TeX, LaTeX and Friends

cis

00:46

@VincentMiaEdieVerheyen See some smaller corrections.

cis

02:16

Wissen Macht Frei

05:05

Anything related to duck must be posted here.

Anonymous

07:24

@cis Wonderful @cis. This is great!

Anonymous

I have included the \SeperationBetweenImages length here:

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

Very very happy with your code!

Anonymous

I wonder if I can adjust the \SeperationBetweenImages in a way that the text of the middle y-rulers overlap exactly, such that they do not need to be double (since my images are of equal height).

Anonymous

07:39

\pgfmathsetlengthmacro{\SeperationBetweenImages}{1.18cm}% seems to exactly overlap the 1cm

Anonymous

It also works if \def\ImageScale{0.51} is changed

Anonymous

But only while \def\ShowHelps{1}

Anonymous

if

\def\ShowHelps{0}

then the Seperation should be different

Joseph Wright

08:04

@UlrikeFischer Looks good

@DavidCarlisle Stats look good :0

cis

I don't test, but maybe it has to be

`\pgfmathsetlengthmacro{\SeperationBetweenImages}{1.18cm+\HelpBoxRule}`

or \pgfmathsetlengthmacro{\SeperationBetweenImages}{1.18cm-\HelpBoxRule}

while `\def\ShowHelps{1}` is used.

Anonymous

08:44

@cis Yes, you are completely right!

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

Now the 1cm is always matched. Now the matching of the 0, 2, 3 and so on remains :)

Anonymous

I guess it will require an extra definition of an \yRuler, perhaps \yRulerBis

Anonymous

09:08

I made a small question about it here: tex.stackexchange.com/questions/561458/…

David Carlisle

@JosephWright S columns :-)

Ulrike Fischer

09:30

@DavidCarlisle ?

David Carlisle

09:47

@UlrikeFischer secrets

Paulo Cereda

@DavidCarlisle ooh

Ulrike Fischer

@PauloCereda and as he is not a duck ...

David Carlisle

@UlrikeFischer you've got mail

Ulrike Fischer

10:13

@DavidCarlisle ;-)

Ulrike Fischer

10:23

@JosephWright yes, quite nice. But it relies on extractbb to be available, with an unknown application you get a "file not found error".

Joseph Wright

@DavidCarlisle >

@UlrikeFischer Well it's the same binary as dvipdfmx, so it should be fine for testing for the dvipdfmx version :)

Ulrike Fischer

@JosephWright yes, but I was wondering if it is safe to detect the version in the format to avoid that every document gets delayed by calling extractbb. dvipdfmx perhaps doesn't exist in minimal tex installations.

Joseph Wright

@UlrikeFischer If you are worrying about non-standard set ups, the format is likely too 'early'

Ulrike Fischer

@JosephWright could one have a \sys_get_shell:nnN that doesn't error if an unknown or not allowed command is used, but gives back a "false"?

Joseph Wright

10:39

@UlrikeFischer There's a TF version

Ulrike Fischer

@JosephWright that only tests if shell-escape is disabled (and if there are quotes), it doesn't catch the case of a typo or missing application.

Joseph Wright

@UlrikeFischer Hard error I think

David Carlisle

@UlrikeFischer I would just have the new code as the dvipdfmx back end and (for testing) have the old code as a separate driver [olddvipdfmx] or whatever, it can set a boolean then use the same file so I don't think there would be a big cost to this. Outside of initial developer testing dvipdfmx only gets updated by th etex system update and that can arrange teh right def file

Ulrike Fischer

@JosephWright this here works fine: so it should it be possible to catch it:

\documentclass{article}
\begin{document}
\newread\mypipe
\openin\mypipe="|extractbbXXX~--version"
\loop
  \ifeof\mypipe
  \show\blub
  \else
    \begingroup
      \endlinechar=-1 % suppress space at line end
      \read\mypipe to \x
      \texttt{\x}\par
    \endgroup
\repeat
\end{document}

David Carlisle

@UlrikeFischer but that requires openin which is fine in this case but if you have wanted \input{|zzzz you'd have to do openin first (like inputiffileexists) but then you run the command twice which may not be desirable

Ulrike Fischer

10:58

@DavidCarlisle yes, I assume we can arrange something. But actually I'm more pondering about the general question. I mean we have tests for versions everywhere "if luatex < X", "if package version < 2020/03/04", "IfFormatLater" and so on. But imho only biblatex actually checks that its external tool biber matches.

