a theorem inside an enumeration gets indented according to the current left margin. Is there an easy way to have it start at the original left margin?

@MarianoSuárez-Álvarez interrupt the enumeration and resume it after the theorem?

I thought of that, but after the theorem there not a new \item, so ti does not work without extra incantations!

@Skillmon I can report that luacode makes the 13 minute code immediately

Sorry, run immediately. Literally, it went from taking ages to make some pictures, to I can do significantly more complex ones in a few seconds

@MarianoSuárez-Álvarez you can always force something but what is a theorem doing inside an enumerate???

@Jasper yep, that was to be expected. Parsing maths expressions in TeX is slow.

@PauloCereda Oh, ducks science.org/content/article/…

@samcarter I happened to go to tex.stackexchange.com/q/736616/52406 just a few seconds after your second comment. 23 seconds inbetween your comments. You are quick!

@mickep duck speed :P

I just read this answer: tex.stackexchange.com/a/736629/47927 on my phone, so I did not see the author. But when I read the fourth paragraph, I knew …

Oh, teh author edited their answer already :)

@Skillmon I will be converting fully to luacode then for most things. It has the data type table too which is nice. Why do people even use pgfmath anymore? Heck, that sorting algorithm you build for me (which was awesome) can be done more easily in luacode as well. lua is easy to learn too, I learned it well enough in a day (to my surprise) to make vector fields.

@JasperHabicht brilliant literary style always shines through in a recognisable way?

@Jasper because that only works in LuaTeX, and some people still want to use different engines for different reasons (I myself mostly use pdfTeX, but I don't do many calculations). Also for historic reasons.

@DavidCarlisle your orthography is like your pronunciation, unrecognisable to some, distinct for others.

@Skillmon ooh pronunciation, try the new pronunciation package, that makes extensive use of one of my better packages.

I observed some strange behaviour of TeX.SE. When I go to the questions with activity tab, it very often shows me questions from years ago. When I check for any comments in 2025 or even bot edits, I see nothing. Why? Do upvotes count as activity?

@MaestroGlanz sometimes the system bounces questions to the front page to encourage activity on them. (Personally I never use that activity view, I always just use the view in date order)

@MaestroGlanz Questions bumped by the community user usually have an answer with 0 votes. If you either up- or downvote the answer, the question won't get bumped again.

New version of easybook on TL!

@egreg about time!

@DavidCarlisle do you want me to remove your name from ebgaramond-maths? I prefer to keep track of why I used code, but I'll delete it if you object.

@DavidCarlisle =3 indeed!

@Skillmon It does not. It only loads ifthen, xparse and tipa if I am not mistaken. It is, in fact, a huge list of words (which might have been better put in some external file maybe for easier maintenance)

@cfr no it's fine (and true) I was just joking. i just to see what the file did and didn't expect to see my name:-)

@JasperHabicht the best part of it is that handling a list of 20 thousand words is very quick as the package gives errors on all use so your job always finishes in a very small time

\documentclass{article}

\usepackage{pronunciation}

\begin{document}

\end{document}

@DavidCarlisle so I suspected, but sometimes people joke about things they mean ....

is it possible to redirect output in a cross-platform way?

@cfr foo > bar.txt does more or less work in windows and *ix shells.

@cfr or if calling from Lua rather than a shell you can grab the output into a lua string then write it anywhere

@DavidCarlisle Let's wait for version 1.0.3

@JasperHabicht I have emailed asking for that

In expl3 code there is this convention for longer braced stuff to put the left brace on a new line with indentation, then the stuff inbetween on the next line with more indentation. Why not just put the opening brace at the end of the line above and remove one level of indentation? Are there certain situations where this is a better choice?

(I ask because I did not adhere to this convention in my code up to now ... I don't want to appear rebellious or change conventions, but I don't think that I can read code better this way ...)

@JasperHabicht as long as you are consistent in your own packages, it dooesn't matter much but I think, looking at some examples in latex.ltx...

  \int_compare:nNnT \tex_lastnodetype:D < 0
    {
      \box_if_vertical:NT \l__mark_box
        { \vbox_unpack:N \l__mark_box }
    }

@DavidCarlisle Ah, I was looking for examples which don't adhere to this.

