@Speravir Just a thought, without colors and boxes

\NewEnviron{simplequest}{%
  \setbox0=\vtop{\itshape
    \setbox0=\hbox{\BODY}
    \ifdim\wd0>\dimexpr\linewidth-8em\relax
      \advance\hsize-8em\noindent\BODY
    \else
      \hsize=\wd0
      \noindent\BODY
    \fi\par
    \xdef\tpd{\the\prevdepth}
  }
  \raisebox{-\dimexpr\dp0-\tpd-.2ex}{\LARGE\quotedblbase}\kern.2em
  \box0
  \kern.2em\raisebox{-1.2ex}{\LARGE\textquotedblleft}\par
}

\begin{simplequest}
Dies hier ist ein Blindtext zum Testen von Textausgaben.
Wer diesen Text liest, ist selbst schuld.
\end{simplequest}

\begin{simplequest}
Dies hier ist ein Blindtext
\end{simplequest}

Instead of \Large you can scale the quotes, so this becomes more independent on the current font size. Also the em in the definition for the textwidth should be changed to fixed units, perhaps.

Good night

@egreg Good night. I have to investigate this :-)

Is there any tool that removes line breaks from paragraphs in the .tex source file? My collaborator likes to break all the lines up at random points which I find very distracting, and I have to manually merge them back all the time.

Is there any tool for this? It's not a trivial find-and-replace type of problem because there are situations when line breaks should not be removed (e.g. at display equations, \item, etc.)

@Szabolcs Is he using version control? Maybe he's using the "1 line per sentence" rule, which makes diff work better.

@AlanMunn No, it's just the editor. He doesn't use diff. Actually this makes it very inconvenient to use diff. But latexdiff is still good.

Is there a way to have long squiggly arrows in LaTeX?

@OrangeHarvester Do you know this site? --> detexify.kirelabs.org/classify.html

@Szabolcs Yes. It only gives \rightsquigarrow. I want something longer comparable to \longrightarrow

@Szabolcs Most editors will wrap a selection or a paragraph, but I assume you want something more automatic than that.

@AlanMunn My editor wraps the paragraph at the window edge, so I like to keep a paragraph "on one line" (i.e. no linebreak characters). This way I can have the window as wide or as narrow as I like and everything is formatted nicely. But when I get back the edited version of the file, line breaks are scattered in it in random positions. I can't really ask him not to do it, unfortunately ... was wondering if there's an automatic way to undo these line breaks, which won't mess up the LaTeX.

But I don't think there is one. If there were, someone would already have replied.

I'll just fix it again by hand.

@Szabolcs I understand the problem, but not what you mean by "I'll fix it by hand". If you mean by manually deleting each newline, there are faster ways than that.

@Szabolcs Your advisor probably uses static word wrap on his editor.

@AlanMunn well, I select the pieces in Vim and press J ... the reason why I can't use one of the many online services written for this purpose is that there are special situations when a newline needs to be preserved: e.g. in front of an \item or a \begin{equation}

@OrangeHarvester I know. What can I do? :-)

Anyway, no big deal :) I'd better go back to working on the paper and not think about unproductive things

@Szabolcs Ok. Just checking. :) So yeah, that's probably as automatic as you can get.

@Szabolcs I think the editor kate has a apply word wrap and join lines sort of thing to deal with it. I am not sure how effective that will be. I guess there should be such tools for other editors too. (especially emacs :P)

@OrangeHarvester Check out the tikz-cd package. It might already have what you need.

@OrangeHarvester: Oh. You are here.

@AlanMunn Okay thanks. I will check it out. Also, just of the top of your heard, will it be fairly straightforward to include tikz in math mode equations without having to care about alignment and positions?

@GarbageCollector Yes. Had a TeX related difficulty.

@OrangeHarvester I think it is, and I've been trying to find you a question that discusses it. I'm pretty sure we have one on the site.

@AlanMunn Okay. Actually I did find one, but it did not re-assure me, (may be because I am ignorant about tikz).

@GarbageCollector Please ask before you comment about my intentions. It makes me appear as if I am not providing complete information, which will possibly piss off people.

@OrangeHarvester Sorry for my bad. :-)

@GarbageCollector Its okay.