@DavidCarlisle if we do it in the format, the extra test would probably not really matter.

David Carlisle

@UlrikeFischer certainly but as Joseph says if it is in the format it catches less developer testing cases where you have both installed and are juggling paths.

@UlrikeFischer but where do you draw the line at testing external applications? If you generate a valid dvi that requires a new dvipdfmx to process it but you have an old dvipdfmx, is that any different from generating a PDF 2.0 document when your pdf viewer only understands pdf 1.5 ?

Joseph Wright

@DavidCarlisle Tricky: stuff 'normally' allowed in the shell is side-effect free

Ulrike Fischer

@DavidCarlisle in my setup it would work: the xetex binaries are different too, so I have two format files. But for a developer setup it is not really needed. I mean I can simply add a file somewhere with the version. The question is if we will need a version check in distributed code to issue at least a warning, that something won't work (assuming that nobody reads the documentation ...).

Joseph Wright

@UlrikeFischer Really, we have the same with other stuff, and it's simple: we don't update macros in release versions until TL has the binaries

David Carlisle

@JosephWright if you did this with repstopdf it would mean that you checked epstopdf is available by actually converting a file (but that's a bad example as you wouldn't normally use that with pipe, which is perhaps what you meant)

Joseph Wright

11:09

@DavidCarlisle Ah,true

@DavidCarlisle Probably not one we can build-in

@UlrikeFischer Really, I feel you are over-worrying

Anonymous

@cis I am trying to adapt the code based on your MWE answer.

Anonymous

Currently I have:

Anonymous

\draw[Ruler] (#1) coordinate(Y) -- (#2);
\foreach \n in {0,1,...,\NoYdiv}{%%
\ifnum\n=0 \ifnum\ifzero=1 \draw[Ruler] ([yshift=\n*\u]Y) -- +(#4+1cm,0) node[#3, inner sep=1pt]{\n};
\else
\ifnum\n=1 \draw[Ruler] ([yshift=\n*\u]Y) -- +(#4,0) node[#3, inner sep=1pt]{ \n\,cm}{
\else
\draw[Ruler] ([yshift=\n*\u]Y) -- +(#4+1cm,0) node[#3, inner sep=1pt]{ \n}{
\fi}
\fi}
}

Anonymous

But it is not compiling.

Anonymous

Maybe it should be (also not compiling) -> ?

Anonymous

11:17

\draw[Ruler] (#1) coordinate(Y) -- (#2);
\foreach \n in {0,1,...,\NoYdiv}{%%
\ifnum\n=0 \ifnum\ifzero=1 \draw[Ruler] ([yshift=\n*\u]Y) -- +(#4+1cm,0) node[#3, inner sep=1pt]{\n};
\else
\ifnum\n=1 \draw[Ruler] ([yshift=\n*\u]Y) -- +(#4,0) node[#3, inner sep=1pt]{ \n\,cm};
\else
\draw[Ruler] ([yshift=\n*\u]Y) -- +(#4+1cm,0) node[#3, inner sep=1pt]{ \n}
\fi}
\fi}

David Carlisle

@VincentMiaEdieVerheyen I don't know what this code is doing, but \ifnum\ifzero=1 \draw[Ruler] can not possibly work. \ifnum needs a number then = then another number, but if \ifzero is true it gets =1\draw

@VincentMiaEdieVerheyen also you open 4 \if.. but only close two \fi

Anonymous

@DavidCarlisle I guess I closed them now:

Anonymous

\draw[Ruler] (#1) coordinate(Y) -- (#2);
\foreach \n in {0,1,...,\NoYdiv}{%%
\ifnum\n=0 \ifnum\ifzero=1 \draw[Ruler] ([yshift=\n*\u]Y) -- +(#4+1cm,0) node[#3, inner sep=1pt]{\n} \fi;
\else
\ifnum\n=1 \draw[Ruler] ([yshift=\n*\u]Y) -- +(#4,0) node[#3, inner sep=1pt]{ \n\,cm} \fi;
\else
\draw[Ruler] ([yshift=\n*\u]Y) -- +(#4+1cm,0) node[#3, inner sep=1pt]{ \n} \fi;
\fi};

Anonymous

@DavidCarlisle I am not sure if I understand this.

