TeX, LaTeX and Friends

barbara beeton

02:57

@mickep -- I've received a fairly extensive answer, but sadly no explicit term. This is too long to post here, but I'll happily send it on to you by email if you'll send me your address. (Mail to me at bnb at tug.org.)

user202729

03:07

By the way... do we "really" don't want opinion-based questions here, or just "because the site inherit the kill button from SE"?
*do the research*
Okay I find [Are "Is this good typography?" questions on topic? - TeX - LaTeX Meta Stack Exchange](https://tex.meta.stackexchange.com/questions/6647/are-is-this-good-typography-questions-on-topic) [Are questions for best practices on-topic? - TeX - LaTeX Meta Stack Exchange](https://tex.meta.stackexchange.com/questions/4369/are-questions-for-best-practices-on-topic) and I think the explanation there are good – opinion-based questions that *is n…

user202729

05:36

(Research a bit more there are comments there for the closure, so never mind)

user202729

05:47

@JosephWright (old message but) for this one I said "implementable in TeX" so the user base wouldn't be a huge issue. (basically scan through the whole document, parse the whole thing and replace \verb with \texttt and similar catcode-independent things, then scantokens the whole thing)

user202729

06:05

(And again, before I reimplement the wheel, did anyone already implement that)

mickep

07:33

@barbarabeeton Oh, nice. You got mail.

Paulo Cereda

09:02

ooh email

Joseph Wright

@user202729 Won't work: you've already lost things like comments

@user202729 We have always been more flexible

user202729

09:29

@JosephWright ... should definitely work, right? Don't grab the whole thing at once, grab tokens one by one

Even if grab the whole thing at once can still work somewhat, by setting verbatim everything and process as string

But have to reimplement TeX eyes, harder

David Carlisle

10:51

@user202729 not possible in general without executing/typesetting, there is no way to just parse

user202729

11:05

Yes I know

What I mean is...

Instead of parsing TeX in the normal way, make a separate parsing rule, pre-parse the whole thing, then execute it normally.

That way you can "easily" have verbatim and other special-structures work within argument

David Carlisle

@user202729 I don't see how you can get anything close to that while being compatible with current tex syntax

user202729

"text syntax"

Well, not strictly compatible...

But after all it seems that there are just a few "special-syntax" things (verbatim-like environments, \verb-like commands)

Plus maybe \ExplSyntaxOn and \makeatletter

Of course not 100% compatible, but not like that the user have to learn a completely new system either

David Carlisle

@user202729 and tikz matrix & and any number of babel language shorthands like "- and ...

@user202729 but perhaps I don't understand your idea. You presumably are still not expecting to support \newenvironment{foo}{\begin{verbatim}}{\end{verbatim}} ... \begin{foo} \whatever\end{foo} ? Just support literal \begin{verbatim} appearing inside the argument of a command in the document?

user202729

Obviously that one one is almost-impossible to support. For that one I'm thinking of letting the user define custom preparser rules for that particular environment

Originally it comes from that every time I want characters with special catcode in expl3 code I have to construct token list → regex replace → execute

Would be nicer to have "put character with catcode X charcode Y right here"

David Carlisle

@user202729 \Ucharcat

user202729

11:15

Hm

David Carlisle

@user202729 or its l3 name if you prefer

user202729

No, not that one

There's already char_generate:nn

David Carlisle

@user202729 yep that

user202729

It doesn't support having characters with special catcode within commands to be executed without elaborate regex replace and what else

David Carlisle

@user202729 in luatex it sounds like you want catcodetables (or an emulation of them in the others)

user202729

11:20

Hm? Not sure I want that either, but I'll take a look

What I mean is to write code like in tex.stackexchange.com/a/622505/250119 without putting the strings in a completely different section / auxiliary macros etc.

Even e.g. \newline_with_other_catcode is not as good as e.g. \cO\^^J of regex, and you have to \unexpanded the rest selectively

By the way the best way to process something preserve spaces for now is to regex replace spaces with something else, process, replace back, right? (or token cycle)

David Carlisle

@user202729 ? "for listings there's \lstinline." that changes catcodes same as teh enviornment form

user202729

No, lstinline has the experimental feature that can be used in argument without cprotect

Although still of course need to escape spaces and what else

David Carlisle

@user202729 well you don't need the macro of course, you could use \char_generate{13}{12} inline if you prefer

user202729

Hm right this case is simpler

But there are complex cases

Like passing a complex constructed token list to a function that takes n-type argument

"can't be that uncommon" right?