@OrangeHarvester The commutative diagram package may not help you, unfortunately. But this question is a bit clearer on the baseline issue.

@GarbageCollector @OrangeHarvester If you both delete your comments about the MathJax then this all stays in chat. :)

@AlanMunn done. :-)

@AlanMunn Thanks for the link. It helps. I now know how to draw a squiggly arrow from another question. The baseline thing explains the rest. Thank you.

@PauloCereda Caveat! kpsewhich --var-value=TEXMFHOME is empty on my system and Ulrike Fischer reported this, too (better: she reported, and so I noticed). kpsewhich --var-value=TEXMFLOCAL points to my MiKTeX main tree. I think, the best would be parsing kpsewhich --var-value=TEXMF. See also additional note in tex.stackexchange.com/a/69484/9237.

@AlanMunn The comment has been removed by someone before I read your chat message.

@Szabolcs Bit late to the party, but tex.stackexchange.com/q/100/86 might have what you need.

@DavidCarlisle Answered a tikz-pgf question; I even used a tikzpicture environment.

@egreg pah! who needs rounded corners:-)

@DavidCarlisle And you could do them also with picture! Actually the tikzpicture is taken from @Speravir code without knowing what it does and some changes almost at random. :)

@egreg it's the only way with tikz questions. so why .66?

@DavidCarlisle It seems to work. Without it there are Overfull \hbox messages. Probably TikZ is adding some space somewhere. Who cares? ;-)

@egreg 0.666em is, of course, twice the default value of the inner sep of TikZ nodes (see page 177 of the manual =)). So if you want to do it "properly", you wouldn't hard code the value, but use \hsize=\dimexpr\dimen@-\pgfkeysvalueof{/pgf/inner xsep}-\pgfkeysvalueof{/pgf/inner xsep}\relax

@Jake Of course! I arrived only at page 176. :)

(surely there's a way to tell \dimexpr to do a-2*b?)

Hehe!

@Jake \hsize=\dimexpr\dimen@-2\dimexpr\pgfkeysvalueof{/pgf/inner xsep}\relax\relax

@egreg Ah, yes! Thanks

@Jake Simpler: \hsize=\dimexpr\dimen@-\pgfkeysvalueof{/pgf/inner xsep}*2\relax

@Jake Thanks!

@egreg Pleasure

TL?

@PauloCereda I would agree (OS dependent PATH environment setting). Vote for it ;-)

@StefanKottwitz Voted. :)

@PauloCereda Closed! Should also be fine for the OP, as he already fixed it.

@StefanKottwitz Danke. :)

@PauloCereda de nada

@StefanKottwitz ooh! :)

@StefanKottwitz cože?

Please, do you have any opinions on this?

@tohecz I think I'd go with the suggestion for Music. If you can't find a site that already has a decent density of questions on lilypond to show where the experts already are, then you're on a fishing mission and part of the key there is not to get thrown out of the pond. Music seems the best bet as a place to start.

@tohecz I'm with @AndrewStacey, as while LilyPond is developed by people with TeX experience, they don't use TeX (for good reasons)

What's the best way to mimic a boolean value?

\newif?

@PauloCereda Depends: are you loading anything else?

@JosephWright No. :)

@JosephWright ok, thanks guys

@PauloCereda In that case 'yes', unless you want to set up your own mechanism

@JosephWright Thanks!

@JosephWright:

If LuaTeX is a valid answer, then

\documentclass{article}

\usepackage{luacode}
\usepackage[T1]{fontenc}

\newif\ifunix

\begin{luacode}
function isUnix()
    if package.config:sub(1,1) == '/' then
        tex.print("\\unixtrue")
    else
        tex.print("\\unixfalse")
    end
end
\end{luacode}
\directlua{isUnix()}

\begin{document}

Hello, I'm running this code in a \ifunix Unix\else Windows \fi environment.

\end{document}

@PauloCereda there are also os.type and os.name functions in luatex

@PauloCereda I'd use ifplatform, but really a doc should not need to know this in general

@michal.h21 oopsie. :)

@JosephWright oopsie². :)

\ifwindows \font\myfont = comicsans at 10pt\myfont \fi

@topskip <3

\ifluatex \font\myfont = some_crazy_font_that_shows_all_crazy_otf_features\myfont\fi

\ifdvi \font\myfont = kmr10\myfont \fi

(kmr10 = knuth modern roman 10pt)

Hi everyone

@ThomasH Hello!

