TeX, LaTeX and Friends

@DavidCarlisle nice. Is marginparsep new in that version?

@UlrikeFischer no 2.09 has marginparsep (and just set all this to fixed lengths) latex2e tried to calculate some so you could set the paper size and have the rest default to something sensible but this looks wrong to me, the first version just didn't take marginparsep into acount in the calculation so if you set it large the margin would go off page

barbara beeton

@DavidCarlisle -- Sorry, at that time I was embroiled in the care and feeding of AMS-LaTeX and the AMS document classes, which have a quite different approach to setting page dimensions, and also don't support metric paper sizes,

@barbarabeeton well clearly it can't be my fault so if not you, we'll have to blame @UlrikeFischer

barbara beeton

@DavidCarlisle -- Oh, be kind to @UlrikeFischer. Else she'll sic Bär on you.

12:34 AM

@DavidCarlisle what was the default marginparsep? 0.4in?

@UlrikeFischer 1994:

\if@twocolumn
\setlength\marginparsep {10\p@}
\else
%<10pt&!bk> \setlength\marginparsep{11\p@}

@DavidCarlisle and nobody complained - I hid my mischief well ...

2

12:49 AM

@UlrikeFischer I doubt we can change a4 or us letter but I wonder if we could change a5 by adding .4in back if marginpar is less than say 10pt and it would still fit in paper width

Q: How exactly does geometry calculate \marginparwidth if only paper size is set? Why 3pt for A5, 57pt for A4 and 65pt for A6?

1:02 AM

@DavidCarlisle I thought it was geometry because I thought article generally ignored paper size. But @yannisl pointed out otherwise.

1

How exactly does geometry calculate the dimensions for page layout if only the paper size is specified? From the documentation, I expected a ratio of 0.7 to be used for the text block and this seems to be the case. The documentation also says the margin note width (\marginparwidth) is ignored by ...

@DavidCarlisle I should delete the question, though. It doesn't make any sense.

@cfr I've raised it on the team list. (yes you could probably delete or change the question)

@DavidCarlisle What does book or twoside do? Because \marginparwidth seems to come out to 51pt for book and a5paper.

@cfr sorry you can look or I'll look in the morning

@DavidCarlisle Deleted.

@DavidCarlisle Nos da ;).

2:39 AM

@DavidCarlisle It's conditional. If the book is double-sided, it only subtracts 0.4in once:

\if@twoside
\setlength\@tempdima        {\paperwidth}
\addtolength\@tempdima      {-\textwidth}
\setlength\oddsidemargin    {.4\@tempdima}
\addtolength\oddsidemargin  {-1in}
\setlength\marginparwidth   {.6\@tempdima}
\addtolength\marginparwidth {-\marginparsep}
\addtolength\marginparwidth {-0.4in}
\else
\setlength\@tempdima        {\paperwidth}
\addtolength\@tempdima      {-\textwidth}
\setlength\oddsidemargin    {.5\@tempdima}
\addtolength\oddsidemargin  {-1in}
\setlength\marginparwidth   {.5\@tempdima}

Paulo Cereda

9:19 AM

@UlrikeFischer ooh ducks are very good at secrets

@cfr @DavidCarlisle The whole algorithm should be revisited IMHO. The algo is based on the font size like ``10pt` and papersize User space should only define left right and marginparwidth. Rest are just mirrored. some of the voodoo magic could go in the .clo files!

9:41 AM

@yannisl sure but that would be in geometry or memoir or some such place, there are really strict limits on what we can change in the standard classes. It is scary even to contemplate changing the 3pt marginpar width in a5paper

9:53 AM

@DavidCarlisle Very true, no-one wants to disturb a flourishing ecosystem!

@yannisl you can't just use left right and margin width though, you need to specify at least one of textwidth or paper width, latex can't know what size paper you are printing on

@DavidCarlisle Yes the papersize is necessary, either from default or a4paper etc. textwidth you do it the other way round papersize-lmargin-rmargin. No hardcode magic numbers based on font and 70-80 characters.

@DavidCarlisle Since user can adjust left and right margin to vary the textwidth

@DavidCarlisle some time back... tex.stackexchange.com/questions/55502/… I was learning tikz when I did those layouts and never published the code, main reason I kept getting errors with margipars when changing between classes and blamed my tkz code.

