@JosephWright Quack. :)

RIP Gene Wilder. :(

I'm gonna leave it here so it's easier to find this time

\definecolor{cffcc00}{RGB}{255,204,0}\definecolor{c008000}{RGB}{0,128,0}
\definecolor{caa8800}{RGB}{170,136,0}\definecolor{cd4aa00}{RGB}{212,170,0}
\definecolor{ce6e6e6}{RGB}{230,230,230}
\begin{tikzpicture}
\draw[white,fill=cd4aa00,line width=1pt] (-102:1.3)to[in=115,out=-84]++(-75:0.61) to[out=
25,in=180]++(10:0.42)to[in=35,out =-90]++(-122:0.365)to[in=-65,out=170]++(150:0.33)to[out
=108,in=-65]++(112:0.63)--cycle; \fill[cffcc00](-2,0)arc(180:230:1.5 and 1.8)arc(240:320:
1.32 and 1.9)-- ++(63:1.25)to[out=105,in=5](0,0.7)to[in=110,out=182](-0.32,0) to[out=-70,

keywords : @PauloCereda Duck TikZ

@Johannes_B Hey! Can I bother you for a question?

@Adeetya Sure, why not?

@Johannes_B tex.stackexchange.com/questions/327119/…

I won't create that ugly thing :-)

Can you give me any pointers?

on ho to do it

@Adeetya titletoc or tocloft and a bit of manually adding things.

Okay thanks. I'll look into it

@PauloCereda seems I don't know his movies, although there are many...

@StefanKottwitz Hard spam attack on LC.

@DavidCarlisle @egreg Is there a reason that \@normalcr is not protected? And can you think of some side effect if one robustify it? (see tex.stackexchange.com/questions/327088/…).

@UlrikeFischer Probably historical

@JosephWright My guess too, but I'm still not sure about possible side effects if something suddenly needs an expansion step more.

@UlrikeFischer Oh, everything in TeX is like that :-)

@Johannes_B thanks, deleted

@UlrikeFischer sort of historical, but unlike other things that we just didn't consider making robust at the time, we went to some lengths to make \@normalcr not robust (for 2.09 compatibilty I think) \expandafter\let\expandafter\@normalcr\csname\expandafter\@gobble\string\\ \endcsname (@JosephWright) not sure if that affects a decision now whether to make it robust anyway.

@DavidCarlisle Ah

why isn't backtick working this morning ?

@DavidCarlisle Did never work for me in the chat.

@UlrikeFischer it normally works for me, including around "this morning" when I asked if it didn't work:-)

@DavidCarlisle Fixed

@JosephWright the newline broke it?

@DavidCarlisle Yes

@DavidCarlisle let's try \@normalcr this morning \expandafter

Oh good ...

@JosephWright fitting, given the subject under discussion:-)

@DavidCarlisle :)

@UlrikeFischer @JosephWright we could consider making it etex \protected\def but I don't think \@normalcr could be made robust via \protect\someinternalversion as for example \def\@parboxrestore{\@arrayparboxrestore\let\\\@normalcr} would only let \\ to the \protect wrapper and you wouldn't really be sure which underlying definition it was picking up.

@DavidCarlisle Wait for upcoming team mail

of course as \\ is (now, since % \changes{v1.2a}{1994/11/11}{(DPC) Make robust}) robust anyway parbox restore should have been \def\@parboxrestore{\@arrayparboxrestore\def\\{\protect\@normalcr}}

@DavidCarlisle This solves the problem with stackengine too and looks like a very sensible idea.

@UlrikeFischer but (a) I'm on holiday and not on the internet and (b) I have to wait for @JosephWright's email (but of course I'm on holiday so not reading email:-)

@DavidCarlisle :-)

@DavidCarlisle Mail sent (though of course you won't see it until after your holiday)

@JosephWright yes I saw that had I not been on holiday I would have seen your email referring to this chat.

@DavidCarlisle :-)

@percusse <3

@yo' He was a nice bloke.

