@PhelypeOleinik Phelype, vá à pharmácia. :)

@PauloCereda Very phunny ¬_¬

@PauloCereda :-)

I am half Italian, so people can make fun too.

@mickep I thought maybe MP has cleaner abstraction capabilities. And it seems it might have proper functions. And possibly variables?

@PauloCereda Seu nome é muito Massa :D

@FaheemMitha Well, I tried tikz, and there were so many options and things to tweak, so it just took me too much time (reading the wonderful, but long manual) to get going. With metapost I more or less have my points, my paths and some equation solver, some path cutting. But I probably prefer it since I'm more used to it.

@PhelypeOleinik /ba dum tss LOL

@mickep PGF/TikZ certainly has a lot of stuff.

@FaheemMitha On the other hand MetaPost does not support 3D (as easily) as tikz/asymptote/pstricks does. Well, I think one should test the options and go with the gut feeling.

@FaheemMitha metapost is fundamentally metafont, extended to produce arbitrary postscript rather than just font glyphs.

@DavidCarlisle ooh metametafun

@PauloCereda \def\meta{\meta}\meta fun

@DavidCarlisle So I gathered. But I don't know MF either.

@DavidCarlisle But you stick to picture mode, don't you?

@PauloCereda I'm more meta than you.

@DavidCarlisle by a large margin, yes. :)

@mickep as do all people of taste and sophistication.

Does MF have real variables? And functions?

@DavidCarlisle 𝔰𝔲𝔯𝔢𝔩𝔶, 𝔪'𝔩𝔬𝔯𝔡

@FaheemMitha MetaFun has (as you might know) the very nice manual with lots of examples to look at.

@FaheemMitha define real. I haven't used it much this century so don't ask me, but if you are looking to avoid Tex's eccentric language idioms choosing a language by the same author produced at the same time, might not be the best choice.

@DavidCarlisle GOD

/Fortran humour

@DavidCarlisle Sadly, I don't have a CS degree, so I can't define "real variables". As already mentioned, I don't have much of an education. I blame my parents.

@FaheemMitha I meant define "real" you asked the question so you must have some definition of variable in mind and some system that you don't consider to have real variables/

@DavidCarlisle ooh bindings

Also eccentricity is relative. And I don't mind TeX's weirdness most of the time. But it doesn't hurt to look at alternatives.

λx.x

@PauloCereda Ohh, lambda!

@mickep ooh <3

@PauloCereda I suppose it gives you an identity:-)

@DavidCarlisle Real variables as in a language like Python. You can assign to them and they have storage allocated to them, and have scope etc.

OTOH, TeX doesn't have variables in the usual sense. Nor does it have functions.

@FaheemMitha some people would moan they're not typed but Ok but so be it

@FaheemMitha I am not sure that I agree with that. Define "usual sense" and define "function"

@DavidCarlisle Python has typing, doesn't it?

@DavidCarlisle ooh a math joke :)

@DavidCarlisle I thought you might not.

@PauloCereda He sounds very much like a mathematician right now, pushing for definitions.

See my comments above, regarding assignment, storage, scope. And again, remember my lack of a CS education.

@FaheemMitha I am sure @PauloCereda will be pleased to tell you that typing in python is what is commonly known as duck typing

@mickep he is definitely plotting something!

/ba dum tss

@mickep That's what it says on my passport

@PauloCereda ohhh ;)

@DavidCarlisle if it walks like a duck and it quacks like a duck, then it must be a duck

And here's Wikipedia's page on variables - en.wikipedia.org/wiki/Variable_(computer_science)

@DavidCarlisle Your passport says you are a mathematician??

@mickep yes

@DavidCarlisle Cool! I wish we had such passports!

@mickep Uk passports have always had an "Occupation" field

I think I read in Mark Kac's autobiography that a colleague, pressed to ask for his profession, responded that he was a topologist.

@DavidCarlisle Really?!

@mickep I wanna be English

And I want duck to be written on my passport

@DavidCarlisle oh nein

I wanna be Swede

@PauloCereda With baked beans for breakfast? :S

@PauloCereda Welcome!