Anyway maybe I'll figure out something next time I need it

UD's DefineVerbatimToScantokens is good enough (I think? I use my own macro but this one looks similar) in this case

David Carlisle

@user202729 possibly although nested verbatim questions far more likely to appear on a tex geek forum like here than in real documents. The whole thing is a non-issue for real documents that hardly ever use verbatim at all and use verbatim nested in a command even less

CarLaTeX

14:17

@DavidCarlisle twitter.com/jeperego/status/…

David Carlisle

@CarLaTeX es fehlt nur die Ananas, dann wäre es gut.

Alan Munn

14:45

Pianist Krystian Zimerman interrupted by rubber duck during concert

►

3

CarLaTeX

14:58

@DavidCarlisle Ich stimme nicht zu

Ulrike Fischer

@CarLaTeX are pineapples out of bound in pasta too?

David Carlisle

@CarLaTeX Du bist Italiener, von dir kann nicht erwartet werden, dass du Pasta und Pizza verstehst

CarLaTeX

@UlrikeFischer Yes, no pineapple on pasta

Ulrike Fischer

David Carlisle

@UlrikeFischer better call a medic for @CarLaTeX

muzimuzhi Z

15:07

May I ask that are there any existing extensions/packages that bring pgfkeys' "style key" to keyval?

There are calls for this feature from time to time, among which the latest one is https://tex.stackexchange.com/q/629489

David Carlisle

15:30

@muzimuzhiZ it depends what you mean by pgfkeys style, l3keys is perhaps the closest to that, plan to converge more on l3keys going forward

@muzimuzhiZ ah just read the post you meant a key like pgfkeys key style, not a generic "keys in the style of pgfkeys" although actually my answer covers that as well:-)

Joseph Wright

15:59

@DavidCarlisle :)

muzimuzhi Z

16:19

The main aim is to extend `keyval` so that packages based on it can define key styles. Seems this does the trick
```tex
\NewDocumentCommand{\define@key@style}{ m m o m }{%
\IfNoValueTF{#3}{%
\define@key{#1}{#2}{\KV@do #4,\relax,}
}{%
\define@key{#1}{#2}[{#3}]{\KV@do #4,\relax,}
}%
}
```

Faheem Mitha

@muzimuzhiZ Why not just use pgfkeys?

I'm not familiar with keyval.

David Carlisle

@muzimuzhiZ we won't be extending keyval (@Skillmon tells me it's already perfect) but will be helping people ease over to l3keys at some point.

@FaheemMitha you are not? You have never used graphicx, or hyperref for example?

Faheem Mitha

@DavidCarlisle Occasionally, for both. But I've never used keyval directly.

muzimuzhi Z

@FaheemMitha Cause there're plenty packages using keyval and users might want to have "key style".

@DavidCarlisle I understand. It is not a feature request to keyval, only casual/personal attempts.

The situation is, pkgA uses keyval, users use pkgA. For me it's easier to patch keyval than to change the key-val package that pkgA is based on. though it's not the ideal way.

@FaheemMitha well, keyval.sty is short enough

Faheem Mitha

16:36

@muzimuzhiZ Oh, I thought you meant for your use.

I've only used pgfkeys. It seems to work for what I've used it for. And the documentation is a bit hard to grasp, but good by prevailing standards.

I had to read it a bunch of times, and then experiment a bit.

Joseph Wright

@muzimuzhiZ Just \define@key the right key name ...

@muzimuzhiZ You don't need to patch anything, just define the key appropriately

muzimuzhi Z

@JosephWright Why the "key name"? It seems to me what needed is to replace the user-specified key value #1 with \KV@do#1,\relax,.

Joseph Wright

@muzimuzhiZ I mean all you need to to have \setkeys{...}{...} as the 'payload'

muzimuzhi Z

16:54

You mean sth like
\define@key{fam}{mystyle}{\setkeys{fam}{<key-val>}} ?

I slightly dislike this way because the family is read and set twice. The pro is it uses only the two public macros \define@key and \setkeys.