David Carlisle

@VincentMiaEdieVerheyen you still have \ifnum\ifzero=1 \draw[Ruler] which is always going to fail.

@VincentMiaEdieVerheyen if ifzero is false your second ifnum test is \ifnum\else which is a syntax error. What do you want to test? Also where are you setting \ifzero to true or false?

Anonymous

11:33

@DavidCarlisle ifzero determines whether or not the 0 gets printed ... tex.stackexchange.com/questions/561458/… but in my case, in fact, I always want to have the 0 printed ... so perhaps I can take out this check/test.

David Carlisle

@VincentMiaEdieVerheyen but if it is false, everything from there to the matching fi is skipped, so then your \ifnum has no content so is a syntax error, conversely if it is true the \ifnum` has =1 which is also a syntax error. there are no conditions under which \ifnum\ifzero=1 \draw[Ruler] can do anything other than error.

Anonymous

I see

David Carlisle

@VincentMiaEdieVerheyen which numbers did you want the ifnum to compare?

Anonymous

I still can't compile the following ... but I guess the syntax should be kind of correct now, as long as you don't know the context ;) ?

Anonymous

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

Anonymous

11:38

@DavidCarlisle In fact I only want to check for the number 1 ... which should be of different length than all the other numbers.

David Carlisle

@VincentMiaEdieVerheyen that has an \else not in an \if so is an error even without looking at the tikz

Anonymous

OK @DavidCarlisle understood. Now I adjusted the code and it is compiling fine! :)

Ulrike Fischer

@JosephWright you mean I got a but of H.O. infection? ;-) Could be, we will see, at least we have a plan, in case we need and if not it is fine too.

Anonymous

@DavidCarlisle What I need to do now is to have the +1cm become +1cm or -1cm based on the negativity or postivity of the value in #4, #4 being somewhere else in the code, as you understand.

David Carlisle

@UlrikeFischer no you didn't save and restore the catcodes of n and C before using them in that delimited argument. You are still OK,

@VincentMiaEdieVerheyen ? \ifnum#4>0 +1cm\else -1cm\fi ?

Anonymous

11:51

@DavidCarlisle I am convinced that your method would work, thank you, but I can't compile:

Anonymous

\draw[Ruler] (#1) coordinate(Y) -- (#2);
\foreach \n in {0,1,...,\NoYdiv}{%%
\ifnum\n=1 \draw[Ruler] ([yshift=\n*\u]Y) -- +(#4,0) node[#3, inner sep=1pt]{ \n\,cm};
\else
\ifnum#4>0 \draw[Ruler] ([yshift=\n*\u]Y) -- +(#4+1cm,0) node[#3, inner sep=1pt]{\n};
\else
\draw[Ruler] ([yshift=\n*\u]Y) -- +(#4-1cm,0) node[#3, inner sep=1pt]{\n};\fi;
\fi;
};

David Carlisle

@VincentMiaEdieVerheyen well what is #4 , is it a number or a length, it's probably a length, so use \ifdim#4>0pt instead

Anonymous

True, it is a length.

David Carlisle

@VincentMiaEdieVerheyen also ;\fi; looks wrong will leave you with double ;;

Anonymous

When I try to compile:

Anonymous

11:55

\draw[Ruler] (#1) coordinate(Y) -- (#2);
\foreach \n in {0,1,...,\NoYdiv}{%%
\ifnum\n=1 \draw[Ruler] ([yshift=\n*\u]Y) -- +(#4,0) node[#3, inner sep=1pt]{ \n\,cm};
\else
\ifdim#4>0 \draw[Ruler] ([yshift=\n*\u]Y) -- +(#4+1cm,0) node[#3, inner sep=1pt]{\n};
\else
\draw[Ruler] ([yshift=\n*\u]Y) -- +(#4-1cm,0) node[#3, inner sep=1pt]{\n};\fi
\fi;
};

Anonymous

I get the error:

Illegal unit of measure (pt inserted).

Anonymous

#4

is either -4mm or 4mm

David Carlisle

@VincentMiaEdieVerheyen the error message gives more information than that.

Anonymous

@DavidCarlisle It only comes up when adding the \ifdim#4>0 it seems ...