10:12 AM

@yannisl yes that's geometry package approach you can specify either and if you specify both and they don't add up you get an error that it's over-constrained

10:27 AM

@yannisl I think it's reasonable that the [11pt] and [12pt] options affect the default text width and users wouldn't have to soecify new lengths when adding [12pt]

10:38 AM

@DavidCarlisle What about 10.5pt or 11.5pt or 12.5pt? With modern engines any size is feasible.

@yannisl the standard classes do not have options with those names, the option names are not font sizes, they are three named layout options

@yannisl modern engines are not the issue, scalable fonts were available in 1993 when this was set up.

@DavidCarlisle True but the gold pot under the rainbow, is to find formulae that relate the layout to textwidth, since the standard classes do this. But I get your point these classes are the bedrock that a lot of other classes are based on and should not be changed.

11:04 AM

@DavidCarlisle would a change only affect a5paper? Isn't that used in other papersizes too?

11:15 AM

@UlrikeFischer a real fix would affect all sizes, but as I say we could choose to only apply it if the current marginpar size is unusably small, say less than 10pt, so it would only affect a5paper, I don't think we can change the default marginpar width in a4 and letter paper after 29 years

@UlrikeFischer meanwhile I git cloned the arlington stuff, I'll see if I can compile their c++ tool later

@DavidCarlisle yes, just saw the team list mail.

[  7%] Building CXX object CMakeFiles/TestGrammar.dir/src/Main.cpp.o
[ 15%] Building CXX object CMakeFiles/TestGrammar.dir/src/ArlingtonTSVGrammarFile.cpp.o
[ 23%] Building CXX object CMakeFiles/TestGrammar.dir/src/CheckDVA.cpp.o
[ 30%] Building CXX object CMakeFiles/TestGrammar.dir/src/CheckGrammar.cpp.o
[ 38%] Building CXX object CMakeFiles/TestGrammar.dir/src/ParseObjects.cpp.o
[ 46%] Building CXX object CMakeFiles/TestGrammar.dir/src/PredicateProcessor.cpp.o
[ 53%] Building CXX object CMakeFiles/TestGrammar.dir/src/LRParsePredicate.cpp.o

@UlrikeFischer ^

@DavidCarlisle this looks like some magic ...

@UlrikeFischer ooh it works (or at least it gives a help message)

TestGrammar_d


Arlington PDF Model C++ P.o.C. version v0.81 built Mar 30 2024 11:21:56 (GNU-C linux debug)
Choose one of: --pdf, --checkdva or --validate.

Usage:
        TestGrammar --tsvdir <dir> [--force <ver>|exact] [--out <fname|dir>] [--no-color] [--clobber] [--debug] [--brief] [--extensions <extn1[,extn2]>] [--password <pwd>] [--exclude string | @textfile.txt] [--dryrun] [--allfiles] [--validate | --checkdva <formalrep> | --pdf <fname|dir|@file.txt> ]

Options:
-h, --help         This usage message.

11:30 AM

where are you running that? in wsl?

@UlrikeFischer yes there is a native windows build file as well but I tried wsl first as thinkgs usually easier there

seems to say something (see skype)

@DavidCarlisle Oh, given up on cygwin? :P

@mickep it comes with a cmake stack and one pre-compiled shared lib in linux and windows flavours, getting it running under cygwin might take more than cmake .; make :-)

@DavidCarlisle Excuses. :)

@mickep also building these pdfs is something like 25% faster in wsl than cygwin (not done exact time comparison)

11:57 AM

@UlrikeFischer you have mail...

@DavidCarlisle yes saw it. But it complains about DA in my test file ;-(. Lets look if it does that too with the original files once I added an appearance.

@UlrikeFischer yes I noticed, I thought you'd be pleased:-)

12:21 PM

@DavidCarlisle I updated the exam paper and added an appearance, I also added an alt text as without it the default alt text wins as tooltip in acrobat.

@UlrikeFischer just warnings no errors, I'll email

@UlrikeFischer I guess you want me to see if I can make you a windows version to play with ....