@mickep then fish and chips for lunch. :)

@mickep ooh <3

@PauloCereda and roast duck for dinner

@DavidCarlisle oh no

@DavidCarlisle The error comes from \includegraphics of course: github.com/pgf-tikz/pgf/blob/…

@DavidCarlisle Here is the failing build: travis-ci.com/hmenke/pgf/jobs/262921782#L2824

@FaheemMitha I would much rather prefer a pull request for typos. This is of course an easy issue to open but time-consuming and annoying for me. You don't even have to clone the repo for that. Just go on GitHub, find the file in doc/generic/pgf/text-en/ and click the pencil in the upper right hand corner. When you are done editing, write a commit message and click submit. This will automatically fork the repo. create a new branch, commit your change to the new branch and file a PR.

@HenriMenke or more exactly the error comes from that file doing mask=\pgf@imagemask using an undocumented mask key that was (@JosephWright reminded me yesterday) apparently in some version of dvipdfmx.def before the def files were centrally maintained, but has never been part of graphics

@DavidCarlisle But why did it start failing yesterday after it worked for years?

@HenriMenke if it's doing something useful we could make an extension request to the graphics package to add it (driver specific back end files really shouldn't be adding user level options)

@HenriMenke it's not been defined in the texlive dvipdfmx.def since about 2015 as far as I can see

@HenriMenke I really can not guess, a change elsewhere caused it to go down some previously unused code path?

@HenriMenke since it's running with pdftex how come it loaded graphics-def/dvipdfmx.def at all?

@HenriMenke or earlier why (../pgfmanual-dvipdfmx.cfg

@HenriMenke oh sorry you are in dvi mode, Ok

How does:

> this documentation comes with a number of slowly-paced tutorials

read?

@DavidCarlisle I found out why it didn't fail for years.

Also isn't "full-blown" more standard than "full-flung"? The latter doesn't get any hits on Google.

@HenriMenke Done. There are a few other suggested changes, all very minor.

@DavidCarlisle I just tried rolling out a new build system and I do TEXINPUTS=../../text-en:../../images: for all engines, but in the old build system for dvipdfmx and dvipdfm it only does TEXINPUTS=../../text-en: which means that the images are never found and therefore not included, hence the mask can't fail.

Please feel free to accept or discard as you feel appropriate.

I assume that that section was mostly written by Tantau.

@FaheemMitha Merged. Thank you.

@DavidCarlisle How can I apply the mask myself, now that this interface is gone?

@HenriMenke no idea, I'd have to look what the old copy was doing. But it would be good to know what changed, it's a bit worrying if an \ifsomething is taking a different course for an unknown reason.

@DavidCarlisle The reason is that the included images were not found previously in which case \pgfimage just emits a warning and does nothing. Now that I fixed TEXINPUTS the images are suddenly being found and that is when mask blows up.

Is it happening only to me? ^^^

@HenriMenke ah... \define@key{Gin}{mask}{\def\Gin@mask{#1}} will define the key

@DavidCarlisle yes, but probably it will do nothing sensible.

@CarLaTeX It's just you.

It happens on Edge, Chrome and Firefox

@UlrikeFischer it used to do:

\def\Ginclude@pdf#1{%
 \message{<#1>}%
  \bgroup
  \def\@tempa{!}%
    \special{pdf:epdf\space
      \ifx\Gin@page\@tempa\else page\space\Gin@page\space\fi
      \ifx\GPT@pagebox\ltx@empty\else pagebox\space\GPT@pagebox\space\fi
      bbox\space\Gin@llx\space\Gin@lly\space\Gin@urx\space\Gin@ury\space
      clip\space\ifGin@clip 1\else 0\fi\space
      \ifx\Gin@scalex\@tempa\else width\space\the\Gin@req@width\space\fi
      \ifx\Gin@scaley\@tempa\else height\space\the\Gin@req@height\space\fi
      (#1)\space

@HenriMenke It worked perfectly this morning

@CarLaTeX I don't have the left sidebar anyway:

@CarLaTeX Firefox 70.0.1 but I also don't have the left sidebar.

Hey, is \write18 able to write content into a txt-file, similar to filecontents?

(I need something to use in a loop.)

@cis well not 18 but yes that's exactly what it does filecontents is a wrapper around that.

@DavidCarlisle @HenriMenke I'll try to turn on and off again...

