TeX, LaTeX and Friends

cis

01:17

5 hours ago, by Vincent Mia Edie Verheyen

@cis page=2 doesn't make a difference when I compile?

@VincentMiaEdieVerheyen Did you compile it? Because if I compile it, it makes a difference whether I use page=2 or not.

\documentclass[]{article}
%\usepackage{mwe} % Dummy Images
\usepackage[margin=1cm]{geometry}
\usepackage{graphicx}
\begin{document}
\includegraphics[width=300px, height=700px]{example-image}

\includegraphics[width=300px, height=700px]{example-image-a4-numbered}
\includegraphics[width=300px, height=700px, page=2]{example-image-a4-numbered}
\end{document}

Where did I know the key?
I looked into the manual ----> texdoc mwe.

And that's the point: you have to find out / learn things using simple minimal examples (that's not a phantasy term); as I said here: https://tex.stackexchange.com/questions/561458/drawruler-how-to-have-ifnum-n-xyz-to-be-drawn-longer-or-shorter/561460#561460
If you always want to figure out everything in the overall code, nothing will work; and in particular, hardly anyone will sit down and try to understand a large amount of code, about 90% of which has nothing to do with the partial problem.

cis

06:44

...
....
........ Mhhhh.....

If I say \def\mydef{} is this a case for NoValueTrue?

Seems not in some cases. :()

....
....
....
.....

@ xparsers

Could someone solve this. It think it has to work with tcolorbox manual p. 448

IfNoValueTF={\thistitle}{ true }{ false }

but \def\thistitle{} seems not to initialize a true-tag.

2

Q: tcolorbox - Define a macro for the title that can be empty

I fail to use a macro that give a possibly empty title for a tcolorbox frame. Here is the code used. \documentclass{article} \RequirePackage{tcolorbox} \begin{document} \def\thistitle{One title} \begin{tcolorbox}[title = \thistitle] Bla, bla, bla.... \end{tcolorbox} \def\thistitle{} \begi...

tcolorbox

I tried

\tcbset{
IfNoValueTF={\thistitle}{}{title=\thistitle},
}

has no effect.

Anonymous

07:23

@cis I mad a MWE here now asking for the looping of this minimal code, for pictures 1-32, 2 images per page: tex.stackexchange.com/questions/561547/…

cis

07:55

https://chat.stackexchange.com/transcript/message/55460874#55460874

@frougon solved it with pgfkeys ---> 0:1 for the xparser, who have nothing done... :()

TeXnician

@cis You are mixing up no value with empty which are independent concepts. If a macro is empty it does not have no value.

cis

08:31

Yes, I was afraid of that.

However, I prefer a pgfkey solution itself. This has the advantage that I understand it. :()

TeXnician

@cis That's your call, I just wanted to clarify that the behavior you observed is absolutely expected (and documented btw ;)).

cis

08:43

@VincentMiaEdieVerheyen
I don't like to say it: In my opinion, you have to simplify and abstractise this a lot more:

You have a set of initial data:

Object Date     North  East  Wight Volume Density ImageNo Image
0001   2020.08  51.2   49.1  56    120    0.65    1       example-image
0002   2020.09  51.2   39.1  77    133    0.65    2       example-image-a
............................
............................
............................

This table now needs to be expanded in a document in a loop.

Object: 0001
Date: 2020.08
North: 51.2
East: 49.1
...
....
....




Object: 0002
Date: 2020.09
North: 51.2
East: 39.1
...
....
....




...
....
....
...
....
....

All the graphic trappings are initially irrelevant and can be integrated later.

Maybe someone will answer that for you.

If someone solves a complicated overall example for you, it was more of a coincidence.