David Carlisle

12:01

@VincentMiaEdieVerheyen which is why I said to use \ifdim#4>0pt you are comparing a length to an integer

Anonymous

Aaaha, I see! Thanks

Anonymous

@DavidCarlisle Looking at the compiled results, actually I think it would look more elegant to just shift the numbers, not shift the ruler-sticks 😅

Anonymous

So I fiddle with the hspace ... it's a bit strange that this worked perfectly for the 0.25cm which is positive, but I had to substract a lot more space, for the negative value (-0.72)

Anonymous

\draw[Ruler] (#1) coordinate(Y) -- (#2);
\foreach \n in {0,1,...,\NoYdiv}{%%
\ifnum\n=1 \draw[Ruler] ([yshift=\n*\u]Y) -- +(#4,0) node[#3, inner sep=1pt]{ \n\,cm};
\else
\ifdim#4>0pt \draw[Ruler] ([yshift=\n*\u]Y) -- +(#4,0) node[#3, inner sep=1pt]{\hspace{0.25cm}\n};
\else
\draw[Ruler] ([yshift=\n*\u]Y) -- +(#4,0) node[#3, inner sep=1pt]{\hspace{-0.72cm}\n};\fi;
\fi;
};

Anonymous

12:12

Anonymous

Next challenge is to remove the number duplicates, because 2 rulers are drawn:

Anonymous

\yRulerBis{LLM}{ULM}{right}{4mm}{2mm}
\yRulerBis{LRM}{URM}{left}{-4mm}{-2mm}

Anonymous

I was able to remove the duplicates of the 0 and the 2 and the 3 and so forth easily,

Anonymous

But I don't know how to remove the duplicate of the '1cm', which is printed twice now:

Anonymous

\draw[Ruler] (#1) coordinate(Y) -- (#2);
\foreach \n in {0,1,...,\NoYdiv}{%%
\ifnum\n=1 \draw[Ruler] ([yshift=\n*\u]Y) -- +(#4,0) node[#3, inner sep=1pt]{\n\,cm};% This is drawn twice now ... I need to find a way to remove the duplicate
\else
\ifdim#4>0pt \draw[Ruler] ([yshift=\n*\u]Y) -- +(#4,0) node[#3, inner sep=1pt]{\hspace{0.25cm}\n};
\else
%\draw[Ruler] ([yshift=\n*\u]Y) -- +(#4,0) node[#3, inner sep=1pt]{\hspace{-0.72cm}\n};\fi;% Removed duplicate
\draw[Ruler] ([yshift=\n*\u]Y) -- +(#4,0);\fi;

Anonymous

12:25

Should be some other positivity check again.

Anonymous

I think it could be something like this, but it doesn't compile:

Anonymous

\draw[Ruler] (#1) coordinate(Y) -- (#2);
\foreach \n in {0,1,...,\NoYdiv}{%%
\ifnum\n=1 \ifdim#4>0pt \draw[Ruler] ([yshift=\n*\u]Y) -- +(#4,0) node[#3, inner sep=1pt]{\n\,cm};\fi;
\else
\ifnum\n=1 \draw[Ruler] ([yshift=\n*\u]Y) -- +(#4,0);\fi;
\else
\ifdim#4>0pt \draw[Ruler] ([yshift=\n*\u]Y) -- +(#4,0) node[#3, inner sep=1pt]{\hspace{0.25cm}\n};
\else
%\draw[Ruler] ([yshift=\n*\u]Y) -- +(#4,0) node[#3, inner sep=1pt]{\hspace{-0.72cm}\n};\fi;% Removed duplicate
\draw[Ruler] ([yshift=\n*\u]Y) -- +(#4,0);\fi;

Anonymous

Is it possible @DavidCarlisle to jus put two checks after one another?

\ifnum\n=1 \ifdim#4>0pt

Ulrike Fischer

@DavidCarlisle ;-). But now that you mention catcodes I realize that I misread the docu, \sys_get_shell doesn't give back a str but a tl.

Anonymous

I guess I found it :) No worries.

Anonymous

12:34

Only now, if it becomes 2 digits, then the numbers are no longer centered.

Anonymous

cis

You can make the number output an optional argument; sth. like

\newcommand\mycmd[5][\n]{n=#1 ......... #2 ... #3 .... #4 ..... #5 .....}

\def\n{111}

\mycmd{2}{3}{4}{5} ---> n=111 ......... 2 ... 3 .... 4 ..... 5 .....