@cis classic tex has 16 write channels 0-15 so \write18 was just a syntactic trick to use the syntax of writing to a file but give it a special meaning of executing stuff.

@DavidCarlisle I tried something like that (without success) :

\documentclass[a4paper]{article}
\begin{document}
\def\aaa{qwertzu}

\foreach \k in {1}{
\begin{filecontents*}[overwrite]{data1.txt}
Line  , 5  \aaa
Line , 6
\end{filecontents*}
}
\end{document}

@cis well that writes everything to the same place so you only get the last output

@cis well \foreach is not defined ...

@cis well it would if foreach was defined

@DavidCarlisle Ouh, yes once more:

\documentclass[a4paper, landscape=false]{article}
\usepackage{ tikz}

\begin{document}
\def\aaa{qwertzu}

\foreach \k in {1,...,9}{
\begin{filecontents*}[overwrite]{data\k.txt}
Line  , 5  \aaa
Line , 6
\end{filecontents*}
}
\end{document}

@DavidCarlisle well no, the content is empty

@cis it is so scary seeing someone load half a million lines of tikz code to define a loop that would tale a couple of lines:-)

@DavidCarlisle BTW: The content \aaa should depend from foreach too...

@DavidCarlisle much better to load lots of expl3 code ;-)

@DavidCarlisle Yes, pgffor.sty

@cis er I'd use \loop myself, which is defined already:-)

@UlrikeFischer Don't have to load that, it's just there ;)

@DavidCarlisle @HenriMenke It works only if I zoom out at 75% :(

@DavidCarlisle Aha, I did not know the command \loop

@HenriMenke Are you the only PGF/TikZ maintainer?

@cis \int_step_inline:nn ;)

I see two other people listed. One is David. The other is Phelype.

@JosephWright Ohhh. I think I have to open a question tomorrow. ;)

@cis er I get the same... @JosephWright @UlrikeFischer can you try the above (I get 9 empty files)

I deleted the branch that was created. Is there any point in keeping the fork around?

@FaheemMitha nobody trusts me to actually do anything to the tikz code though:-)

@cis \foreach can't work, the content is an argument and can't change catcodes. Try \foreach \k in {1}{\verb+$+}.

@DavidCarlisle Yup

@DavidCarlisle @HenriMenke without the left bar it works

@JosephWright blame Frank?

@DavidCarlisle Sounds like a plan: he's not here

@DavidCarlisle where, for the foreach? As I wrote: if doesn't work for verbatim.

@UlrikeFischer ah yes, but odd that you get no warning and it just makes empty files.

@DavidCarlisle Can one test if one is in an argument where catcode changes don't work?

@UlrikeFischer OK.

@UlrikeFischer not really but normally you'd expect that the macros fail and give at least some spurious fragile command in moving argument kind of error, not just silently write empty files

@FaheemMitha The current maintainers are Till Tantau, Christian Feuersänger, David Carlisle, Joseph Wright, Stefan Pinnow, and me. But in practice, I'm the only one actively working on it.

@HenriMenke I'm working on the expl3 re-implementation ;)

@HenriMenke Till's already managed to palm beamer off on me ...

@HenriMenke BTW, that 'expl3' comment's not entirely facetious: if/when I get the backend code fully-working, I might have some suggestions for pgf

@JosephWright You have push access to the pgf-tikz/pgf repo and you are a member of the “maintainers” team.

@cis

\documentclass[a4paper, landscape=false]{article}
\newcount\filecount
\newwrite\cisout
\begin{document}
\def\aaa{qwertzu}

{
\endlinechar=\newlinechar%
\filecount=1 %
\def\aaa{file number \the\filecount}%
\loop%
\immediate\openout\cisout=data\the\filecount.txt %
\immediate\write\cisout{%
Line  , 5  \aaa
Line , 6
}%
\immediate\closeout\cisout %
\advance\filecount by 1 %
\ifnum\filecount<10 %
\repeat %
}%


\end{document}

@HenriMenke :)

@JosephWright If you could repair the xetex driver, that would be great.

@HenriMenke @DavidCarlisle and I (on behalf of the team ...) seem to have picked up similar access to almost all of CTAN ...