@ThomasH Hello

I'm writing my thesis using R, knitr, ctable, tabularx, pgfplotstable, pgfplots, siunitx, and lots more, but I'm constantly running into trouble with one or the other not working together. Anyone have a best practice approach for this?

@ThomasH IMHO the best approach would detect the conflicting parts, maybe reducing the whole code to a minimal working example.

Then I think it would be easy to solve. :)

For siunitx, let's blame @JosephWright. :)

@ThomasH: what kind of trouble do you get?

!!/cricket

@Paulo that's what I've been doing so far, but I'm wondering what the best general approach is, i.e. should I use xtable or hmisc to put tables directly into Latex, or should I write the table to a file and then typeset with pgfplotstable ...

@DavidCarlisle Hold on. :)

I've been using xtable, but it doesn't seem to use booktabs correctly, because my pgfplotstable (using top-, mid- and bottomrule) looks markedly different

but pgfplotstable doesn't seem to do tabularx columns

@ThomasH Ah I see.

!!/help

and no one seems to be able to do tablenotes

@ThomasH Which document class do you use for your thesis? :)

@David: let's test it.

!!/cricket

Psmith, the TeX bot, in fixed font mode: Let's take a look at the last cricket results:

- Victoria 536/9  v Queensland 322/10 &  60 *
- South Australia 240/10  v New South Wales 157/10 &  141/7 *
- Mountaineers 475/7  v Matabeleland Tuskers 142/4 *
- Southern Rocks 235/4 * v Mashonaland Eagles 352/10
- Madhya Pradesh 263/8  v Vidarbha 195/10 *
- Railways 231/10  v Rajasthan 232/3 *
- Andhra 228/10  v Karnataka 229/9 *
- Goa 219/10  v Kerala 223/4 *
- Hyderabad (India) 236/10 * v Tamil Nadu 262/9

report

You'd think Psmith might mention the odd international match in amongst all the local indian results bbc.co.uk/sport/0/cricket/21515161

@ThomasH what do you want to do with notes? if you wrap the table in a minipage any foortnotes come at the bottom of the minipage below the table

@ThomasH tabularx columns wouldn't make much sense for numerical tables anyway as they are for multiple line pargraphs of text, tabular* should be enough to allow inter-column spaces to stretch if needed

@DavidCarlisle remarks on individual items in tables (e.g. where results are based on less subjects because someone didn't answer a question)

@DavidCarlisle I have a few tables where some of the column names are too long and I want to wrap that text

while keeping all columns the same size

@DavidCarlisle @PauloCereda Thanks for the feedback so far. I gotta run but I'll be back in an hour or so and will read over the chat log

@ThomasH well as I say \footnote should work Ok for that although you might have to locally define it to be nothing at the point that you try to access the table data numerically if that's what you are doing. You should be able to have the main column be specified with dcolumn or siuntix or whatever then use \parbox[t]{some width}{your heading}

The main reason I haven't used the footnote/minipage workaround so far is because it adds another line above the footnotes right under the bottom line of the table, which isn't exactly pretty, but i'm guessing there is probably a way to suppress that

@ThomasH \let\footnoterule\relax after starting the minipage

@PauloCereda What did I do?

@JosephWright Nothing. :)

@AndrewStacey Thanks for the pointer, it looks like a nice tool! But I'm not looking for a one sentence per line format at the moment. I was glad to learn about that though, maybe I'll use that format next time.

@Szabolcs Looking at the code, you could fix that simply by removing "Rule 1" in the output routine.

WTF those flags!

@PauloCereda I hope, you got a notice of my message!?

@Speravir I did, it's in my TODO list. :)

@PauloCereda Fine. If you need feedback, how it looks in MiKTeX, askk us MiKTEX users … or set it up yourself in a VM or sandbox. ;-)

@Speravir Thanks. :) I need to work on more features as well. Heiko made some suggestions as well. :D

@PauloCereda Speaking of suggestions: How about outsourcing the arara directives into an own file and making some kind of input syntax? You could even hardcode, that the input file must have the string arara-directives or similar. Now the “issue” is, that on every change of arara directives the save date of the main file changes, too. There are cases, where this file date is evaluated, though.