\mycmd[]{2}{3}{4}{5} ---> n= ......... 2 ... 3 .... 4 ..... 5 ..... ---> (void)

Anonymous

@cis Wanted to share this code, out of enthousiasm. I am happy I now centered the 2 digit numbers in the middle. ;D Now I will look at your suggestion, I am curious to understand it.

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}

David Carlisle

@UlrikeFischer yes which is why all the spaces went:-)

Anonymous

12:46

@cis I see. Your suggestion is to let the numbers stop appearing at some point, is it?

cis

If it gives the desired result...

I see: node[#3, inner sep=1pt]{\hspace{0.16cm}\n}
this is not how to use TikZ.

----> node[#3, inner sep=1pt, xshift=0.16cm]{\n}

Anonymous

@cis Thanks. Will correct.

David Carlisle

13:07

@VincentMiaEdieVerheyen not exactly wrong but using cm lengths with fonts looks very weird \hspace{0.25cm}\n compared to \phantom{0}\n which is I think what you want here. (different for the actual rulers as there you want real-world units not font dependent lengths.)

Anonymous

13:24

cis found the TikZ solution:

node[#3, inner sep=1pt, xshift=0.16cm]{\n}

Would that be OK, or do you still mean cm looks weird, should I change it into pt then?

Anonymous

Aaah you mean change it into

node[#3, inner sep=1pt, xshift=\phantom{0.16}]{\n}

?

Torbjørn T.

@VincentMiaEdieVerheyen That won't work, xshift needs a length.

@VincentMiaEdieVerheyen David meant node[#3, inner sep=1pt]{\phantom{0}\n}

Anonymous

I think the \phantom{0} (as far as I understand it is empty) is not really necessary. ... I did need a shift in the x coordinate.

David Carlisle

@VincentMiaEdieVerheyen it looked like you were trying to align 9 and 10 by shifting single digit numbers

Anonymous

@DavidCarlisle Aaah yes, indeed, interesting! Will try that.

David Carlisle

13:31

\phantom{0} is a box the width of a 0 so will work whatever font you use, if you shift by a fixed length you are just doing it by eye and it will be wrong if you change fonts

Anonymous

Perfect.

Torbjørn T.

@VincentMiaEdieVerheyen Having looked briefly, I'd do node[#3, inner sep=1pt, minimum width=1cm]{\n}; for both 1 and 2 digit numbers, with an appropriate value for the minimum width. (1cm is a bit much, perhaps 2em is ok.)

Anonymous

@cis Maybe that could be very interesting if the Images were of different heights?

I posted the small adaptations of your code here as well, for our reference: https://tex.stackexchange.com/a/561482/67761

Anonymous

13:55

OK thanks, I have tried to compare with the first digits, and I think

minimum width=2.15em

would get them in the middle.

David Carlisle

@VincentMiaEdieVerheyen you can right align the text in the node

Anonymous

@DavidCarlisle I had:

Anonymous

\foreach \n in {0,1,...,\NoYdiv}{%%
\ifnum\n=1 \ifdim#4>0pt \draw[Ruler] ([yshift=\n*\u]Y) -- +(#4,0) node[#3, inner sep=1pt]{\n\,cm};
\else
\draw[Ruler] ([yshift=\n*\u]Y) -- +(#4,0);\fi
\else
\ifdim#4>0pt \ifnum\n<10 \draw[Ruler] ([yshift=\n*\u]Y) -- +(#4,0) node[#3, inner sep=1pt, xshift=0.25cm]{\n};
\else
\draw[Ruler] ([yshift=\n*\u]Y) -- +(#4,0) node[#3, inner sep=1pt]{\phantom{0}\n};
\fi
\else
\draw[Ruler] ([yshift=\n*\u]Y) -- +(#4,0);\fi;
\fi;
};

Anonymous

Now I have:

Anonymous

\foreach \n in {0,1,...,\NoYdiv}{%%
\ifnum\n=1 \ifdim#4>0pt \draw[Ruler] ([yshift=\n*\u]Y) -- +(#4,0) node[#3, inner sep=1pt]{\n\,cm};
\else
\draw[Ruler] ([yshift=\n*\u]Y) -- +(#4,0);\fi
\else
\ifdim#4>0pt \ifnum\n<10 \draw[Ruler] ([yshift=\n*\u]Y) -- +(#4,0) node[#3, inner sep=1pt, minimum width=2.15em]{\n};
\else
\draw[Ruler] ([yshift=\n*\u]Y) -- +(#4,0) node[#3, inner sep=1pt, minimum width=2.15em]{\n};
\fi
\else
\draw[Ruler] ([yshift=\n*\u]Y) -- +(#4,0);\fi;
\fi;
};

Anonymous

13:58

@DavidCarlisle Not sure what to right align.

