Windows JVM development environment to help us debug things.

@TeXnician I suppose I could admit I have a windows jvm or two on this machine:-)

@DavidCarlisle This seems like an issue in JAR production where a few classes are omitted from the JAR which Windows requires for native access but the other OSs don't. So we would need someone to actually debug on Windows to identify where in our process things get eliminated too early, including a few rebuilds and trial and error. So in addition to the Windows JVM we need some JVM dev experience. We could also call it the “revenge of noone answering our calls for joining arara development” ;)

My main suspect right now is this part which was introduced by @PauloCereda and about which I know nearly nothing. So blaming the duck seems fine right now :)

@TeXnician yes not sure how much time I have and my Java is a bit rusty but I have compiled and packaged jar before (of my own java, not just typing make somewhere:-) so in principle I could probably help

@TeXnician I have lots of experience in blaming ducks

@DavidCarlisle Thanks for the offer, I'll check with @PauloCereda if he has some ideas what to change and maybe we'll come back to you for testing a few variations :)

@AlanMunn Not just you. Until a few weeks ago, I'd never heard of clsguide or usrguide.

@AlanMunn and @cfr: seems like I'm doing something wrong, because I knew of both, but I guess I'm not as long in the TeX business as you two...

In case you didn't know, there is also fntguide which can turn out as an interesting read

(though more for non-Unicode engines)

@AlanMunn I second this. An introductory guide would be good, but a single document with pointers to things would be fantastic. I'm thinking of trying to make an index for myself just to try to keep track of where things are and what there is. It will inevitably be continually out-of-date, but it might be something.

@Skillmon That I did know of. It might also depend how you learnt TeX in the first place. I've never seen a copy of Lamport's book since looking at it before I started learning in a bookshop. (I'm guessing that was the other book in the university bookshop. There were two. One I found unintelligible. The other was entirely intelligible, so I bought that one.) MacTeX also offered some of its own resources and I was focused initially on a class based on amsbook, though I don't remember any docs.

@cfr you could make it available and into a coordinated community effort to provide such an index.

It's just possible the other book was the TeX Book, but I'm pretty sure it had LaTeX in the title, so I'm guessing it was Lamport's. After a while, I bought The LaTeX Companion, but I didn't find it terribly useful.

@cfr I must admit I don't own any book about (La)TeX. I read parts of the TeXbook as texbook.tex...

@Skillmon My internet access at home was a dial-in modem pool which provided a non-graphical interface to a unix-ish system (I forget which) and lynx. It was slow and very fragile. I had proper access on campus or I couldn't have downloaded the software, but a book of some kind was invaluable.

@cfr Well, I learned (La)TeX at most 10 years ago...

@Skillmon QED

Is it expected that using \SetHorizontalCoffin in the preamble will give an error?

@cfr probably even less... I started studying 11 years ago, and I think I learned TeX 2 or 2 and a half years later... My first "real" documents in LaTeX were notes for an exam I and a few friends took.

Thinking about it, I can't really believe that all this is already this long ago...

@Skillmon My first document was my PhD thesis. Word kept throwing my work away.

I'm not sure if I exactly learnt LaTeX. I learnt enough LaTeX, but only enough.

@cfr some of use learned latex before the uk was on the internet

@cfr any typesetting in the preamble is risky as fonts (and things) are only fully set up at begin document

@DavidCarlisle I wasn't in the UK.

@AlanMunn -- Your comment has led me to see if I can find a reference to these "core" documents (which are all in the main LaTeX package on CTAN), and I didn't find anything obvious by following the "obvious" links on latex-project.org (@JosephWright and @DavidCarlisle please take note). Adding a suitable pointer on the home page (maybe titled "Documentation for package writers") would be useful, I think.

@cfr so you had a few extra years of internet access I guess:)

@DavidCarlisle So it is expected? I just wasn't sure.

@cfr expected may be too strong but not surprising (also Im not sure I know SetHorizontalCoffin is that a renamed \hcoffin_set:Nn ?

@DavidCarlisle Actually, the UK was on the internet before I left. Technically, we all had email addresses at university. Just only one person ever read it, so the contents were transmitted via word-of-mouth.

@cfr we had email but janet not internet:-)

@DavidCarlisle What's janet?

@cfr joint academic network and email addresses were in the correct order [email protected]

@DavidCarlisle Maybe that's what we had. Since I never used it, I never knew.

\NewDocumentCommand \SetHorizontalCoffin { m +m }
  { \hcoffin_set:Nn #1 {#2} }

@cfr my crystal ball is working, I see

[From the .sty in l3experimental.)