@DavidCarlisle Actually it solves only one part of the problem (the error) but gives wrong output -- but this is probably something stackengine should solve.

So I have submitted my cstypo package to CTAN. \o/

@DavidCarlisle The following gives with lualatex from TL15 (and pdflatex all versions) true, but false with lualatex TL16. See also tex.stackexchange.com/questions/327153/…

\documentclass{minimal}
\usepackage{pdftexcmds}
\usepackage{graphicx}

\begin{document}
\makeatletter \GPT@IfFileExists{cours algebre.pdf}{true}{false} \makeatother
\end{document}

(naturally assuming that the pdf exists ...)

@egreg ^^^^^ -- in this case, the warnings are innocuous, but they shouldn't be there at all. the "q" is not requested from that font, only from \mathfrak.

(cont'd) the warnings do not appear with \boldsymbol or mathfrak or \widehat alone, or with a combination of any two. they appear only when all three are combined in the order \widehat{\boldsymbol{\mathfrak{q}}}. (if \widehat and \boldsymbol are swapped, no warning. i don't understand.)

@barbarabeeton You know that @DavidCarlisle is to be blamed

@egreg -- of course, but ... i've tested a bit more. if amsfonts is loaded instead of amssymb, no warnings, even with the order $\widehat{\boldsymbol{\mathfrak{q}}} $. so now i'm looking in amssymb.sty to see what can be determined. since \widehat is defined in the basic latex code, \boldsymbol comes from bm, and \mathfrak comes from amsfonts, i am for the moment totally baffled.

Could someone please confirm this with xcolor (I'm still using 2.11, so wanted to know if it was fixed) \definecolor{t}{HTML}{bb0000} works, as does \definecolor{t}{HTML}{BB0000}, but \definecolor{t}{HTML}{ff0000} failes and \definecolor{t}{HTML}{FF0000} works?

@daleif fails with TL15 but works in TL16 (xcolor 2016/05/11 v2.12). There is somewhere a question about it with an answer of egreg imho.

@UlrikeFischer thanks, I'll leave it alone then. Isn't using TL16 in production yet, want to finish a particular manuscript first. I'll just uppercase

Philosophical question: If your main text is in English but contains paragraphs in Czech, should the Czech bits be typeset using Czech typography rules or do the English typography rules apply even to the Czech language bits? :)

@wilx Sorry for the silly question, do they differ much?

@wilx I'd say you are mixing up language and locale: you are publishing this presumably in the Czech Republic so the local conventions apply, even if for example all of the text is not in Czech

I'd say, writing in another language binds the typographic rules to the language.

@wilx I'd say it depends on the typographic rules you are applying. When our students write their thesis in English and include a preface in Danish, I usually just recommend enclosing the Danish text in \begin{otherlanguage}{danish}...\end{otherlanguage}

A good example might be French vs English and the colon, doesn't French colon have extra spaces around it. There I'd say French rules apply to French text, and English to English

@daleif My thoughts exactly.

@daleif Interesting isn't it: we entirely disagree :-)

It's like writing a RTL language using LTR rules. :)

@JosephWright boo <3

@JosephWright So if You include a full paragraph (say a quote) in Danish in an English text, you'd use English hyphenation rules for the quote?

@daleif No, not hyphenation (which is a feature of language), but I'd use the conventions of the locale in which the book was being printed, for example in relation to "." vs "," for the decimal marker

@daleif My starting point is 'If I'd gone 100 years ago to a hot metal shop and asked them to set the work, what would they have done?'. My view is they'd use whatever 'rules' they always did, so based on the physical location of the shop not the language of the text.

@JosephWright hehe, eventhough Danish rules say , for decimal marker, I always instruct our students to use .

@daleif For technical usage I'd agree with that: simplifies things somewhat :-)

@JosephWright oooh

Metrical vs. Imperial... I like the sound of it.

