quack

@PauloCereda oh I had roast duck last night, and you're still here?

@DavidCarlisle oh no

@UlrikeFischer @JosephWright if we were looking for a default setup I suppose it ought to at least in principle work from miktex so it might yet be better to extract the version string and check it's at least 1.12 as a number. There doesn't appear to be any way to get the equivalent of tex.luatexversion without splitting up the banner string from texlua unfortunately

@DavidCarlisle the version number doesn't work: in miktex I have luahbtex 1.12 (used currently by lualatex) and luatex 1.12 (used by lualatex-dev). In luaotfload I have a test string.match (luatexstatus.banner,"MiKTeX") to setup things for miktex.

@UlrikeFischer blurg I suppose you are right, in that case the test above is probably OK as in '2019' == texlive 2019 and anything else including miktex assume a new setup with lualatex using luahbtex...

@DavidCarlisle ooh blurg

@DavidCarlisle Imho for now the main point is to get over the freeze period which means travis that is still using texlive 2019. miktex tests can wait.

@UlrikeFischer sure . Although since travis setup installs the whole thing anyway it could use the pretest as easily as tl2019

@UlrikeFischer it's good to have documentation so you can say it's not your fault tex.stackexchange.com/questions/531851/…

@DavidCarlisle ooh

@DavidCarlisle I was wondering about this. But wouldn't it put too much strain on the pretest mirrors?

@DavidCarlisle does arara work now?

Assuming that one point in its timeline it has worked before...

@PauloCereda it works as well as arara 4, I never got beyond arara --version with either:-)

@PauloCereda arara --version works for me now.

@DavidCarlisle I can relate to that. :) I actually use arara way less than a typical arara user. :)

@UlrikeFischer yay!

@UlrikeFischer travis would cache the install so in most cases it would only be a tlmgr update

@PauloCereda GNU make ftw!

@Skillmonlikestopanswers.xyz bah :)

@PauloCereda usually my bigger projects have many dependencies, every plot is done in an individual file, I create tables using Python, etc. Good luck building that without make.

@Skillmonlikestopanswers.xyz well... :)

% arara: make: { targets: [ foo, bar ] }

@PauloCereda what's the advantage of calling arara to call make?

@Skillmonlikestopanswers.xyz Because REASONS

<3

@PauloCereda maybe I could take a look into this for the future. Is arara capable of creating a LaTeX format with a built-in rule?

@Skillmonlikestopanswers.xyz I've no idea. :) How is a LaTeX format created?

$(FMT): $(PREAM) $(FILE)
	$(FOT) -ini -jobname="$(NAME)" "&$(TEX)" mylatexformat.ltx """$(FILE)"""

@PauloCereda ^^^ that's my make target.

@Skillmonlikestopanswers.xyz seems doable. :)

@AlanMunn twitter.com/jpdowling/status/1235629424634195973

@PauloCereda classical misuse of the author field, people should use AND.

@Skillmonlikestopanswers.xyz Yes. :)

Anyone know which font to use with beamer for xelatex/lualatex+Russian with T2A fontenc (I'm assuming the latter is not relevant)

@daleif As soon as you said Russian, I was out.

@daleif apart from the standard "don't use legacy font encodings with xetex" answer?

@daleif There are a few options here: tex.stackexchange.com/questions/816/cyrillic-in-latex

@PhelypeOleinik @DavidCarlisle even this small example does not work with xelatex or lualatex

I only get ! . in the output

Presumably because of these

Missing character: There is no Р (U+0420) in font [lmsans10-regular]:+tlig;!

@daleif sure latin modern doesn't have cyrillic (there is a clue in the name:-) but you can use fontspec to pick up any other font, most system text fonts will have cyrillic (the link @PhelypeOleinik just gave has an answer suggesting cmu or linux libertine)

I usually can never set those to work on my linux box. Preaumablt fc-cache or something is not set up properly.

@daleif I Like Fira GO:

\documentclass{beamer}
\usepackage[russian]{babel}
\usepackage{fontspec}
\setsansfont{Fira GO}
% \setsansfont{Noto Sans}
\begin{document}
\begin{frame}
ПРИВЕТ! Этот текст написан шрифтом без засечек.
\end{frame}
\end{document}

@PhelypeOleinik none of those work on my system, they cannot find the font.

Correction lualatex could find Noto Sans

@daleif You have to download it from here: github.com/bBoxType/FiraGO. I put the .otf files under ~/.local/share/fonts and Lua/XeLaTeX can pick them

But @DavidCarlisle prefers Arial ;-)

@PhelypeOleinik it's just a good choice if you are not sure of the script as it has quite large coverage these days:-)

Hi, can you please try checking whether this gives an error?