@DavidCarlisle But it still puts the opening brace in a new line (in line 2). Well, yes I do keep a consistent style of course, I think, it more or less differs in only this single point

That shows the form you describe and has the possible advantage that the argument is always on the next line whether (like the inner one) it uses the compact form { \vbox_unpack:N \l__mark_box } or (like the outer one) it uses the spacy form

but 5 lines later is a different style

   \__mark_debug:n { \iow_term:x
      { Marks:~ mark~ extraction~needs~ recursion~
        \msg_line_context: } }

@DavidCarlisle True. But I think the "official" expl3 syntax allows for both.

where the multiline argument starts on the same line as the opening { and the closing } is on the same line.

@DavidCarlisle Oh, okay ... hm

:67168240 you could always check with git blame who to blame if you copy that style:-)

Sorry. No, I won't copy this style =D

\__mark_debug:n {
    \iow_term:x {
        Marks: ~ mark ~ extraction ~ needs ~
        recursion ~ \msg_line_context:
    }
}

This is how I would probably style it

Well, whatever ... I just wondered where this "single brace on new line" comes from

@JasperHabicht I think the l3 style is arguably clearer as it aligns opening and closing braces if they are not on the same line. Although I still like

\def\@ltab{\@ifatmargin\ifnum\@curtabmar >\@firsttab
      \global\advance\@curtab \m@ne \global\advance\@curtabmar\m@ne\else
      \@badtab\fi\else
      \@latex@error{\string\<\space in mid line}\@ehd\fi\ignorespaces}

where it is so clear where the if.. starts and where it ends and what code is in each branch....

@DavidCarlisle True! Well, for TF constructs, I often have a line with just } {. And yes, the opening brace is "hidden" at the end of the previous line, but the indentation shows that there was an opening brace as well

@DavidCarlisle I think you are at something here.

@DavidCarlisle I would need >> to work. I can't figure out how to catch it into a lua string.

@DavidCarlisle I know it's not ideal. This is a long text explaining the results of a paper, and using a list for indentation of items really helps with making the structure obvious, but the results are theorems so… I've tried several different ways of reorganizing the whole thing and converged to this.

@yo' @PauloCereda not my fault....

@MarianoSuárez-Álvarez Start the new enumeration with an unnumbered item

@cfr stackoverflow.com/a/60478584/1158383 (although I thought there was something simpler)

@cfr ah thi si think is what i had in mind stackoverflow.com/a/60609341/1158383

@DavidCarlisle unless listings changed something you need the processletter-stuff tex.stackexchange.com/a/25396/2388

@DavidCarlisle Thanks -- I commented

@UlrikeFischer yes once I got an example I tried without:-)

@yo' your hint about the editor was key, the OP can upload the isabelle output (which he doesn't want to edit anyway) and input that into the document.

@JasperHabicht I prefer it for expl3 code. also, it's easy to configure vim. but I don't use indentation for the braces ....

@DavidCarlisle Exactly!

@DavidCarlisle I cannot get it to work, but thanks.

@cfr I have some code somewhere doing that (but can't remember where) If I find it I'll ping you

@JasperHabicht that was a joke

@egreg I'll do the AllTheDiff tomorrow

@Skillmon I know, it still made me curious

@DavidCarlisle thanks. I think I need something like io.popen...

but that is not cross-platform, it seems ...

@JasperHabicht no, the official L3 style guide only allows the style with arguments on new lines and further indentation, see l3styleguide.pdf

@DavidCarlisle we should git blame that and tell the author off :)

@Skillmon but that is for latex contributors, isn't it? there is no expectation contrib packages will follow, is there?

@JasperHabicht I prefer it when brace pairs are in the same column.

@cfr yes, you are quite free to do what you want.

@cfr that's what i was looking for:-)

@cfr I even commented on it, l3build/l3build-uload.lua -- use popen not execute so get the return body local exit_status=os.execute(ctan_post .. "validate")

@DavidCarlisle I guess I got what i deserved for being fair in what I write :-)

@yo' I'd be even more annoyed if I couldn't use my name, but I assume there are ediitor issues that are causing this in the background.

@DavidCarlisle The Editor would be fine, but users are users -- we have to aim at reasonable use-cases and make sure that if a user copy-pastes an actual binary file (don't ask why), things do not explode. So IIRC we only accept the basic unicode plane stuff in editable files.

@yo' ah OK, but that does cut out a chunk of Asia as well as large parts of mathematics, as people start using Unicode characters as input with unicode-math this is going to hit hard as it's not just "exotic" characters it is the standard math italic that will be removed.

@DavidCarlisle Indeed -- we will have to adapt at some moment.