Joseph Wright

10:23

@UlrikeFischer Working on DeviceN - I've realised a few aspects of colorspace are sub-optimal :)

@UlrikeFischer Now I have to work out the PostScript for conversion - it's pretty 'fun' (@DavidCarlisle)

Joseph Wright

10:58

Two xstring questions within minutes of each other

Weird

David Carlisle

@JosephWright standard answer: don't use xstring:-)

Joseph Wright

@DavidCarlisle Well that's likely a bit harsh, but I understand your point: people see 'string' and don't understand what is meant (it means a string of tokens), nor expandablity

David Carlisle

@JosephWright well for the one I answered where there is no top level use case just "how do I use this in an edef" not much you can say other than don't do that.

Joseph Wright

11:13

@DavidCarlisle Fair

David Carlisle

@JosephWright will be only sporadically online next couple of days

Joseph Wright

@DavidCarlisle Got it

Phelype Oleinik

11:39

@JosephWright I've pushed a fix to the ConTeXt issue. I wonder if we should have some tests for generic mode...

Joseph Wright

@PhelypeOleinik We could, but as ConTeXt doesn't get routinely updated in TL, we'd likely still miss things

@PhelypeOleinik Should I release?

Phelype Oleinik

@JosephWright Yes, that is a problem, but I think most of what breaks with ConTeXt will also break in Plain (this one did, for example), so testing only Plain would catch most of the problems, I think

@JosephWright Sure

Joseph Wright

@PhelypeOleinik That would help, yes: we probably just need a very simple test

Phelype Oleinik

@JosephWright I need a different .lua build file for that, right?

Joseph Wright

@PhelypeOleinik config-plain.lua or similar

@PhelypeOleinik See the l3build repo itself

Phelype Oleinik

11:46

@JosephWright Got it. Thanks!

@JosephWright Ah, you'll also need to update m3expl009 in the 2e repo: I switched from \@latex@error to \__kernel_msg_error:nnnn, so the message will say LaTeX3 Error instead of LaTeX Error

Joseph Wright

@PhelypeOleinik Ah, oh goody

@PhelypeOleinik Until we change the message text ;)

Phelype Oleinik

@JosephWright It was that or defining a \@latex@error in non-LaTeX formats. I figured it would be better

Joseph Wright

@PhelypeOleinik Anyway, first I have to build the tagged version, then build LaTeX2e, then release :)

@PhelypeOleinik Definitely

Phelype Oleinik

@JosephWright Yes, I had that in mind. I hope Frank won't be upset if the message says LaTeX3 for some time ;-)

@JosephWright Sounds fun :-)

Ulrike Fischer

12:35

@JosephWright I don't doubt it, imho it is combining xcolor command and other stuff in a quite confusing way.

Ulrike Fischer

13:07

@MarcelKrüger on the texlive list was just mentioned that lua-alt-getopt ctan.org/pkg/lua-alt-getopt is a dependency of luaotfload(-tool). Do we need that?

Anonymous

@cis I seem to have found something that prevents me from doing the loop: tex.stackexchange.com/q/561583/67761

Anonymous

There seems to be some incompatibility with how \pgfmathsetlengthmacro is used now.

cis

13:29

@VincentMiaEdieVerheyen I told you that you are going in a difficult or impossible direction.

With a representation that only a few people are likely to notice. They don't know the background. And if you deal extensively in a novel-like manner in threads, most of them are mentally out.

It's not that hard to ask a smart question:

I have a set of datas

Object Date     North  East  Wight Volume Density ImageNo Image
0001   2020.08  51.2   49.1  56    120    0.65    1       example-image
0002   2020.09  51.2   39.1  77    133    0.65    2       example-image-a
............................
............................
............................


How can I read out these table-rows and expand them in a document (pagewise)?

Like:

Object: 0001
Date: 2020.08
North: 51.2
East: 49.1
...
...
\newpage

Object: 0002
Date: 2020.09

Anonymous

@cis I have only 1 something that needs to change within a document, and that is the image. The images need to be looped. All the other info can stay the same. There is only 1 object per document.

