TeX, LaTeX and Friends

David Purton

01:45

@DavidCarlisle Ha! I found it. Not related to fonts at all. Some strange interaction with one of my hooks and the RTL language. Explicitly clearing the hook before exiting the group (and language settings reset) fixes things. I was clearing the hook at the start of the main environment, but somehow this wasn't enough. That's too hard for me to track down exactly what was going on.

Paulo Cereda

08:42

ooh a secret CTAN mirror selector (@MarcelKrüger)

David Carlisle

09:16

davidcarlisle.github.io/latexcgi/…

@JosephWright ^

Joseph Wright

@DavidCarlisle :)

mickep

09:27

@yo' Maybe something for you (Overleaf is mentioned...): tex.stackexchange.com/q/720462/52406

Ulrike Fischer

10:23

@JosephWright did we released latex-dev with the hotfix too?

David Carlisle

11:02

@UlrikeFischer a sorry it's all my fault answer:-)

Joseph Wright

@UlrikeFischer No, it's a bit laborious to do that, and it's also not really needed I think

Ulrike Fischer

@DavidCarlisle ?

David Carlisle

@UlrikeFischer lang=

Ulrike Fischer

@JosephWright the problem is that the finalstrut breaks tagging github.com/latex3/tagging-project/issues/81, which means we can not recommend to use latex-dev anymore ...

David Carlisle

@JosephWright trouble is we mostly tell people trying tagging to use lualatex-dev (even just after a release when it's essentially same as lualatex) so now most tables won't work

Ulrike Fischer

11:17

@DavidCarlisle ah yes ;-)

David Carlisle

@UlrikeFischer snap

Joseph Wright

@UlrikeFischer, @DavidCarlisle I see the point but I also see it's dev code, so there's always the 'ah well, currently bust, use the release version' approach

David Carlisle

@JosephWright I think we should release latex-dev/base we are just about getting people to look at table tagging (or existing arxiv dcuments) and almost everywhere we say use lualatex-dev as that has latest fixes, as most of the time historically lualatex woul dnot work

Joseph Wright

@UlrikeFischer, @DavidCarlisle The admin issue is we call the first pre-release '0', so if I call this 'PR1' with no new features, it's a bit odd ...

David Carlisle

@JosephWright feature is bug fix compatibility with main release

Joseph Wright

11:29

@DavidCarlisle OK, I'll sort it

David Carlisle

@JosephWright without tagging, tabularx, tabulary, colortbl, nicetabular all break, with tagging, all tables break, so the current lualatex-dev is not really usable

@JosephWright thanks

yo'

@mickep Thanks for pointing this out, but it seems it's on Overleaf Community Edition, which is self-hosted hence out of any our control. We can't help with that really, even if we wanted :-( (and for that reason, we do not officially provide support services for that).

But I wouldn't be against having a tag overleaf-self-hosted that we would watch. (We in general watch for issues, in case we actually messed up. But this seems like an inconsistent texlive update...

Joseph Wright

@DavidCarlisle Building now - takes about an hour

daleif

12:05

Below is some code to be used in a rewrite of my private rewrite of fancyref. It seems to be working ok. Two questions (1) Are the naming conventions correct? (2) Are there better ways to make this checker macro? Splitting into a seq for further processing seems to be a lot more flexible than what I had been using.

\documentclass[a4paper]{article}
\usepackage[T1]{fontenc}
\usepackage{array}

% also works under french babel
%\usepackage[french]{babel}

% and mathtoolc with centercolor
%\usepackage{mathtools}
%\mathtoolsset{centercolon}

\ExplSyntaxOn

\seq_new:N    \l__daleif_list_seq
\tl_new:N     \l__daleif_tmp_tl
% this holds the status of the label parsing, 0 = ok, otherwise:
% 1 = blank label, 2 = no :, 3 = empty prefix, 4= seq length = 2 and right is empty
\int_new:N    \l__daleif_badness_level

% The macro is being used in a feature like the fancyref package. This

The code should explain what is going on

mickep

@yo' Oh, so they only borrowed the name... Sorry for the ping.

yo'

@mickep No, not really, it's legit. They install this: github.com/overleaf/overleaf

Ulrike Fischer

@daleif sorry don't have the time to really look at the code, but the int should end with _int.

daleif

12:31

@UlrikeFischer right, fixing in own version

Skillmon

@daleif (removed)

David Carlisle

13:07

@JosephWright thanks

@daleif \seq_clear:N \l__daleif_tmp_tl looks wrong, shouldn't that be in a seq var not a tl?

daleif

13:25

@DavidCarlisle that should be a tl_clear as it is a tl var in both cases.