@daleif Taking the French example, if I was in Paris and asked a local hot metal worker to set some English text, I'd expect they'd space ":" as they would for French

@barbarabeeton I made a smaller example (in the sense that it doesn't load an ams package). With it I even get errors is I remove one of the math before the one with \widehat. The problem seems to be the \frozen@everymath\@emptytoks in \@mathmeasure. Without it everything is fine -- looks as if the math is not correctly setup.

\documentclass{article}

\DeclareSymbolFont{AMSb}{U}{msb}{m}{n}
\DeclareMathAlphabet{\mathfrak}{U}{euf}{m}{n}
    \SetMathAlphabet{\mathfrak}{bold}{U}{euf}{b}{n}

\makeatletter
\providecommand*{\@mathmeasure}[3]{%
  \setbox#1\hbox{\frozen@everymath\@emptytoks
                 \m@th$#2#3$}}

\@ifundefined{@emptytoks}{\csname newtoks\endcsname\@emptytoks}{}

\xdef\widehat#1{\noexpand\@mathmeasure\z@\textstyle{#1}%
      \noexpand\ifdim\noexpand\wd\z@>\tw@ em%
      \mathaccent"0\hexnumber@\symAMSb 5B{#1}%

@JosephWright I would not.

@JosephWright -- but the hyphenation rules are not nicely separated from other "language" features by babel, unless i am sorely mistaken. and english-style hyphenation (either british or u.s.) is just plain bad when applied to many other languages.

@daleif This is admittedly a complex area as deciding what is driven by 'language' here is not easy

That would be arrogance ;-)

@barbarabeeton I didn't say babel gets this right: we'd discussed it to some extent in the team (my view is perhaps different from those of other team members)

@daleif Hmm...

@barbarabeeton I guess I see 'hyphenation' as one 'axis' here

@barbarabeeton, @daleif Of course, one issue is that today the idea of works being set in a particular physical location is not the same as it once was

@JosephWright -- i don't disagree with that. i'm just not sure how it's easily accomplished with the tools currently available.

@UlrikeFischer -- thanks. \frozen@everymath is defined in latex.ltx. i think it's time to call in @DavidCarlisle.

@barbarabeeton He's on holiday, giving bread to ducks and stepping on Legos. :)

@PauloCereda -- i know that, but i'm reasonably patient. (and really, bread is bad for ducks.)

@barbarabeeton <3

@barbarabeeton Well the problem is imho that the ams code is setting it to \@emptytoks (that means clears it), so it looks like it is your problem ;-)

@UlrikeFischer -- while it's most probably true that it's an ams problem, there's no hope that any of the ams packages will be fixed by ams in the reasonable future, so i'm hoping to get help from somewhere it might be effective.

@barbarabeeton I'd be more worried by this

\documentclass{article}
\usepackage{amsmath}
\usepackage{amssymb}
\usepackage{bm}
\begin{document}
$\widehat{\boldsymbol{\mathfrak{q}}}$
\end{document}

! Use of \reserved@a doesn't match its definition.
\reserved@b ->\relax
                     \ifmmode \else \non@alpherr \mathfrak  \fi \use@mathgro...
l.6 $\widehat{\boldsymbol{\mathfrak{q}}}
                                        $

@egreg -- yes, i'm seeing that in one of my experiments. and the question here is whether it's amsmath or amssymb that's the culprit; i suspect both, since that's what i got with @Ulrike's example when i swapped boldsymbol and \widehat (and no ams packages at all loaded, but the code for \@mathmeasure matching what's in both amssymb and amsmath. the documentation from the .dtx files is not helpful, at least not to me at the moment.

In a comment the op tells that he uses "sweave" and "RStudio". I do not know sweave, what is that? Is that a tex distribution?

Last line is great: Remember that many sources on the internet can be wrong.

@Kurt -- from the documentation: "Sweave is a tool that allows to embed the R code for complete data analyses in latex documents." it's included in tex live.

@barbarabeeton Ah, I see, thanks. Well, perhaps you can answer it?