Anonymous

I think I found a good loop now, which I described here: tex.stackexchange.com/questions/561583/… only I do not seem to be able to include the \pgfmathsetlengthmacro ... so I can't compile

cis

13:58

I had to withdraw my comment there. I do not get it.

As I said, I would do it differently. For me sth. like that starts with a table that contains all the essential data, and this then expands to the document.
But everyone as he likes it.

Joseph Wright

14:13

@UlrikeFischer Working on it slowly

@UlrikeFischer I'm getting the 'backend' stuff done, then I'll have to think about mixing: might well have some queries for you at that stage

Marcel Krüger

@UlrikeFischer We could avoid using it, but I see no reason why we should. Is there a problem with lua-alt-getopt?

Ulrike Fischer

14:30

@MarcelKrüger not that I know of (it seems to be there since many years). I was only surprised that there is such a dependency.

Paulo Cereda

Friends, arara 6.0 (dev) is now 100% Kotlin!

4

Ulrike Fischer

14:55

3

Paulo Cereda

@UlrikeFischer oh no

The juice is orange, though

Joseph Wright

15:19

@UlrikeFischer Probably goes back to well before any of our involvement; I remember looking at the opt code when I started l3build, and it was already many years old

Ulrike Fischer

@JosephWright yes probably. How can I test if the dvipdfmx backend has been loaded? \g__sys_backend_tl is an internal variable.

Paulo Cereda

@DavidCarlisle quack

Joseph Wright

@UlrikeFischer Well the idea was that all that sort of stuff should be in l3backend ...

Ulrike Fischer

@JosephWright but this means that you can test if pdftex is used, if pdfmode is active and so on, but you can't test if dvipdfmx or dvisvg or dvips is the target. I'm just trying to expand the "new-transparent.sty" to xetex/dvipdfmx but can't detect the second case.

Joseph Wright

15:47

@UlrikeFischer Reasoning is of course that l3backend is intended over time to provide a backend-neutral API for all this sort of stuff

@UlrikeFischer Of course, I accept the point: probably we should make it public

Ulrike Fischer

@JosephWright yes, if possible. But I doubt that really everything can be done and what can be done, will need some time. There is quite an amount off driver code in use in various places. E.g for the texttransparent stuff, dvips simply doesn't work, and for xetex/dvipdfmx it is much easier to use the new special than to go over color stacks (which aren't in the binary yet anyway).

Anonymous

16:38

@UlrichDiez I can not use your great code tex.stackexchange.com/a/561593/67761 along with % Images Ruler x of the bottom code at tex.stackexchange.com/a/561482/67761

Anonymous

The code goes great with your answer, as such:

Anonymous

% arara: pdflatex
% arara: pdflatex
% arara: pdflatex

\documentclass[a4paper, landscape]{article}
\pagestyle{empty}
\usepackage[showframe=false,
margin=20mm,
]{geometry}
%\geometry{paperwidth=15cm, paperheight=30cm, margin=2mm}% optional
%\usepackage{mwe} % Dummy Images
\usepackage{textcomp} % \textcopyright

\usepackage{tikz}
\usetikzlibrary{calc}
\pgfkeys{/tikz/savevalue/.code 2 args={\global\edef#1{#2}}}
\pgfdeclarelayer{background}
\pgfdeclarelayer{foreground}
\pgfsetlayers{background,main,foreground}

Anonymous

But as soon as you want to inlucde, for example the

% Images Ruler x

section within the \begin{tikzpicture} it goes wrong .... it doesn't compile anymore.