\listfiles \batchmode

\documentclass{article}

\usepackage{svg}
\usepackage{markdown}

\begin{document}

Hello, World!

\end{document}

@DavidCarlisle It's interesting that people put development effort into things like Arial

To me it gives:

\pgfutil@addpdfresource@extgs ...urce@extgs@plain
                                                  {#1}
l.8 \begin{document}

@yo' Same here

(and it shows this string in the PDF itsef: TRP1¡¡/ca 1/CA 1¿¿)

@PhelypeOleinik thanks. Can you post the listfiles please?

 *File List*
 article.cls    2019/12/20 v1.4l Standard LaTeX document class
  size10.clo    2019/12/20 v1.4l Standard LaTeX file (size option)
     svg.sty    2020/01/13 v2.02e (include SVG pictures)
   iftex.sty    2020/03/06 v1.0d TeX engine tests
 scrbase.sty    2020/01/24 v3.29 KOMA-Script package (KOMA-Script-independent b
asics and keyval usage)
  keyval.sty    2014/10/28 v1.15 key=value parser (DPC)
scrlfile.sty    2020/01/24 v3.29 KOMA-Script package (loading files)
pdftexcmds.sty    2019/11/24 v0.31 Utility functions of pdfTeX for LuaTeX (HO)

@yo' There you go ^^^

@yo' oh did he finally provide a mwe?

thanks a lot!

@DavidCarlisle in a sense, yes.

@yo' I know this, let me just find the issue.

@PhelypeOleinik large corporations based in Seattle, not people

@UlrikeFischer ah ok. (Btw it seems to boil down to transparent.sty)

Ein Bier, bitte

@yo' only in part. The problem is csvsimple here loading only a subset of pgf and this confuses transparent. See github.com/ho-tex/transparent/issues/3

Pivo prosím

@UlrikeFischer I thought I'd seen that somewhere:)

@PauloCereda Enten-Ananas-Pizza bitte.

@DavidCarlisle oh nein

/quacks in despair in German

@UlrikeFischer Oh thanks a lot. So \usepackage{pgfcore} is the cure :-)

@yo' or \usepackage{pgfsys}.

@UlrikeFischer Yeah, doesn't make much difference. (Now just someone to blame is needed. Any idea, @DavidCarlisle ?)

Anyway, thanks for the help!

@yo' PDF Resource handling, no question: Blame @UlrikeFischer

@DavidCarlisle ooh

@JosephWright I have a curious failure in xetex tests. With tl19 I get this trailer:

<</Root 1 0 R/Info 2 0 R/ID[<4af12ce19144e087d0d2f1e8a81b959a><4af12ce19144e087d0d2f1e8a81b959a>]/Size

and with tl20 this one

<</Root 1 0 R/Info 2 0 R/ID[(h,t\304a\363\373\034\361:8eTxN')(h,t\304a\363\373\034\361:8eTxN')]/Size

Somehow the epoch setting seems to go wrong.

@UlrikeFischer is that a valid string in a different encoding or bad pdf?

@DavidCarlisle somewhere in the multiverse, the PDF is valid. :)

@PauloCereda For each PDF file, there exists a planet where it's valid by the local PDF specification.

@DavidCarlisle As string it is imho valid. I'm not quite sure if the content of the /ID has to have a specific format, but it wouldn't really matter in the test suite. The main problem is that is the tests fails.

@yo' ooh

@PauloCereda #8405

@yo' ooh COMBINING CLOCKWISE ARROW ABOVE

:)

:)

gotta go now, it's very late lunch time

@UlrikeFischer yes but the promise of the epoch setting is only that if you run it again you get the same data, not that the data in one implementation is encoded in the same format as the previous release used

@UlrikeFischer /Skype poke :)

quack

@CarLaTeX no smartworking for me, at least for now

@DavidCarlisle well the curious thing is that it doesn't happen all the time. If I rename the file I get the "normal" output. But I will take it for now and hope it doesn't indicate a problem.

@UlrikeFischer Er, that's odd

@JosephWright yes, but I don't think that it is a l3build problem. I can reproduce it too if I set SOURCE_DATE_EPOCH etc on the command line and compile then. But only for one file currently.

@PauloCereda "Never trust bibtex"? More like never trust humans.

@AlanMunn Also, BibTex :)

@PauloCereda bibtex is the executable... :)

@AlanMunn I meant the literal capitalization the bloke used. :)

@PauloCereda Oh I thought you were complaining about me.