@UlrikeFischer or do you have wsl installed?

@UlrikeFischer so that just leaves @JosephWright's other example with just tex AF (It would be simple to add mathml as well but it's a good test case as it is)

@DavidCarlisle I have wsl, I played around with it some time ago. But I need to check in which state it is and remember how to use it ...

@DavidCarlisle can I use a git repo that I pulled normally or must that be inside the wsl?

@UlrikeFischer you can use your existing files (that's what I'm doing) it's easy for wsl to see anything in C: (or other) windows drive but it is (intentionally) hard for windows applications to see wsl files as they are in a hidden folder, if you allow windows applications to write in that area they can easily mess up the filesystem, but going the other way is fine

@UlrikeFischer but as I say there is a visual studio build file for windows I'll look, I might have to install too much though I might not have all the build tools needed

samcarter

3:15 PM

Is "Ducks are good at finding stuff" a valid answer to tex.meta.stackexchange.com/questions/10163/… ? :)

3

barbara beeton

4:21 PM

@samcarter -- Ducks, you say? Take a look at the 27th Annual Pawcatuck River Duck Race: tiktok.com/@lucky.the.duck.pawcatuck/video/7351579692941593899

@DavidCarlisle It doesn't even need changing for all a5paper. For book, at least, it only affects the single sided layout. I don't know what it might break but isn't there already code to check if \marginparwidth comes out too large (> 2in)? So something symmetrical for too small?

samcarter

@barbarabeeton Oh, fast ducks (not to be confused with fast food)!

@cfr we are discussing it but we might just remove the duplicated line, it's pretty obviously a cut and paste error even if you failed to report it for 29 years

4:45 PM

@DavidCarlisle Interesting that nobody noticed during all the years.

@mickep they probably just thought "that's a small margin" and \setlength\marginparwidth{10cm} to make it bigger, but yes surprising it's lasted this long

@DavidCarlisle Maybe not so many used a5 for some odd reason.

@mickep yes well the a5 setting is unusual (and not to worrying to change as anyone with margin paragraphs must be setting it so won't be affected) what is more the issue is whether to fix the settings for a4 and us letter (ie 99.9999% of all latex documents ever) which by default have marginparagraphs .4in too narrow in one side settings

@DavidCarlisle I usually suggest changing, but regarding this I would not change.

@mickep but there are variants between change or not change, eg only change if \DocumentMetaData is added (so it's a new document) or only change if there's a full moon or ...

4:52 PM

@DavidCarlisle Or add a note about it in some documentation.

@mickep of course because everyone reads that

How much would it change for a simple a4 document?

@DavidCarlisle Well, somebody can at least say that is documented.

@mickep just the width of marginal nodes so not any line or page breaking in the main document, but the notes will then typically have smaller vertical size so may move a bit

@DavidCarlisle I've used a5paper a bunch of times. But always two-sided for making up booklets from A4.

@DavidCarlisle Smaller vertical size? Why?

@cfr I use a5 for my lecture notes.

4:55 PM

@mickep Because the notes will be wider so less tall.

@mickep wider paragraphs have longer lines so fewer lines per paragraph

@cfr area = height * width is tricky math for @mickep

Oh, indeed.

@DavidCarlisle Well, I thought you meant the available height was changing, and that I did not understand.

@mickep excuses

@DavidCarlisle That wouldn't benefit most new documents, though?

@cfr well we have plans....

4:58 PM

@DavidCarlisle \setuppapersize[A5][A5]? :P

@cfr the only "good answer" would be "don't make a typo 30 years ago" so every possible fix now has some bad aspects

@DavidCarlisle I was worried you might.

@DavidCarlisle $$ (not a typo...)

@DavidCarlisle instructables.com/…

Or there's instructables.com/Making-a-Time-Machine, but that seems too much trouble.

rdrg109

I have been able to display all characters that belong to `\l_char_special_seq` using.

```
\documentclass{article}
\begin{document}
\ExplSyntaxOn
\seq_map_inline:Nn \l_char_special_seq{\texttt{#1}}
\ExplSyntaxOff
\end{document}
```

I'm now trying to display the `\catcode` of all tokens that exist in `\l_char_special_seq`. I have tried the following

```
\ExplSyntaxOn
\seq_map_inline:Nn \l_char_special_seq {\number\catcode#1}
\ExplSyntaxOff
```

but I get the error: `Missing number, treated as zero.`. Any ideas on how to display the catcode of all tokens in `\l_char_special_seq`?