Anonymous

Ah, you mean the double digits?

Anonymous

I am a bit confused now, as to what is the best method to achieve centered numbers.

David Carlisle

@VincentMiaEdieVerheyen if you right aligned this node node[#3, inner sep=1pt, minimum width=2.15em]{\n}; you would not need the test for < 10

@VincentMiaEdieVerheyen or center if you prefer that.

Anonymous

14:16

I am feeling lost with all these ifs ... I guess I should comment the code for my own sake.

Anonymous

@DavidCarlisle I have now removed the test < 10:

Anonymous

\foreach \n in {0,1,...,\NoYdiv}{%%
\ifnum\n=1 \ifdim#4>0pt \draw[Ruler] ([yshift=\n*\u]Y) -- +(#4,0) node[#3, inner sep=1pt]{\n\,cm};% 1cm (from left image): bar + text
\else
\draw[Ruler] ([yshift=\n*\u]Y) -- +(#4,0);\fi% 1cm (from right image): bar
\else
\ifdim#4>0pt \draw[Ruler] ([yshift=\n*\u]Y) -- +(#4,0) node[#3, inner sep=1pt, minimum width=2.15em]{\n};% all numbers lower than 10 (from left image), excluding 1: bar + text
\else
\draw[Ruler] ([yshift=\n*\u]Y) -- +(#4,0);\fi;% all numbers (from right image), excluding 1: bar

Anonymous

But I find that align=right creates no difference?

Anonymous

15:03

@cis Maybe it is worth testing the answer using these images

Anonymous

cis

@VincentMiaEdieVerheyen \includegraphics[width=300px, height=700px]{example-image}

For colored pics

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

Marcel Krüger

20:34

@DavidCarlisle @UlrikeFischer I fixed the bug which introduced the empty lines in the l3 banner. We could push out a hotfix release if this is considered a significant problem.

David Carlisle

@MarcelKrüger I doubt it's a problem at all to anyone except it bothered me:-) thanks for fixing it though, can go out next tine

Anonymous

@cis Thanks.

Anonymous

@cis Next challenge is to have multiple pages and loop through them. So, I am curious now to try out your page=2 setting here :)

Anonymous

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

Torbjørn T.

@VincentMiaEdieVerheyen If you include a PDF with multiple pages with \includegraphics, it should.

Anonymous

20:51

@TorbjørnT. Trying to achieve something as such:

Anonymous

\begin{document}

\foreach \i in {\PictureIndexStartsAt,...,\PictureIndexEndsAt}{ %%%% Looping through these pictures
\foreach \image/ in {
{./Pictures/Image-\i}
}{%%%%%%%%%%%%%%%


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



\tcbitem[......] \UseImage
\tcbitem[......] \UseImage



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



\end{tikzpicture}

\newpage
}%%%%%%%%%%%%%%%%%%%%%%%%

\end{document}

Anonymous

But I can't compile it. ...

Anonymous

I don't know why ... but I can't even compile the following:

Anonymous

\begin{document}

\foreach \i in {\PictureIndexStartsAt,...,\PictureIndexEndsAt}{ %%%% Looping through these pictures

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


\end{tikzpicture}

\newpage
}%%%%%%%%%%%%%%%%%%%%%%%%

\end{document}

Anonymous

I would like to achieve this with the bottom code at tex.stackexchange.com/a/561482/67761

Torbjørn T.

20:58

@VincentMiaEdieVerheyen Too late in the evening for me to look much into that now.

Anonymous

OK no worries, have a good sleep :)

Ulrike Fischer

21:10

@MarcelKrüger looks better, but imho we should also try to shorten the messages and get rid of the descriptions. @DavidCarlisle should the message in ltluatex be changed or should luaotfload define its own function?