@Speravir Oh my, I have to think about it. :) arara is growing up too fast. :)

@Speravir The main quality of arara is that the directives are in the main file.

@egreg I'm speechless, Claudio read the whole manual and wrote a review. :)

@PauloCereda Step by step, making it available to MiKTeX is more important! :-) BTW also @egreg I wanted to point you to a LaTeX package, where IIRC the file date is important, but I can’t find it now.

@egreg Let’s say, the main quality is, that it is in the source and can be manually influenced.

@egreg and @DavidCarlisle I will accept egreg’s answer, but if you have a solution with picture, add it as well. The rounded corners for me are not a “must have”. @egreg Also your solution from chat.stackexchange.com/transcript/message/8180396#8180396 worked, but the placement is not intuitive:

}

@Speravir That was only a sketch, but the main idea is the same.

Oops copy & paste with enter?

@egreg In both solutions parts are Voodoo for me. OTOH: You can safely remove the TikZ library positioning. It is not needed with your code.

@Speravir I like best the trick with \prevdepth that helps putting the opening quotes at the right height above the last baseline.

@Speravir Look on the site for "prevdepth user:4427` for other uses.

We can't really answer this

Or am I mistaken?

@egreg Will do. I just was going to write, that prevdepth is one of the Voodoo things, as you entered the message. :-)

@JosephWright The answer would be "possibly a very outdated TeX distro." :)

@JosephWright Since you logged in: Can you please remove chat.stackexchange.com/transcript/message/8193279#8193279 and chat.stackexchange.com/transcript/message/8193282#8193282? There went something wrong with copying and pasting.

@PauloCereda My point is that a lot of the code is proprietary, so even where TeX is involved there may be wrappers we can't know much about

@JosephWright Ah yes, you are right. :)

@Speravir Zapped

@JosephWright Thanks. I see now, that there is a lonely brace left. ;-)

@Speravir Do you want me to edit?

By the way, guys: FF19 has a built-in PDF viewer. :)

@JosephWright Yes, if not complicated.

@JosephWright I have a very similar impression about that one. This cannot be constructive, and no answer can really answer the question.

@PauloCereda It was included earlier, but it now activated by default. I would love a setting, where I could open PDF with the internal viewer, but when needed, could switch to a plugin.

@JosephWright I posted a comment for both the question and the (IMHO wrong) answer. Now excuse me, it's 3.3°C outside and I go on a bike in short-short outfit

@tohecz In my area the big people do use 3B2, I have to say, and although I used to see some dvips PDFs this is now rarer

@JosephWright yeah, my problem with that answer is probably the question if you know what I mean :-/ well, bye and see you later

@JosephWright I promised lover of structure that you'd solve all his/her problems, I hope that's OK tex.stackexchange.com/questions/96821/…

@DavidCarlisle xgalley's not bad, although further review is needed. You are free to work on xor.

@JosephWright yes but is there any point unless all constructs are using it?

@DavidCarlisle Of course not, but that's why I am keen to get things usable and out there

@JosephWright good plan

@DavidCarlisle I really would appreciate someone else looking over xgalley/l3galley: Bruno does not really know much about this area

@Speravir Before closing the \vtop, I store in memory the depth of the last (or only) line, so that I can shift the opening quotes disregarding it. Thus the placement will be (barring unusual lines) the same in all uses of the environment with respect to the last baseline. It's almost like having two reference points: the top and the bottom baselines; the former is for the closing quotes, the latter for the opening.

@JosephWright yes I need to get back to a state where I can breath, and then there's a minor issue about a project at work that's seriously overrunning and ....

@DavidCarlisle Understood

@Joseph would it make sense to migrate this now off-topic question to GraphicDesign?

@tohecz Probably still has the 'big list' issue

@JosephWright yeah, it seems to me so, quite.

34,000

@egreg Ah, thx.

@Speravir The result, as you can see is not identical to yours, but probably some tweaking with the TikZ parameters can help. However, as long as the multiline objects have the same width, one or to points more or less are irrelevant, arent't they?