5:06 PM

it would be more idiomatic to use expl3 rather than the tex primitives \number\catcode but anyway \catcode takes a number not a character so you need `#1 if all the tokens are not expandable or a bit more if some of them may be

@rdrg109 oh they are not character tokens but csnames but OK you still need a back tick

\documentclass{article}
\begin{document}
\ExplSyntaxOn
\seq_map_inline:Nn \l_char_special_seq{\texttt{\tl_to_str:n{#1}}~\char_value_catcode:n{`#1}\par}
\ExplSyntaxOff
\end{document}

@rdrg109 ^^^

5:32 PM

@rdrg109 that code doesn't show the characters, it shows whatever the commands expand to which can be quite different. \\ is denoting a backslash but you are displaying it as a newline. see the version I just posted which displays the character preceded by a backslash and the current catcode

but you probably want

\documentclass{article}
\begin{document}
\ExplSyntaxOn
\def\foo{\seq_map_inline:Nn \l_char_special_seq{\texttt{\tl_to_str:n{##1}}~\char_value_catcode:n{`##1}\par}}
\ExplSyntaxOff
\foo
\end{document}

@rdrg109 note the different answers for \ , \_ and \~

rdrg109

6:26 PM

@DavidCarlisle Thanks for showing how to do it.

I have noticed that the documentation of \l_char_special_seq in `$ texdoc expl3` mentions:

```
Used to track which tokens will require special handling when working with verbatim-like material at the document level as they are not of categories ⟨letter⟩ (catcode 11) or ⟨other⟩ (catcode 12).
```

In the image that you sent, `\"` shows the catcode 12. Therefore, I believe that the documentation is not consistent because it explicitly mentions that the characters in `\l_char_special_seq` don't have catcode 11 or 12.

@rdrg109 as I show with the two versions, the seq just lists the characters it does not change as catcodes are changed. " is in the list because it is often catcode 13 in documents (for babel shortcuts) the fact that it happens to be 12 at the point you print this is not relevant

@rdrg109 vvv

\documentclass{article}
\usepackage[ngerman]{babel}
\begin{document}
\ExplSyntaxOn
\def\foo{\seq_map_inline:Nn \l_char_special_seq{\texttt{\tl_to_str:n{##1}}~\char_value_catcode:n{`##1}\par}}
\ExplSyntaxOff
\foo
\end{document}

@rdrg109 now it's not 12

rdrg109

Thanks for the explanation

6:57 PM

@DavidCarlisle @JosephWright l3sys-query is now in miktex but it only works with --shell-escape currently.

7:07 PM

@UlrikeFischer that's fair enough, Karl was more in the loop for security vetting it. I guess we should probably ask that it is added in miktex

@UlrikeFischer do you think it's in hand or should I open an issue at the miktex github?

@UlrikeFischer this I assume

github.com/MiKTeX/miktex/blob/…

7:28 PM

@DavidCarlisle it would probably not harm to ask.

Alan Munn

@UlrikeFischer In true @DavidCarlisle fashion, he should now tell you that he did ask. :)

@AlanMunn I have asked

Alan Munn

@DavidCarlisle :)

@DavidCarlisle githib?

7:44 PM

@UlrikeFischer github.com/MiKTeX/miktex/issues/1473

7:55 PM

@JosephWright did we intend to return catcode 12 spaces, it has some advantages in that you can handle names with repeated space but it makes printing filenames a bit cumbersome as you get this by default

Joseph Wright

10:00 PM

@DavidCarlisle Er, probably missed that: we read as catcode 12 to avoid losing spaces, but e should likely return catcode 10

@JosephWright I just happened to have the above file in a test directory where I checked the example I put in the above miktex issue

@JosephWright of course you could use a font with a space in the space slot but that's not so easy in pdftex:-)

Joseph Wright

@DavidCarlisle I’ll look at it when I get a moment: might be Tuesday as have plans tomorrow and Monday