@cfr ah, I never look at team code, very unreliable

@DavidCarlisle :)

@cfr oh, a half-closed interval :)

@DavidCarlisle It's just that naïvely, I'd expect it to work a little bit like saving something in a box, but it doesn't. I complains \__color_backend_reset: is undefined. If you try to define it first, it complains \__color_backend_reset: is already defined.

@cfr -- Maybe "Guide to LaTeX" by Helmut Kopka and Patrick Daly. An excellent intro and reference manual, although now sadly a bit long in the tooth. I do wish that Patrick Daly had been willing to update it after Helmut Kopka's death.

@cfr backend code is set up at the end of the preamble but you can force it to be loaded earlier

@barbarabeeton Definitely not that one. That was the other one, which I bought. So it is the one book I know it wasn't.

@cfr mwe?

@cfr better to set it in a hook so it runs after begin document

\documentclass{book}%
\usepackage{xcoffins}

\NewCoffin \PagenumberCoffin

\ExplSyntaxOn
% \cs_new_protected_nopar:Nn \__color_backend_reset: {}% error as \__color_backend_reset: is already defined
\ExplSyntaxOff

\SetHorizontalCoffin \PagenumberCoffin {somewhat}% error as \__color_backend_reset:n is undefined

\begin{document}

\paragraph{A subsubsubheading} rather than a paragraph.

This is a new paragraph.

Super Text is in another new paragraph.

\end{document}

\documentclass{book}%
\usepackage{xcoffins}

\NewCoffin \PagenumberCoffin



\AddToHook{begindocument/after}{%
\SetHorizontalCoffin \PagenumberCoffin {somewhat}% error as \__color_backend_reset:n is undefined
}
\begin{document}

\paragraph{A subsubsubheading} rather than a paragraph.

This is a new paragraph.

Super Text is in another new paragraph.

\end{document}

@cfr ^

@DavidCarlisle I'm more curious. In the actual example, it turned out not to be useful to set it in advance at all. (It might be more efficient, but maybe not.)

I just wondered why it thinks \__color_backend_reset: is both undefined and defined already?

@cfr you definined it too early so when the backend code loaded it gave an error. what you should do instead is load the backend earlier (for which there is a specific command I forget at this moment)

@cfr \sys_ensure_backend:

@DavidCarlisle Thanks. I'm being an idiot.

@cfr better than being a duck

@DavidCarlisle Thanks.

@DavidCarlisle I might be both for all you know.

@cfr we could use your pot for duck stew

@DavidCarlisle It's a cauldron.

@cfr we could use your cauldron to magic up some duck stew

@DavidCarlisle No you couldn't. It's vegan.

Isbegindocument/after better than \AtBeginDocument? Is the latter too early?

@DavidCarlisle

@cfr well it's later so safer (we added the new one as it's after all the typesetting setup almost as if you had the code really after begin document), whereas \AtBeginDocument happens sort of in the middle of the setup

@cfr we could use magic to produce duck stew without killing any ducks?!

@Skillmon And without killing any other sentient beings or using animal products? I don't have that spell, but you could use this new email thing?

@cfr wait, no other animal products? Not even honey or milk?

@Skillmon A strictly vegan cauldron wouldn't use honey. I'm not strictly vegan because I eat honey. But definitely no milk.

@cfr I like meat. I can survive on a vegetarian diet, and don't eat much meat usually. But vegan (even if not strictly vegan) is too much for me... :P

@Skillmon I'm not trying to convert you ;). And I use other things which a strict vegan wouldn't. (E.g. leather shoes.) It's a compromise. The closest I come to attempting to convert people is making vegan desserts they don't think can be vegan. (But there is a lot more vegan stuff around now than when I started doing that, so I suspect what people believe could be vegan has shifted. Though people still seem astonished chocolate can be vegan, even though the best stuff always was.)

@Skillmon I'm not letting @DavidCarlisle use my cauldron for unethical purposes, though. He can make his own with picture!

@cfr oh, don't worry, I wasn't feeling being "converted" :) And yes, @DavidCarlisle should really create a picture mode cauldron!

@DavidCarlisle In this case \AtBeginDocument works but neither begindocument/begin nor begindocument/after.

@Skillmon OK, good. It's just some people are quite self-righteous about it and I think it's both ethically dubious and counterproductive.

\documentclass{article}
\usepackage{color}
\begin{document}

\begin{picture}(100,100)
\put(50,50){\oval(50,20){}}
\put(70,65){\circle{20}}
\put(35,50){\line(1,0){30}}
\put(70,65){\circle*{2}}
\put(75,67){\line(6,-1){10}}
\put(75,63){\line(6,1){10}}

\thicklines
\put(50,57){\oval(70,60)[b]}
\put(5,57){\line(1,0){10}}
\put(85,57){\line(1,0){10}}
\end{picture}

\end{document}