Anonymous

That section is, for example, the following code:

Anonymous

\newcommand\xRuler[5]{
\path let \p1=($(#1)-(#2)$) in
\pgfextra{ \pgfmathsetlengthmacro{\RasterboxesWidth}{veclen(\x1,\y1)} }
[savevalue={\w}{\RasterboxesWidth}];
\pgfmathsetmacro\NoXdiv{\w/\u}  % Number of x dividions

\draw[Ruler] (#1) coordinate(X) -- (#2);
\foreach \n in {0,1,...,\NoXdiv}{%%
\draw[Ruler] ([xshift=\n*\u]X) -- +(0,#4) node[#3, inner sep=1pt]{\ifnum\n=1 \n\,cm \else \n\fi};
}%%
\foreach \n in {0.1,0.2,...,\NoXdiv}{%%
\draw[] ([xshift=\n*\u]X) -- +(0,#5);
}%%
}

Anonymous

16:40

Anonymous

Ohhh ... I guess there is another \pgfmathsetlengthmacro causing problems :( :(

Anonymous

I have no idea how to solve these issues. this will take me all week! :D

Anonymous

17:10

@UlrichDiez I have asked a minimal question about it here: https://tex.stackexchange.com/questions/561612/why-can-i-not-newcommand-inside-a-tikzpicture-of-this-macro I seem to think that I can not use

\newcommand

inside the \NiceForEachElement loop ...

Ulrike Fischer

17:20

@VincentMiaEdieVerheyen \newcommand inside a loop is not a good idea, after all a loop is normally processed more than once, and so the second \newcommand doesn't define a new command anymore.

Anonymous

17:46

@UlrikeFischer Yaay you solved it :)
Now I can compile everything. Problems solved. good night sleep tonight ;)

Anonymous

Only have to figure out now how to make the right image not be the same as the left image. Instead the index of the image should become i+16 ... somehow.

Anonymous

% arara: pdflatex
% arara: pdflatex
% arara: pdflatex

\documentclass[a4paper, landscape]{article}
\pagestyle{empty}
\usepackage[showframe=false,
margin=20mm,
]{geometry}
%\geometry{paperwidth=15cm, paperheight=30cm, margin=2mm}% optional
%\usepackage{mwe} % Dummy Images
\usepackage{textcomp} % \textcopyright

\usepackage{tikz}
\usetikzlibrary{calc}
\pgfkeys{/tikz/savevalue/.code 2 args={\global\edef#1{#2}}}
\pgfdeclarelayer{background}
\pgfdeclarelayer{foreground}
\pgfsetlayers{background,main,foreground}

Joseph Wright

@UlrikeFischer Sure: we have an aim but it ay be a while

@PhelypeOleinik The change in message seems to be meaning I get absolute paths in the .tlg files :(

Ulrike Fischer

@JosephWright I pushed a new version of the "new-transparent.sty", it works but it also shows a bit what is still missing (it needs the resources for pdfmode, and the new dvipdfmx).

Joseph Wright

18:06

@UlrikeFischer Should I look to start l3backend-extgstate.dtx?

Ulrich Diez

@UlrikeFischer Afaik \foreach of tikz/pgf is an exception: Each iteration takes place within its own local scope/group, thus a \newcommand performed within an iteration will be gone with the local scope of that iteration...

@UlrikeFischer Afaik - if you do macro-definitions within tikz/pgf's \foreach, you need to double hashes.

Ulrike Fischer

@UlrichDiez yes I know I didn't go into the details and only wrote "normally" as it doesn't matter much why it actually failed: a \newcommand doesn't belong in a loop.

Anonymous

@UlrichDiez I have a compiling code (code 3) now here: https://tex.stackexchange.com/a/561482/67761 the only challenge that remains is to have the image on the right-hand side differ by a constant from the image on the left hand-side. That is:

\ImageRight

to have its index differ by a constant +16 of the counter/index of the left-hand side image.

Phelype Oleinik

18:51

@JosephWright Odd... The path should or should not show up either way (I think). Let me check

Ulrich Diez

@VincentMiaEdieVerheyen I just added an answer... Let me know if things work out as you intend them to.

Anonymous

19:13

@UlrichDiez Thank you. I had also created this mini-question in the mean time: tex.stackexchange.com/questions/561629/…

Anonymous

@UlrichDiez The constant in your answer tex.stackexchange.com/a/561628/67761 works great, but I feel the scaling is no longer working. Let me send you a screenshot.