David Carlisle

@daleif OK I didn't really follow the code flow just looked at the names:-)

daleif

14:00

@DavidCarlisle As far as I can see from the docs \seq_get_right puts the item into a tl var.

David Carlisle

@daleif yes sure thats fine, but \seq_clear:N expects a seq var (or conversely you should use \tl_clear:N to cleara tl)

Joseph Wright

14:16

\text_reverse:n anyone? tex.stackexchange.com/questions/720238/…

Ulrike Fischer

@JosephWright not sure yet, imho the question is very unclear. I wonder if they simply want to change the direction.

Joseph Wright

\ExplSyntaxOn
\cs_new:Npn \text_reverse:n #1
  {
    \text_map_function:nN {#1} \__text_reverse_aux:nw
    \__text_reverse_end:n { }
  }
\cs_new:Npn \__text_reverse_aux:nw #1#2 \__text_reverse_end:n #3
  {
    #2
    \__text_reverse_end:n {#1#3}
  }
\cs_new:Npn \__text_reverse_end:n #1 {#1}
\ExplSyntaxOff

@DavidCarlisle Indeed, but if they do just want it reversed ^^^

Ulrike Fischer

14:31

@JosephWright how would that e.g. handle a char with a following combining accent? Is that a grapheme?

daleif

@DavidCarlisle I know, the error with \seq_clear:N \l__daleif_tmp_tl l was that is whould be \tl_clear:N \l__daleif_tmp_tl l, there is only one \seq_clear but there should be several \tl_clear

Joseph Wright

@UlrikeFischer That is indeed a grapheme

@UlrikeFischer Anything that's a 'user perceived char'

@UlrikeFischer \tl_show:e { \text_reverse:n { Spın̈al~Tap } }

Ulrike Fischer

@JosephWright I had forgotten and was to lazy to check ;-)

Pablo González L

15:38

Hi, There is a problem in appropriating the name of a dimension for a \itemwidth package if it is not (or I cannot find it) defined in some TeX Live package $ grep -rn 'itemwidth' /usr/local/texlive/2024/texmf-dist/

Alan Munn

15:49

@PabloGonzálezL Is this a question?

Pablo González L

@AlanMunn Yes: The dimensions are defined globally, and the name is not occupied by any official package... so the question is whether I can take ownership of it...

Joseph Wright

16:12

@UlrikeFischer :)

David Carlisle

@PabloGonzálezL well if you do ifx\itemwidth\@undefiend \newdimen\itemwidth, even if another package (or a future latex format) defines it as a register you'd still be ok. But if you feel you need to ask about a name, that's probably a hnt that you shpuldn't use it

@PabloGonzálezL I interpreted this as a question but you have stated it as an assertion that there is a problem (which is why @AlanMunn asked, I assume)

Pablo González L

@DavidCarlisle @AlanMunn Sorry if I expressed myself wrong :( ...the truth is that I can use another name but this one fits, unfortunately I can't do "scope" like with the commands

David Carlisle

@PabloGonzálezL as I say in practice it's unlikely to be defined as anything other than a dimen register so it's probably safe enough

Pablo González L

@DavidCarlisle Great, I didn't want to run into any problems or conflicts with any official package :D

@DavidCarlisle Another thing, I made a couple of requests in the latex2e repo and they are already resolved... should I close them?

