@DavidCarlisle Would be nice if you had to use a special flag to use obsolete things. 'pdflatex --old' or something, that also turned off everything SINCE that version.

@Canageek it wouldn't have worked. It took a lot of persuasion (and some years) to get people to switch over to 2e, "almost everything will still work" was a major selling point. Also basically it's not possible to hide (in 2e) the majority of things that you shouldn't use eg plain tex and primitive commands.

@DavidCarlisle Sigh

@DavidCarlisle Would you hide \ooalign? ;-)

Morning quack!

It's so amusing the fact that using TeX to write scientific papers somehow gives more confidence to the presented content. Math formulas, environments, fonts, alignment, TikZ...

@egreg If it would annoy you I'd undefine \ooalign and make % have catcode 12.

@DavidCarlisle of course, you could distribute pdflatex09 etc....

@PauloCereda You remind me of: one can't unsee http://uncyclopedia.wikia.com/wiki/Duct_tape#See_also:

@yo' oh no

@PauloCereda How much duck tape is needed to write a thesis?

@yo' o.O

I was not expecting this question. :)

^^ I wonder how long would a thesis PDF be :-)

@yo' :)

@yo' That makes me think: what would happen if we use \usepackage{forest} and \usepackage{savetrees}? :)

and also: en.wikipedia.org/wiki/File:Book_music_playing.jpg

@PauloCereda that's like having a right-wing party together with the greens :-)

@yo' Ah, a classical merry-go-round organ :)

@yo' :)

@JosephWright I've seen a girl in France whose family is still building them, moreorless to keep the tradition up!

@Joseph, @yo' I have a punched card as a souvenir. Ironically, the coded instruction is exit. :)

@PauloCereda I've got some as well...

@yo' That's what I used when I started

@DavidCarlisle, @egreg Would appreciate any thoughts on github.com/josephwright/etoolbox/issues/9. My feeling is the current behaviour is wrong, but before I change it I want to be sure I'm not missing something. (Particularly as Frank has decided etoolbox is acceptable as kernel-like for bug reports!)

@DavidCarlisle, @egreg Need to remind myself of all of the possible definitions of robust commands first ...

@JosephWright yes bit odd that \newcommand[][] makes things robust rather differently than declarerobustcommand, but done at different times:-)

@DavidCarlisle The \newcommand version is less complex, yes? But am I right thinking the logic in etoolbox is wrong: seems odd but ...

@JosephWright will look later need to make my program not segfault first:-(

@DavidCarlisle :)

@DavidCarlisle C?

@PauloCereda well xslt generating C from XML which calls fortran which is then compiled to a mex file which is called from matlab, which then calls the mex to call the C to call the fortran which then executes one of the arguments as a callback function which calls a C function which marshalls the arguments and calls a matlab function which passes back to the C and segfaults, so yes.

@DavidCarlisle Fortran programs never

Program received signal SIGSEGV: Segmentation fault - invalid memory reference.

Backtrace for this error:
#0  0x10542ee42
#1  0x10542f60e
#2  0x7fff8d7895a9
#3  0x10542575e
#4  0x105425975
#5  0x105425d0e
Segmentation fault: 11

@DavidCarlisle gdb session? :)

I don't know: you look at a beamer issue and decide to use etoolbox to fix it, then you find at least two bugs in etoolbox!

:)

@DavidCarlisle On that etoolbox business, I notice that \ifdefltxprotect is a bit complex to explain: it only comes out TRUE for things set up with \DeclareRobustCommand not those using \newcommand. (You can imagine how I'd like to fix that!)

@JosephWright I don't think we ever overly documented that newcommand makes things latex-robust just if it is used to declare a command with an optional argument:-)

@DavidCarlisle Possibly not

@DavidCarlisle I guess \ifdefltxprotect is only actually used by etoolbox anyway so not a big issue (the code paths are correct)

@egreg: I tried to buy Pringles yesterday and the market had some sort of generic one, it was Mister Potato. :)

@PauloCereda Mister Potato?!

@JosephWright oooooh

@JosephWright that's a big potato!

@PauloCereda Yup

"He's a super potato"

Where's the fox bloke? Freddy, I guess...

@PauloCereda Baa says Suzi Sheep!