Anonymous

This is what the output would look like without the constant-code, using code 3, or using code 2:

Anonymous

Using

\pgfmathsetlengthmacro{\u}{15.07425pt}% unitlength
\def\ImageScale{0.7}

Anonymous

But using the constant-code of your answer tex.stackexchange.com/a/561628/67761 it outputs:

Anonymous

19:18

Anonymous

The images are scaled too large.

Anonymous

19:33

I compiled as many times as I could (using pdflatexmk) but the scaling problem remains.

Anonymous

I wonder if we could make the code faster .... a very old version of this 2 picture page was this one: and it was so much faster: ->

Anonymous

\documentclass[a4paper, landscape]{article}
\usepackage[margin=0cm, showframe=false]{geometry}
\geometry{paperwidth=15cm, paperheight=30cm}
\usepackage{mwe} % Dummy images

\usepackage{comment}

\usepackage[skins]{tcolorbox}
\tcbuselibrary{raster}

\usepackage{eso-pic}
\newcommand\myvhcenter[1]
  {%
    \begingroup
      \sbox0{#1}%
      \raise.5\dimexpr\dp0-\ht0\relax\hbox to 0pt{\hss\usebox0\hss}%
    \endgroup
  }
\newcommand\myhcenter[1]{\makebox[0pt][c]{#1}}

\usepackage{tikz}
\usepackage{tikzpagenodes} % current page text area.center

Gaussler

20:01

What are the best practises for when to use \tl_if_eq:nnTF and when to use \str_if_eq:nnTF?

Phelype Oleinik

@Gaussler Only when differences in catcodes are important you use \tl_if_eq:nnTF, otherwise if you just want to know if two “strings” have the same characters (close to what other programming languages would consider as characters), use \str_if_eq:nnTF

Alan Munn

A hyperref question: are all the PDF metadata elements engine-independent? Or at least independent of {pdf|xe|lua}latex?

Ulrike Fischer

@AlanMunn ?

Alan Munn

@UlrikeFischer If I use \hypersetup{pdfauthor=Me, pdflang=..., pdfkeywords=...} does this work independent of the engine I'm using?

@UlrikeFischer We're now getting bugged about accessibility of documents we post for classes, and unfortunately TeX doesn't do this well yet. But some things can be done (whether they're useful or not is a separate issue.)

Ulrike Fischer

@AlanMunn yes, more or less, internally there are differences how they are encoded, and some engines add stuff, but the info should be there for all engines.

@AlanMunn I know, we are working on it. And if your document is not to complicated I could help you to make it accessible.

Alan Munn

20:14

@UlrikeFischer Unfortunately linguistics documents are not on the low end of complicated. Two especially tricky problems are example glossing (of the sort that cgloss4e or expex do) and trees, (also phonetic symbols, perhaps).

Ulrike Fischer

@AlanMunn can you use lualatex?

Alan Munn

@UlrikeFischer Yes, that's what I mainly use these days, and anything old can be adapted I'm sure.

@UlrikeFischer I've been reading the documentation to your tagpdf package, but it's a bit overwhelming. (Not a criticism, just there's lots to learn.)

Ulrike Fischer

@AlanMunn yes, but lots of stuff I wrote actually for me, so that I can look up details again. I would say the main question is if you need real accessibility or only need to pass some formal test. Which document class are you using normally?

Alan Munn

@UlrikeFischer Mostly article, but with titlesec. I have one textbook manuscript which uses memoir.

@UlrikeFischer As for passing tests, it's hard to tell whether we'll reach a point of being forced to pass tests. But I would be interested in real accessibility where possible. Although for grant applications PDF/A is increasingly required.

Phelype Oleinik

@JosephWright Hm, right, so that test picks up the l3build format file and the one in the system tree (which is good: the darn thing works :-). Not sure why it didn't pick that up before though...