@Speravir Some low level TeX programming can save labor. :)

@egreg Yes, I also think, this small difference not so important.

@egreg Yes, users like me are happy, if someone packs this in a good package.

@egreg Speaking of TikZ: Do you want to remove positioning yourself or should I do? This library was used in my code for above=<wide> of <anchor> or below ... respectively. In your solution these are not necessary anymore.

@Speravir I can remove it. I just didn't dare doing it as I never know what bits of TikZ one should load. And, as I told Jake, I'm still at p. 176 of the manual. ;-)

@egreg ;-) poisitioning is p. 185 … (typo intentionally not removed)

@JosephWright have you seen this:

@Speravir What did I say? :) Really, I'm happy that such a powerful tool exists, but mastering it requires too much labor on my side. I prefer to study LaTeX3, at the moment.

@bloodworks That colortbl is involved in the problem is highly unlikely. ;-)

@egreg Should we tease DavidC? ;-)

@Speravir no

@Speravir I've already had the privilege of finding one of the rare bugs in his packages.

Unfortunately, he didn't adopt Knuth's policy of sending checks to the bug finders. :(

@egreg I could follow the new Knuth policy of sending virtual cheques for virtual money of a bank I just made up, I could encode them in tikz but then you'd never be able to print them out

@DavidCarlisle Nor cashing them, either. ;-)

oops @Jake spotted @egreg being rude about TiKz and just flew in to defend its honour

@DavidCarlisle Maybe someone flagged my post as offensive.

@egreg Please take a look at xgalley :-)

anyone know what package the OP is using here `        \begin{subfigure}[t]{0.3\textwidth}` I tried `\usepackage{subfigure}` but get ! Missing number, treated as zero.
<to be read again>
                   }
l.10         \begin{subfigure}[t]{0.3\textwidth}

@DavidCarlisle subcaption

@DavidCarlisle TikZ! And PGFPlots! How will they make nice graphs?

@egreg ah.

(actually, only PGFPlots, it loads TikZ automatically)

Interesting blog post from EL&U about closing rates: english.blogoverflow.com/2013/02/…

But read the comments :-)

according to data.stackexchange.com/tex/query/98700/… we close about 9% of our questions...

@MartinSchröder now remove all obvious duplicates from that

@tohecz I leave that to the db gurus :-)

@JosephWright why does beamer redefine \reset@color to do an \unskip it makes a mess of leaders:(

@DavidCarlisle Don't ask me :-)

@DavidCarlisle Till set up all sorts of questionable stuff which I'm now stuck with

@DavidCarlisle What I said. :)

@JosephWright actually you could argue it's user error to use raw \leaders all latex ones protect themselves with a kern after for same reason (otherwise things like \newline which also do an unskip would eat rules (but a lot of users would probably moan:-)

@JosephWright In this answer the OP and the editor seem to be the same person; the system doesn't allow flagging twice an answer, so I report here: tex.stackexchange.com/a/98989/4427

@egreg lol, I've flagged in between, just in case

@egreg BTW @JosephWright and others I see a box below this answer “We're looking for long answers that provide some explanation and context. Don't just give a one-line answer: please explain why you're recommending it as a solution. Answers that don't explain anything will be deleted. See Good Subjective, Bad Subjective for more information.” Is this a bug or a new, unknown feature?

@Speravir Special mod message we can add

@Speravir Probably because the answer has been flagged as "low quality". But it doesn't apply any more.

Ah, thx both @JosephWright and @egreg. So, a till now unknown feature.

@DavidCarlisle After writing beamer apparently Till Tantau wrote pgfkeys and pgfmath rewrote TikZ and everything else. Then he droppped the support for beamer to Joseph and others in the current team. So probably he also knows that beamer needs to be rewritten completely.

@percuße yes I know, so no point blaming Till, simpler just to blame @JosephWright

@DavidCarlisle Oh a blame-a-mod-within-reach button.

@egreg: I suppose, that’s something you not disagree :-)

@Speravir Upvoted both comments. :)

@egreg Playing Paulo: <3

@NicolaTalbot glossaries.sty uses \equals in \glsgetgrouplabel, which should be \equal.