@JosephWright at least she's not a clever clogs. :)

@PauloCereda Indeed

@DavidCarlisle: advanced pig-related discussion. ^^

I asked this question a month or so ago. It didn't receive enough attention to get any comments or answers. I put a bounty on it, but alas, it still didn't recieve much attention. I am aware that there are properly very few people in the world who know the answer. Do you have any idea where to go from here to get an answer? tex.stackexchange.com/questions/334147/…

@neic The Gnuplot input is simply 'dumped' by LaTeX to file: I'm not sure there's any obvious way to have access other than reading back

@JosephWright I guess that PL wanted to protect against code that uses the \foo<space> trick (there are several instances around). From my point of view, the “outer” macro should be made \protected, with replacement text from the inner one. Or, possibly, that path was the easier one.

@JosephWright I'm not sure I understand you meaning. I'm looking for gnuplot output, not input.

@JosephWright By the way, the kernel \MakeRobust macro can fail: if a package defines \cite as a normal macro, \MakeRobust\cite will not work, because a definition for \cite<space> exists.

@neic not sure how else other than a file that could work, all interaction between tex and gnuplot is via files isn't it? (I'd have the \newread in the preamble not in the body of the document but other than that your code looks about right to me)

@egreg you could try our new latexbug.sty package

@DavidCarlisle The package in question was the bug. ;-)

@egreg even better (although I meant re makerobust)

@neic OK, but as @DavidCarlisle says the only path for data is using a file

@egreg Yes, but they only get picked up at all for that part of the code if the outer definitions matches the \protect \foo[space] model used by \DeclareRobustCommand

@egreg I have a simple solution to that ;)

@egreg I guess in seriousness we could fix that (though simply using \protected is pretty tempting to me)

@JosephWright not sure you'll persuade @egreg to use context

@egreg What happens is exactly what you say, it's just that the clean-up is not I think quite right. Try for example

\RequirePackage{etoolbox}
\DeclareRobustCommand\foo[2][a]{}
\show\foo
\expandafter\show\csname foo \endcsname
\robustify\foo
\show\foo
\expandafter\show\csname foo \endcsname

@DavidCarlisle :)

@DavidCarlisle, @egreg Full command is

\newrobustcmd*{\robustify}[1]{%
  \ifundef{#1}
    {\etb@error{\string#1 undefined}\@eha}
    {\ifdefmacro{#1}
       {\ifdefltxprotect{#1}
          {\letcs\etb@resrvda{\expandafter\@gobble\string#1 }%
           \@tempswatrue}
          {\let\etb@resrvda#1%
           \@tempswafalse}%
        \ifdefparam\etb@resrvda
          {\etb@ifscanable\etb@resrvda
             {\etb@robustify\etb@resrvda
              \let#1\etb@resrvda}
             {\etb@error{Failed to robustify \string#1}
                {The command is special and cannot be

which as I read it means that if \foo[space] exists as part of a robust macro it is always used as the new 'main' macro

@DavidCarlisle @JosephWright. Ok. Thank you. I'll try to code a solution where the filename and reader name are generated automatically so the user (not me) don't have to keep track of all the reader and filenames.

@yo' -- several (looong!) file drawers, i think. but probably better with spooled tape, maybe 10 or so 20cm spools.

@barbarabeeton Or in @PauloCereda's case: half an inch

@JosephWright Which doesn't really solve the problem I showed before

@DavidCarlisle oy

@barbarabeeton @Paulo ya bed'r numberz teh tapez or your screewed

@yo' :)

@yo' -- no kidding! and don't drop them either. rewinding is a miserable job; better have patch tape available too. (been there, done that. even got a calibrated hole-driller as a reward for my efforts.)

@barbarabeeton I quite wasn't kidding, I'm sure that catalogization and proper storage of the tapes was a nightmare!

@yo' -- not easy, for sure, but better than dropping a tray of punched cards that didn't have sequence numbers in the last 8 columns.

@barbarabeeton "jigsaw" puzzle level 1000? probably almost like DNA sequencing!

@yo' -- yeah. and if the card deck wasn't printed (never seen a duplicator that had a print head), you can up that puzzle level by a couple orders of magnitude. at one time, i could read by what holes were punched, but never fast, and if the code was binary, forget it. (at least, at the time, recycled cards had a very generous recycle value.)

"A Dragon is a highly unusual biologic version of a lighter."

@egreg huh? If you mean a non-robust \cite, it wouldn't have the form \protect\cite[space] at the top level so any 'stray' \cite[space] won't affect things

@JosephWright The problem was an old package that does \def\cite{<whatever>}. Then, if \MakeRobust takes the current definition of \cite<space>, the command \cite will do quite a different thing from the expected.