BTW, in the keyval package manual,

% A declaration of the form:\\
% |\define@key{family}{key}{...}|\\
% Defines a macro |\KV@prefix@key| with one argument. When used in a
https://github.com/latex3/latex2e/blob/c7ce5c12653100d7677362cabd27748cb3ab3de3/required/graphics/keyval.dtx#L178

It the `\KV@prefix@key` a typo? To me it should be `\KV@family@key`

David Carlisle

@muzimuzhiZ most of the code uses prefix rather than family see \csname\KV@prefix\@tempa\endcsname and similar lines. I think I changed to use family in the documentation as it's less obviously a "prefix" when used as a separate argument. But it's all a long time ago...

muzimuzhi Z

17:11

The following lines explaining the `[default]` uses `family` consistently,

% A declaration of the form:\\
% |\define@key{family}{key}[default]{...}|\\
% Defines a macro |\KV@family@key| as above, however it also defines the
% macro |\KV@family@key@default| as a macro with no arguments, and
% definition\\
% |\KV@family@key{default}|.

David Carlisle

@muzimuzhiZ yes the user doc part should use family as you say. I was just explaining it's not so much a typo as a remnant of the original naming

muzimuzhi Z

got it

bertieb

17:35

(hey, you guys' chat box area in the footer has transparency? when did you get that feature?)

David Carlisle

@bertieb ?

bertieb

@DavidCarlisle if you scroll back in chat history, you can see chat (ie messages' text) behind the footer- at least that's how its rendering here!

Quick Q about question formulation- I have a Q along the lines of "repeat until a page is filled" (eg tex.stackexchange.com/questions/6608/…, tex.stackexchange.com/questions/217861/…), except it's about a single-page table. I'm having a bit of trouble with my MWE as I'm not that au fait with tables- how much will I anger the local deities if I also sneak in a query about adjusting the table I'm trying to(1/)

repeat (2/)

David Carlisle

@bertieb not here tried firefox and chrome

bertieb

...there's a certain delicious irony in SE chat making me break my message at the final word in the LaTeX room

David Carlisle

ah I thought you meant the text box itself was transparent, I think messages have always shown behind the coloured band of the footer

bertieb

17:44

oh right apologies for not being clear! I don't recall it (tho have not been in here often) and RA doesn't get transparency in its footer

@bertieb basically there's a couple of things that I have yet to figure out how to make a table do (give a column width despite no content; and span a table to the width of a page); neither of which are terribly difficult but I haven't set found the right QA on the site to explain it to my tired brain

barbara beeton

@mickep -- mail sent.

mickep

@barbarabeeton Well received. Thank you very much! I was very happy to read the extensive answer. I liked in particular the quote by Mark Twain and the discussion on aperture vs. gap, where I agree that aperture sounds nicer. Again, thanks for asking your friends.

bertieb

@bertieb so I'm here to check if saying "is there a way to make this nearly-contentless table repeat (so I can print a nice empty table to fill in); oh and also how would I make it look better by adjusting the column width and the table width?"

CarLaTeX

18:29

@UlrikeFischer Oh no!

@DavidCarlisle I need them

Rmano

@UlrikeFischer ugh, first time I see this. On the other hand, my mother used pineapple in her pot-roasted pork --- the pineapple has a very powerful tenderizer for dry meats, and the result was really nice. (Sorry @CarLaTeX, I hope you can survive this... ;-))

CarLaTeX

@Rmano Mia mamma una specie di arrosto ripieno di prugne. Considerando che sono lombarda, non c'è Natale senza mostarda di frutta (col lesso)

Rmano

@CarLaTeX :-) hmmm, mostarda. yummy. (although lately, the one I buy in Italy is less and less spicy ...)

CarLaTeX

18:45

@Rmano The best is Sperlari, if you don't have a gastronomia di fiducia :)

Rmano

@CarLaTeX yes, that is the one I have. But I remember it hotter (in the sense of "piccante") when I was younger... (maybe it's my taste that evolved ...)

CarLaTeX

@Rmano Maybe you're right, but I'm getting older, too. Comunque il mandarino non riesco a mangiarlo :D

Canageek

19:09

@DavidCarlisle Thanks

Also: Thanks everyone. SOrry for the taking ages to reply, I got sick right after my defence and am just getting back to work on comments

I decided I needed a break after that, since I'd missed the sememster change deadline anyway

Skillmon

21:16

@JosephWright no, \setkeys isn't nestable (well, technically here it doesn't hurt as the set will be the same set, but still...)

Skillmon

21:36

(better formulated: yes, fine as long as the nested \setkeys uses the same family, else havoc!)

@samcarter sorry, it took me way too long :) In my defence I travelled a lot this weekend!

samcarter

23:18

@Skillmon I'll better consult your travel schedule before making a joke again :)

Transcript for

$TeX - LaTeX$ TeX, LaTeX and Friends