@Kurt -- just because i know what it's for doesn't mean i know how to use it! i think @daleif is already on the right track.

@barbarabeeton Oh, okay, no problem. Thanks for answering.

@JosephWright If it's on the net it's surely right. /me taking the place of @DavidCarlisle who's not on line ATM

@egreg -- if i make just a couple of adjustments to that, it goes through without a whimper: change \usepackage{amssymb} to amsfonts (no names needed here), and move \usepackage{bm} to load before amsmath. once bm precedes amsmath, amsfonts can be in any relative order. clearly something is fishy, but it's going to take digging to figure out what.

@barbarabeeton @barbarabeeton is sweave actually included in Texlive? Had not seen that before usually users have to manually copy the sweave package from their R dist.

Damn, I'm sick again :( high temp, joints pain, backbone pain. I'm definitely donna get the lyme disease tests done

@yo' Oh no! :(

@PauloCereda yes, even last week I was considering it could be this, and the recurrence of the symptoms makes me worry a lot

@barbarabeeton I wrote the SASnRdisplay package that can handle the stuff that usually require the sweave package

@daleif -- i didn't check for details, but texdoc sweave brought up what looks like the "readme" file. however, sweave is not recognized at the command line. (on a linux platform with tex live 2016 fully installed.) so probably only the "support" is included.

@yo' -- for goodness sake, get yourself checked! and treated for whatever it is! the longer you wait, the harder it is to get rid of.

@barbarabeeton i do know

@yo': using Lilypond for typesetting percussion! It's very nice!

@PauloCereda it is, in both modes

@yo' The generated midi is also very nice! :)

@barbarabeeton @barbarabeeton You're basically disabling \bm

@StefanKottwitz Thank you. How many posts have there been at the end?

@Johannes_B I think it didn't raise after the 35. Stupid spammers, usually they would try to keep a low profile.

@Johannes_B I did not count as I removed the account with all its posts.

@egreg -- okay. i don't doubt that. what i'm trying to do is compose an understandable and helpful bug report that can be turned over to @DavidCarlisle, who is now the poor victim drafted to make sense of amsmath. since bm has his fingerprints all over it, this may be just deserts.

@StefanKottwitz I checke dnew psots and saw 32 new. Moment later it were 33. I pinged you here, got back. 35 posts. At this moment, i decided to write you a mail so you can act more quickly. I thought the spammer would go on and on and on and on and on ...

@Johannes_B at least it was not raising after I saw your mail

@StefanKottwitz Out of interest, can you bulk-delete? Or do you have to delete one by one?

@Johannes_B I can delete the account and its posts together (in some clicks)

@Johannes_B you can ban too btw.

@StefanKottwitz Would be insane otherwise. Click delete. Sure? Click yes. Ok, deleted.

@StefanKottwitz I think i can warn a user.

spammers don't deserve warnings, they won't notice it

@StefanKottwitz By the way, vietnamese spam. Seems to be over.

@Johannes_B yes, replaced by standard spam that is fewer

the question/answer captcha can prevent all bots, but not human spammers (otherwise nobody could enter)

@StefanKottwitz I asked a freind who is from vietnam to translate: Hi, spar dir die Energie, wir löschen alle Spamnachrichten. I sent a PM to every spammer. Might have been the same person, if the person ever read them.

@Johannes_B You could have said earlier: I could have put that sentence in Vietnamese to the registration page. (I got it similar in English there.)

@StefanKottwitz Oh, i didn't know.

@UlrikeFischer hmm spaces in filename pdftex is using \pdffilesize to test for existence which is one of the primitives the luatex manual helpfully says is removed as can be implemented in lua, without suggesting an implementation. luatex.def is using the code pdftex.def would use if \pdffilesize not defined which is to use \IfFileExists which apparently is falling over the space in the filename. I should probably fix that once I'm home:-)

@DavidCarlisle Miss you! <3

@PauloCereda you could use that as the basis for an answer to

@JasperLoy Thanks! :-)

@JasperLoy What about me? I really look like that.