@AlanMunn never. <3

How quickly they forget.

@Moriambar Since today smart-working mandatory until further notice

@AlanMunn oh

@PauloCereda Fooled by someone from Novo Hamburgo. tex.stackexchange.com/q/531897/2693 (see comments).

@AlanMunn ooh fake Deutsch :)

@PauloCereda Indeed. :)

@JosephWright (and the other moderators) -- apologies for flagging a question that was in the correct queue; I thought I was in meta when I acted on it too quickly.

@DavidCarlisle ooh EAS-4-TbSync

@PauloCereda :)

@JosephWright we ducks are good at finding stuff. :)

@PauloCereda doesn't really work though:-)

@DavidCarlisle oopsie

@DavidCarlisle No?

@JosephWright main thing I was missing with the extension I had before was accepting meeting invites and that's explicitly listed as not yet implemented with this tbsync based one. So it works for what it claims to do but....

@JosephWright looking in TB's addon list it says:

@DavidCarlisle Saw it

@JosephWright looks like it's in the pipeline, so hopefully it'll come. Otehrwise have to open up the office 365 web interface to accept invites (or more often don't bother and just turn up without having accepted:-)

@DavidCarlisle That's me :)

ooh chemmacros updated

@PauloCereda who needs chemistry?

@DavidCarlisle Englishducks. :)

I had to kill a LuaTeX process and the shell said:

> fish: “luatex test.tex” terminated by signal SIGTERM (Polite quit request)

So polite :-)

@PhelypeOleinik so not kill -9 # die now and good riddance then?

@PhelypeOleinik oh no

@PhelypeOleinik bah who uses fish shell :)

@DavidCarlisle :-)

@PauloCereda Fishes?

@PauloCereda shellfish

@PhelypeOleinik :)

@DavidCarlisle ooh ligatures

@PauloCereda selfish ducks keep shellfish shells in shelf-ish places.

@DavidCarlisle oh :(

I want a semicolon in math - what's the escape or symbol?

(how are you all doing, by the way?)

:)

google is failing me...

@AaronHall ;

does mathjax not handle it properly?

I want to do:

$$2 = \sqrt{2F(0.90;2,8)}$$

ok, added spaces around it, it now works...

@AaronHall spaces around the ; ? wouldn't be needed in tex, your code produces this in plain tex:

ah - I'm working from my notes in a Jupyter Lab notebook...

looks like I found an idiosyncrasy...

@DavidCarlisle thanks!!!!

@UlrikeFischer might try to merge in ctan.org/pkg/ltabptch although clashes a bit with some of the changes already added

@DavidCarlisle Can I ask you for a comma placement judgement?

@AlanMunn you can ask, but I may not answer:-)

Suppose the sentence is: One should not assume that just because I asked, you should answer. would you put a comma after 'that' ?

@DavidCarlisle ctan is full of surprises. Where are the clashes?

@UlrikeFischer oh he's re-arranging LT@start.....

@AlanMunn no

@AlanMunn should I assume that because you asked, I should have added a comma?

@DavidCarlisle No, my Portuguese speaking co-author is comma-happy. :)

@DavidCarlisle I just wanted confirmation of my intuition from another English speaker.

@AlanMunn we linguists know about commas

@AlanMunn real reason I said no was that it looked wrong, but trying to rationalise it I think perhaps that with two commas it looks like One should not assume that you should answer with just because I asked being in apposition or whatever it's called, which isn't the right sense here.

@DavidCarlisle Well their intuition is that the "just..." phrase is a modifier and should be set off on both sides by commas.

@AlanMunn don't trust people who put random dots and twiddles over letters.

@DavidCarlisle îñdéèd.

@DavidCarlisle BTW, correcting page proofs by using PDF comments (which is what I'm doing) is a nightmare, especially when you're trying to correct typesetting errors.

@AlanMunn we used to do that a lot at work (using pdfxchange) but use an html annotation system these days

Just one example: we use $^\cup$ and $^\cap$ as functions, but they sometimes forget the ^, which gets the meaning completely wrong.

Last find in a package:

  \def\pb@xargi\relax
  \def\pb@xargii\relax
  \def\pb@xargiii\relax
  \makebox[0pt]{}

@DavidCarlisle in the sty he is mentioning the double \global. He seems to have told you about it in 2004 ;-).

@UlrikeFischer la la la I can't hear you

@egreg presumably the author needed an intricate definition of \pb@xargi

@PhelypeOleinik you do know you could use one line of sed before running tex:-)

@DavidCarlisle s/line2/where's the fun in that?/g ;-)

@egreg Ouch!