Marcel Krüger

@UlrikeFischer I agree, I am working towards that right now. (Especially since I believe that the individual module messages are meaningless for almost everyone anyway) I already switched the code to use a custom function, the next step is not calling luatexbase.provide_module there and instead doing a version test.

Ulrike Fischer

@MarcelKrüger yes, I don't think that the descriptions are important (and one can always look in the file), I'm really only interested to see file names and a version - it helps to find an error if you know which files have been loaded.

@MarcelKrüger and I see it is already called 3.16-dev ;-)

Marcel Krüger

@UlrikeFischer My current approach would drop the file names too, but I can add them back. But why would they be needed? After all, the code always loads all of the luaotfload files anyway, doesn't it?

Ulrike Fischer

@MarcelKrüger well you could the same for many tex files. Why should anyone care to have all the pgf-files in the log? But actually it helps. I often have a test version of some file in the local folder and change it, and sometimes it gets forgotten, and the same is true for other: many people copy styles and similar files to their document folder and end with some mix.

Marcel Krüger

21:28

@UlrikeFischer Yes, but do determine that we would need the full path and currently we only have the module name. Knowing that we load "luaotfload-harf-define" does not really tell allow me to differentiate if it is the texmf version/a local file/etc. But most of these issues should be solved by checking the version number anyway. (Basically: If a module has a different version than the main file, complain and show exactly which modules are involoved)

David Carlisle

@UlrikeFischer I think ltluatex could change if it's a general issue, what would you do, not print the description by default?

Ulrike Fischer

@MarcelKrüger hm, normally if I change a file locally to debug something I change the version or the file description so that I can see in the log that it is not the original. If this would error it would no longer be possible to do such tests.

Marcel Krüger

@UlrikeFischer Maybe issueing a warning instead of an error then? This could also be printed to the terminal to make it more obvious that changes code is used.

Ulrike Fischer

21:43

@MarcelKrüger yes that would be perfect.

@DavidCarlisle I don't think that it is really needed. But it would affect other packages (e.g. luatexja and luatexko), so perhaps a new "short" function is better? or an additional long function for the package that miss the description?

Marcel Krüger

@UlrikeFischer I don't think a separate function is necessary here. If a package does not want to print the description, it can simply omit the description field and write a description in long_description or something. Then the metadata is still saved but not printed.

Ulrike Fischer

@MarcelKrüger I meant the other way round: if we no longer print the description but some package absolutely want it in the log. But perhaps we should simply wait if someone complains.

Marcel Krüger

22:09

@UlrikeFischer I pushed a draft for this: It does not register submodules (so the only logged module is luaotfload as a whole) but if versions are inconsistent, you get

Ulrike Fischer

which lead to the inner family discussion if one should use the new hic sunt dracones or the roman hic sunt leones ;-)

@MarcelKrüger can one make a new line after the "at"s?

Marcel Krüger

@UlrikeFischer Definitely "dracones". I always wanted to meet a dragon.

@UlrikeFischer Yes, but it will still have the (luaotfload) prefix.

Ulrike Fischer

22:26

Nov 11 '19 at 14:57, by Ulrike Fischer

@MarcelKrüger imho it will look better as this line breaks in the middle of a path

Marcel Krüger

@UlrikeFischer Done, but e.g. the /home/ path is still broken.

Ulrike Fischer

@MarcelKrüger imho it is fine for normal cases, less wordy. Is there an option to get a full list of all lua files if needed? E.g. by setting a debug level in the conf-file?

@MarcelKrüger yes, in some cases it is unavoidable, some of my pathes e.g. are always too long.

@MarcelKrüger depends on the dragon ;-) I have read quite a number of books with dragons over the years and sometimes there were quite vicious.

Marcel Krüger

22:44

@UlrikeFischer The issue with that is that some files are already read when the config file gets loaded. Of course we can always load a list of all luaotfload modules after the fact with something like for k,v in next, package.loaded do if k:match"^luaotfload-" then print(k) end end which we could run at the end of luaotfload.lua or we can restrict this to the modules loaded after reading the configuration. Which modules these are might change soon though.

Ulrike Fischer

@MarcelKrüger and what about an environment variable? Say luaotfloaddebuglevel or options (MiKTeX has this for its tracng). That can be read early and is easily set and unset.

Transcript for

$TeX - LaTeX$ TeX, LaTeX and Friends