@egreg I'm not thinking about \MakeRobust, I'm thinking about \robustify - the former is a different problem

@egreg isn't this test supposed to make sure it detects that \@ifundefined{\expandafter\@gobble\string#1\space}% but maybe I don't understand the issue (and didn't try a test:-)

@DavidCarlisle Silently fails to do anything

\show\cite
\def\cite#1{Hello}
\MakeRobust\cite
\show\cite

Contrast

\RequirePackage{etoolbox}
\show\cite
\def\cite#1{Hello}
\robustify\cite
\show\cite

not that silent just log only LaTeX Info: The control sequence `\cite' is already robust on input line 3. yes it doesn't cover the case that someone has hidden a previously definition.

@JosephWright I suppose it could check that it has the form \protect\foo as well...

s/previously definition/previously robust definition/

@DavidCarlisle Yes, that's what etoolbox does

@DavidCarlisle Indeed, though I think 'just use etoolbox and make engine robust' might today be better advice

@yo' @PauloCereda sklepde.plakatoteka.pl/img.php?D=1&P=17194274

@UlrikeFischer ooh!

@UlrikeFischer: I'd happily make a tee of this image, but it's quite dangerous to wear a red tee around here. :)

@PauloCereda It will look in green good too imho.

@UlrikeFischer ooh

@UlrikeFischer ^^

@PauloCereda You could fit it to the colors of Palmeiras. Or Borussia Mönchengladbach ;-)

@UlrikeFischer ooooh I like those ideas. :)

pdfquestion: Anybody knows how to flatten a pdf-form completely? I've tried with Preview and PdfExpert, PDFwriter, and managed to flatten all textfields, but the checkboxes still remain clickable.

no worries, got it fixed through LaTeX-magic instead. no need to flatten it.

Is there some way to convince tikz nodes in beamer slides to be invisible instead of missing when covered? (i.e. take up space and be available for reference when using relative positioning) I can of course wrap them in \visible<>{} but as they are already overlay aware, I imagine there might be a way to change the default behavior.

From the tikz and beamer documentations I have not even been able yet to figure out who defines the overlay aware versions of \node etc anyway...

@UlrikeFischer These colors?

@egreg Entebierhaus

@barbarabeeton odd it wasn't mentioned to the stix beta testers:-)

@egreg The current beer in the stadion are bitburger and bolten.

@UlrikeFischer Hannen was the first Altbier I drank; our German friends from Krefeld brought it.

@UlrikeFischer Bitte ein Bit. :-)

@barbarabeeton Nice! If I want to have the slashed 0 for the empty set, all my summation signs will be small also in display style. STIX Two Math, ss07.

@barbarabeeton Sorry, but it's hilarious.

@DavidCarlisle -- that's a very good question! will investigate. (i do think there was a mention sometime back there that release would be scheduled for december 1 or so. but i hadn't heard confirmation.)

@barbarabeeton you can't seriously expect that anyone took any notice of forward predictions of stix release dates:-)

@egreg -- what!!??!!?? i'll investigate.

@DavidCarlisle -- touch`e.

Ok, finally made TeXStudio more readable by turning latexSyntaxMistake highlighting off, as apparently every custom command or one I use by a package it doesn't know is a mistake.

@barbarabeeton Minimal example

\documentclass{article}
\usepackage{unicode-math}

\setmathfont{STIX Two Math}
\setmathfont{STIX Two Math}[version=test,StylisticSet=07]

\begin{document}

$\sum\displaystyle\sum\emptyset$

\mathversion{test}

$\sum\displaystyle\sum\emptyset$

\end{document}

@barbarabeeton I can understand a stylistic set with small symbols for summation and product. But what's the relation with the empty set?

@PauloCereda The investigation reveals some very sad facts, this is so bad! And exactly the type of things that shall never ever happen, not in the so much controlled and secured industry! :-(

@PauloCereda try out the newest version of tikzpeople:

\documentclass{report}

\usepackage{tikzpeople}
\usetikzlibrary{shapes.callouts} %
\begin{document}
\begin{tikzpicture}

\node[name=duck,duck,evil,minimum size=5cm]{};
\node[ellipse callout, draw, xshift=3cm,yshift=2cm,
callout absolute pointer={([xshift=0.5cm]duck.mouth)},
] {Quack};
\end{tikzpicture}
\end{document}