TeX, LaTeX and Friends

8:14 AM

@barbarabeeton some packages have code that is so transparently clear that no documentation is necessary.

8:37 AM

@DavidCarlisle I particularly like code parts starting with

\begingroup\expandafter\expandafter\expandafter\endgroup
\expandafter\ifx\csname pdf@escapestring\endcsname\relax

@JosephWright didn't xetex get \pdfescapestring?

@UlrikeFischer No, it's not in (u)pTeX either so I did not do anything about it. As we can do it from the macro layer, I'm reluctant to ask/work on more primitives

@UlrikeFischer that's easy to read it is read as "Heiko's if undefined" :-)

@DavidCarlisle ;-).

@JosephWright ah, this means that if the experimental driver should work in the end with xetex too, I either need to put Heiko's code back, or find some alternative which does

\def\Hy@pstringdef#1#2{%
    \begingroup
      \edef~{\string~}%
      \xdef\Hy@gtemp{\pdf@escapestring{#2}}%
    \endgroup
    \let#1\Hy@gtemp
  }%

But it looks as if it should work with l3str-convert.

9:03 AM

@UlrikeFischer we should replace them with \@ifundefined (which in its new \ifcsname form is a better test anyway.

@DavidCarlisle I'm doing it already, when I get to one of them, at least in the driver code.

How does one do \edef~{\string~}% in expl3?

9:17 AM

@UlrikeFischer Yes, it should

@UlrikeFischer It's \char_set_active_eq:NN, with a suitable throw-away definition

@JosephWright yes just found it and pondering about the "throw-away definition".

:50371072 why \~?

@UlrikeFischer Actually, that won't work ...

\char_set_catcode_other:N \~
\char_set_active_eq:NN ~ \c_tilde_str

@UlrikeFischer ^^^ Can't have ~ as an N-type argument on it's own

@UlrikeFischer What's that code actually for?

@UlrikeFischer if ~ was etex \protected you wouldn't need to stringify it before the edef would you?

@JosephWright It's used in a lot of places, e.g. url, destination names, form fields.

@DavidCarlisle looks like it, with \protected\def~{\nobreakspace{}} I can use it directly in \edef\tmp{\pdfescapestring{abc~xyz}}\show\tmp.

9:32 AM

@UlrikeFischer Right, but it should be covered by l3str-convert

@JosephWright nice, \convert\tmp{ Café~some } { } { utf16/string } seems to do the right thing without any need to redefine ~.

@UlrikeFischer yes but question as always is do you want to pre-expand, so \def\foo{Café} then convert \foo~some if you want to expand \foo but not ~ there has to be some expansion control somewhere

9:50 AM

@DavidCarlisle I can find things like \Hy@pstringdef\Hy@pstringName{\HyperDestNameFilter{##3}}, so expanding is probably necessary.

4 hours later…

daleif

1:33 PM

Here is a similar example of that tikz vs spaceskip issue

\RequirePackage{fix-cm}
\documentclass[a4paper]{article}
\usepackage{tikz}
\begin{document}

\begin{tikzpicture}

\node[text width=15cm] at (0,0) {
\fontsize{25mm}{27mm}\sffamily\raggedright
15th Congress for\par
};

\node[text width=15cm] at (0,-8) {
\fontsize{25mm}{27mm}\sffamily\raggedright
\spaceskip0pt%
\xspaceskip0pt%
15th Congress for\par
};
\end{tikzpicture}
\end{document}

When I'm doing something fast, I never consider adding the font stuff as options to the node.

0xC0000022L

2:03 PM

Would someone be so kind to fix the link in this answer: tex.stackexchange.com/posts/151806/edit ... following line:

_You may also like to read further in my german blog: [TeX-Beispiel des Monats „Layout II: Kopf- und Fußzeilen“](http://tobiw.de/tbdm/layout-2)._

... where the “ ended up in the URL field instead of the link text. It's below the silly 6-character-minimum since I lack the reputation on this SE site. (Mine already contains the necessary correction)

Phelype Oleinik

2:23 PM

@0xC0000022L Done :-)

0xC0000022L

Thanks

2 hours later…

4:31 PM

any professional journal typesetters here?

5:02 PM

@EmilioPisanty Not at the moment, but you can ask.

@egreg my journal of choice has turned all my \hats into \widehats

the proofs system allows me to edit the TeX for the equations, so I just reverted all of them back to \hats

is this the kind of thing that's going to get the typesetter super annoyed at me?

it was mostly $\widehat{\mathbf u}$, $\widehat{\mathbf e}$, $\widehat{\mathcal G}$ type of stuff

@EmilioPisanty Possibly. Apparently they're the kind of people who thinks that large things look better (they don't).

@egreg thanks =).

I mean, I'm right, right?

@EmilioPisanty Definitely

I'll just roll those dice and see what happens

thx =)

5:08 PM

@EmilioPisanty ^^^^^ Nobody should have doubts, except, perhaps, for the calligraphic G.

@egreg maybe they really like Asian conical hats?

@EmilioPisanty That might be the reason! :-D

there was also a $\tilde{\mathbf E}$ to $\widetilde{\mathbf E}$, which, seriously

@egreg bf, not bb

5:12 PM

@EmilioPisanty ^^^ No doubt either that the left one is better. With \mathbf it's essentially the same.

hmmm. $\widetilde{\mathbf E}$ vs $\displaystyle \widetilde{\mathbf E}$

dunno

at the very least it looks different on MathJax

ultimately what matters is the journal pdf, which isn't super sexy either

@EmilioPisanty It is font dependent, that's true.

the journal pdf

5:39 PM

@EmilioPisanty WOW!

Phelype Oleinik

5:56 PM

@EmilioPisanty The \sim above the E looks an eyebrow and makes the E looks really angry with what they done to him :-)

StrongBad

6:19 PM

@EmilioPisanty why not just do \let\widehats\hats and then they can decide if they really want it

1 hour later…

7:21 PM

@StrongBad if only.....

2 hours later…