@JosephWright I'm tempted to move that test file to -disabled for good. But to avoid being drastic, maybe just do \str_set:Nx \c_sys_engine_exec_str { zzz } in it to avoid any .fmt file being found.

Alan Munn

20:27

@UlrikeFischer For trees I assume the minimum requirement would be to patch forest to add an alt text for each tree, for example. Glossing seems to me to be the biggest difficulty. Partially because it's not clear to me how a visually impaired person would actually want to read a glossed example.

Ulrike Fischer

@AlanMunn pdf/A itself is not really a problem if one of the b-variants is enough. The problem is tagging. And real accessibility would need someone who can test and give feed backs.

Alan Munn

@UlrikeFischer I see. And yes, it's definitely hard to know what to do in the abstract.

Ulrike Fischer

@AlanMunn Useful would be examples of (accessible) web pages. The tagging code should be similar.

Alan Munn

@UlrikeFischer I have a former student (now a VP at Google) who is blind, and has spent his whole career working on accessibility issues. He might have some helpful pointers.

Ulrike Fischer

@AlanMunn that would be nice.

Alan Munn

21:08

@UlrikeFischer Are there any sample documents tagged with tagpdf other than the docs themselves?

Ulrike Fischer

@AlanMunn I have only a few small experiments. I know of a few people who seem to use it, but I with one exception I didn't see the results. (The last half year I haven't done much here, apart updating the package along with the coding of the pdfresource project.) But if you want to try you can ask me if you have questions -- and you should probably install the development version along with the pdfresource code.

barbara beeton

21:34

@AlanMunn -- Your former student would be an invaluable resource. Although different students may learn material in different ways, someone who is both blind and familiar with the subject matter should have valid insights into how to present material such as glosses and other linguistic notation. Do try to enlist his aid.

Alan Munn

21:54

@barbarabeeton I'll email him and see what he says.

@UlrikeFischer So the develop branch from your github repo? And what's the pdfresource code?

Ulrike Fischer

@AlanMunn no, you need the branch called "splitting", and the "splitting" branch from github.com/latex3/pdfresources. That is the newest stuff. And you need the newest latex-dev and should compile with lualatex-dev. So you are warned: it is everything quite experimental ;-) (and can still change ...) . Oh, and you should add the following before the \documentclass as it activates the pdfresource stuff.

\RequirePackage[patches]{pdfresources}
\DeclareDocumentMetaData{pdfversion=1.7,lang=en-UK}

Alan Munn

22:14

@UlrikeFischer Ok. By "newest latex-dev" do you mean the current prerelease as updated by tlmgr?

Ulrike Fischer

22:26

@AlanMunn yes, that should work, mostly the hook system is needed. (If you use unicode-math: load filehook-ltx before, or you will get errors, as filehook isn't adapted yet to latex-dev).

Alan Munn

@UlrikeFischer Ok thanks. I'll play around and ask you if I get stuck.

Ulrike Fischer

@AlanMunn you will perhaps need a validator, but you are on macos aren't you?

Alan Munn

22:57

@UlrikeFischer Yes, I have a full version of Acrobat which can do validation.

@UlrikeFischer I'm also curious to see what our course management systems validation does with documents. Right now everything I post gets 5% compliance score!

@UlrikeFischer But I'm not really interested in figuring out how to trick the system; it would be useful to people generally to learn how to tag linguistics documents properly.

@UlrikeFischer There are quite a lot of us linguists who use LaTeX.

Ulrike Fischer

@AlanMunn yes, only to trick the system wouldn't be worth the effort.but I don't know enough about the inner workings of the gloss packages to know how easy it would be to inject the structure.

And titlesec isn't the easiest package to patch either

Alan Munn

23:14

@UlrikeFischer Yes, I know. I was browsing the hyperref docs today and discovered that it doesn't work well with it either (I don't use hyperref too much, mainly just for inserting links, so I guess I'd never noticed.)

Transcript for

$TeX - LaTeX$ TeX, LaTeX and Friends