David Carlisle

16:27

@PabloGonzálezL well if we wanted the name, you having a package with a register of that name may not stop us using that name, but we'd warn you (and as it would probably be a dimen register anyway you probably would not need to do anything)

@PabloGonzálezL things usually get fixed in latex-dev first and just get a fixed-in-dev tag, then we close them when we do a release from the main branch.

Pablo González L

@DavidCarlisle OK, the only thing missing was an s in (latex-project.org) in (github.com/latex3/latex2e/issues/1374)...but everything was OK

Paulo Cereda

18:34

Alan Munn

@PauloCereda What do you expect from a company from Bielefeld.

Paulo Cereda

@AlanMunn lol

Skillmon

@AlanMunn must be a covered operation base! Don't trust anything claiming to be from Bielefeld.

Alan Munn

@Skillmon You mean it's a stealth Margherita?

Skillmon

@AlanMunn in case you're not aware: en.wikipedia.org/wiki/Bielefeld_conspiracy

Alan Munn

18:47

@Skillmon Yes, I'm aware, otherwise I wouldn't have made the original comment...

Skillmon

@AlanMunn good, good

@AlanMunn I thought you were just stressing the fact its a German company having no authority about what pizza is.

Alan Munn

@Skillmon I mean that's true too, but being from Bielefeld is the icing on the cake. (Or in this case the pineapple on the pizza).

Pablo González L

19:22

@Skillmon It is possible to implement something like what is asked here (tex.stackexchange.com/q/693947/7832)?

Skillmon

@PabloGonzálezL not with documented interfaces, no.

David Carlisle

@karlh you'd have to ask on the biber repo or perhaps someone knowledgable about biblatex/biber will pass by here. it looks wrong to me but I don't know that code well. — David Carlisle 41 secs ago

@JosephWright ^

Pablo González L

@Skillmon sniiiiif :(

Skillmon

@DavidCarlisle could be a pseudo-length (so a length expression stored as a tl) as well.

David Carlisle

@Skillmon well then @PabloGonzálezL's package would break, but as it's not my package, I wouldn't care. (I'm kind like that)

@Skillmon not even a "it is easy to do in expkv" response?

Skillmon

19:41

@DavidCarlisle because it really isn't as expkv doesn't have a notion of submodule

David Carlisle

@Skillmon you could have said it anyway, it would have looked good and no one would have noticed it was not true.

Skillmon

@DavidCarlisle :P I tend to provide exemplary code along with my claims.

cis

20:59

"And improving this has currently a rather low priority."

"It's none of my business," but:

If there is an international file format (called *.pgn), then it would be 'my' first priority to be able to read these 'external files' into the LaTeX package....
Well, that's just my opinion...

PS:
I recently communicated with chess people. The aim was to visualize *all* the images of a game.

My solution was: No problem with "xskakloop"...
Their point of view: no, no desire to use LaTeX if you can't use the pgn files without post-processing (as usual on some websites).

samcarter

21:25

@cis If reading pgn files is important for you, you could create a package

cis

@samcarter No, I can't. I am dumb... :X

Alan Munn

@cis But then don't presume to know what others should be doing with their time. Almost all TeX packages are written on a totally volunteer basis by people who have other things in their life. So telling people who produce software on a totally volunteer basis what they should do isn't usually very appreciated.

5

cis

It's just an interjection... I already said "it's none of my business" (German: "geht mich nichts an")...

Not everyone responds grumpily to suggestions for improvement. Good work is also inspired by the test persons....

For example, I don't get offended when someone tells me that I made a miscalculation - I say: "Thanks for the tip"...

Alan Munn

@cis I'm sorry, but if you have to say "It's none of my business" you're pretty much guaranteeing that you're saying something that you probably shouldn't otherwise you wouldn't have used that preface.

cis

21:44

Maybe, but it doesn't change the technical or content details. This is a debate about polite phrases.

"It's none of my business" means "as long as I don't do it completely myself, it's not my place to judge other people's work" - I didn't do that either.
But I think one can make suggestions for improvements with regard to future versions etc.