@HenriMenke Hmm. That's not a whole lot of manpower, then. Just like all the other free software projects out there.

@JosephWright oberdiek's keeping me busy again this evening....

@HenriMenke I have wondered if the team should ask CTAN to promote a formal '+ the team as fallback' approach to access/licensing, as single-person maintainers have a habit of vanishing

@DavidCarlisle I'd seen

@DavidCarlisle I'm on l3text and the business case ...

@DavidCarlisle Although day today was mainly 'do a viva voce'

@JosephWright I'm trying to formulate a plan to disentangle hyperref from oberdiek (as a first step to sharing back end code with expl3)

@JosephWright @UlrikeFischer crazy that an empty document loading hyperref does

$ pdflatex hy1 | grep oberdiek
(/usr/local/texlive/2019/texmf-dist/tex/generic/oberdiek/ltxcmds.sty)
(/usr/local/texlive/2019/texmf-dist/tex/generic/oberdiek/infwarerr.sty))
(/usr/local/texlive/2019/texmf-dist/tex/generic/oberdiek/kvsetkeys.sty
(/usr/local/texlive/2019/texmf-dist/tex/generic/oberdiek/etexcmds.sty
(/usr/local/texlive/2019/texmf-dist/tex/generic/oberdiek/kvdefinekeys.sty)
(/usr/local/texlive/2019/texmf-dist/tex/generic/oberdiek/pdfescape.sty)
(/usr/local/texlive/2019/texmf-dist/tex/latex/oberdiek/hycolor.sty

@DavidCarlisle Yeah: it's a shame that Heiko didn't get involved in expl3 when e was writing all that generic stuff ...

@PauloCereda Oh, my...

@DavidCarlisle UiUi, that works. Thx. But not a very easy code. ;)

@PhelypeOleinik terrible, isn't it? :)

GOOD NEWS, EVERYONE

@PauloCereda brexit is cancelled?

@DavidCarlisle oh wait

@PauloCereda ALL YOUR BASE ARE BELONG TO US?

@JosephWright oh no

@DavidCarlisle and almost everything looks like things that should be in the kernel/expl3

@UlrikeFischer Yup: all coming

@UlrikeFischer Need to get Bruno to do big integer division ... or work it out myself

@UlrikeFischer I'm working on the text-to-string business ;)

@DavidCarlisle I tried to figure out what this xcolor-patch.sty is about but couldn't find any reference which problem is it meant to resolve.

@UlrikeFischer yes, but I was hoping to go in small steps, separate first and switch back end later, but I suspect that doesn't make sense and switching to expl3 at same time is going to be more manageable

@JosephWright that will be very useful. We need a generic approach to this "convert user input into ...".

@UlrikeFischer Yes, I know: Chris finally seems to be happy with me calling it text ..

@UlrikeFischer We have Bruno's string stuff, so UTF16/escape is OK

@UlrikeFischer The format-stripping part is easy enough, it's the math mode ...

@DavidCarlisle I'm thinking of it for csquotes as I mentioned earlier ...

@DavidCarlisle Once 2020-02-02 is out?

@UlrikeFischer @JosephWright did you see the expl3 math actualtxt question today?

@DavidCarlisle Er, yes, but that was ... odd

@DavidCarlisle yes, but I don't really know how \collect@body works and had no time to figure out if he is using it wrong or not.

@JosephWright yes I was hoping to separate things now to make teh scope clearer and then switch in expl3 in 2020 but I'm coming round to the idea of leaving it more or less as it is now until 2020 as isolating all those packages just seems work for little gain

@CarLaTeX Looks like the site's bugged for me too (but only some pages): i.sstatic.net/ryIzg.png

@PhelypeOleinik same here for me.

@Skillmon @CarLaTeX I just found this: meta.stackexchange.com/q/339651/383578

It's a sign of the end times. Hugh Grant just made a political ad.

@JosephWright That is why you and David have push access. So that in case I have to give up, you can at least merge PRs and triage issues.

@HenriMenke Sure

@PhelypeOleinik Thank you! Indeed, this morning worked perfectly.

@CarLaTeX You're welcome :)

@CarLaTeX Seems to be fixed...

@PauloCereda Shouldn't the opposite of “ice” be “water” rather than “fire”?