Alan Munn

@cis Well that might mean what the German phrase means when it's used, but I would not say that is what the phrase in English usually means. In English it has a pragmatics very similar to expressions like "I'm not a racist, but ..." or "I'm not a sexist but ..." which are almost certainly followed by some sort of racist or sexist statement.

David Carlisle

@cis no sorry that is not how that is read by a native english speaker, the phrase "it's none of my business, but" means "what follows is intentionally rude"

Skillmon

22:22

New stupid idea: Begin sentences with alert phrases just to state something completely harmless and correct. Example: I'm not a sexist, but LaTeX is a great typesetting system.

Alan Munn

@Skillmon bbc.com/news/articles/cd11gzejgz4o

Skillmon

@AlanMunn I think I read the original source for that non-toxic glue on pizza once. I'm certain I remember this, but can't remember where I read it. And it was certainly before AI overview.

Paulo Cereda

@AlanMunn @barbarabeeton frobnicate: man7.org/linux/man-pages/man3/memfrob.3.html

barbara beeton

22:51

@PauloCereda -- Hmm. new word for me. Sounds subversive.

NaCl

23:19

Hey, I have a soft question which is probably better discussed in chat here. I'm doing programming language stuff and have some macros for that, but I'm now reconsidering/surveying if there are better ways (without compiling from custom markup to TeX). I have two options right now:

David Carlisle

@NaCl ?

NaCl

Either build (maybe abstract) syntax trees explicitly, for example, `print 3+(2*f)`, where `f` is a construct outside of the programming language, so usually indicated by not using colors (or a different one) may look like `\xprint[\xcol]{\xadd[\xcol]{\xcol{3}}{\lrpars{\xmul[\xcol]{\xcol{2}}{f}}}}`, or just rely on syntax and define shorthand macros that are locally scoped, like `\xlang{\xprint \n3 + \l( \n2 * f\r)}`.
Both have their pros and cons.

I like the first, because it enforces at typesetting-time the arity of operations. I hate it, because the TeX turns into Lisp and I have yet to find a good editor config for essentially writing Latex as a somewhat readable Lisp. (notably, editors automagically closing a brace is big pain)

On the other end of the spectrum, I like the second, for exactly the opposite reason: it's more readable, easier to debug and also write. However, it's easy to introduce typos and have the document compile fine, which makes everything more difficult

My question is whether you know some kind of in-between or have otherwise some tips on how to improve my situation

David Carlisle

@NaCl as you see chat isn't a good forum for code based questions.. I am not sure I understand your question, why not just use print 3+(2*f) as the input syntax and have it parsed while typesetting cf listings or minted or using luatex and a custom parser in Lua for your input form? on editors emacs only auto closes a brace if you want it to,

NaCl

ok, I should probably invest some time to identify the option in my editor to turn that stupid auto-close feature off or have it be smarter :-)

I'm constrained to pdflatex, so no lua. The code contains mathy expressions, can be inline and ~~can~~is expected to itself occur in math envs

I thought about writing a parser in TeX itself, so I could just write print 3+(2*f) and it expands to the respective macros or throws some kind of error

But I'm both worried about the amount of time I need to spend to learn how to do that vs the benefit as well as the amount of time the document needs to compile when employing such a tactic

David Carlisle

23:38

@NaCl why the pdftex constraint? longer term unicode, accessibility, ... a fundamentally 8 bit system won't be feasible. You could of course write a parser in tex, I wrote a more or less complete xml parser including entity declarations, namespaces etc in tex but it would be a lot easier in lua.

NaCl

I have collaborators who insist on pdflatex

David Carlisle

@NaCl sometimes, people are just wrong.

NaCl

So, I reckon there is not much to do here, is there

David Carlisle

@NaCl you can use pdftex as I say listings would get you a reasonable way or minted (which uses shell-escape and a parser written in python, so it depnds if shell-escape is acceptable. I don't think I'd want to hand write either of the two forms that you suggest for any expression bigger than the one you show, so emitting that from a parser eithert an external tool or written in tex seems the only viable option

NaCl

Thanks! Makes sense

Transcript for

$TeX - LaTeX$ TeX, LaTeX and Friends