« first day (3595 days earlier)      last day (1633 days later) » 

02:54
@PauloCereda -- Wow! Never seen one of those!
 
3 hours later…
cis
cis
06:00
@VincentMiaEdieVerheyen I added some clearer help lines tex.stackexchange.com/a/560909/46023
cis
cis
06:31
If you look closely, the lines are sometimes slightly offset. That should be such a muddle with "pgflinewidth". In the end it doesn't matter, because the cyan line is set to 0mm. The specified numerical values ​​should be correct.
07:18
@yo' Does the layout of youtube.com/c/texusersgroup/videos looks odd to you too? I see an enormous logo at the top.
yo'
yo'
@UlrikeFischer yep, and it's being discussed internally \end{secret}
cis
cis
@yo' Mmmhhhh, was the projector presentation created with beamer.sty?
(Otherwise I can't look the movie) :()
yo'
yo'
07:35
@cis I dunno, you'd have to ask Paulo.
cis
cis
07:46
@PauloCereda Was the projector presentation chat.stackexchange.com/transcript/message/55433506#55433506 created with beamer.sty?
@JosephWright we can check for it, but should we? If it wants to undefine \pdfoutput shouldn't cslatex.ini delay it to e.g. \everyjob? I mean we requesting a number of primitives and imho \pdfoutput is one of them.
08:12
@UlrikeFischer politically we probably should. arguably we only require the functionality not the name, eg \outputmode, so \oripdfoutput just about meets requirements.
@UlrikeFischer do you want us to release oberdiek and hypdestopt ?
TikZ going for expl3 space rules: github.com/pgf-tikz/pgf/commit/… :D
4
@DavidCarlisle Robert seems to be satisfied. I hope I didn't forget anything ;-). I'm just looking if one can replace auxhook, imho it should be possible.
@yo' oh no :)
@cis the keynote was made with beamer, yes, modulo video thingy. :)
08:29
@DavidCarlisle I was thinking much the same
@UlrikeFischer but auxhook is already split so that can go out separately whenever it's ready?
@DavidCarlisle yes, no problem, and I should probably contact Markus, he seems to be the main user (or he works around it, didn't really checked yet).
cis
cis
@TeXnician That means, that I will get more rarely 'Dimensions too large'?
@TeXnician I had already introduced this before but it broke usage of pgfkeys in functions which obliterated pgfplots. This time it's optional.
@HenriMenke Ah, right, must have missed that at that time. Still, using ~ for spaces in TikZ looks unfamiliarly expl3-ish to my eyes ;)
08:36
\pgfkeys{/pgf/declare function/ignore spaces=true}
\pgfmathdeclarefunction{test}{1}{
    \pgfmathparse{sin(#1)}
}
\pgfkeys{/pgf/declare function/ignore spaces=false}
@TeXnician Yes, I definitely stole the idea from expl3.
@cis Not really. As linked in the linked issues it only solves some spacing problems (shifting plots).
@HenriMenke Definitely a good thing not to introduce yet another syntax :)
@HenriMenke hi! We need to write to you, expect an email from us soon. :) (the wacky people at the Island)
@PauloCereda Heralding the most informal way of communication, an email. This must be very important!
@PauloCereda Looking forward to your messaege :)
@HenriMenke you are important. :)
@PauloCereda (˵ ͡° ͜ʖ ͡°˵)
@PauloCereda What are your plans for visiting Germany?
@PauloCereda I'll be back there in about 2 weeks.
08:41
@HenriMenke as soon as the pandemic is over and it's relatively safe for international travels (along the lines of myself not getting trapped inside closed borders), I will go. :)
@PauloCereda Let me know when you're there, then I'll come to meet you.
@HenriMenke YAY
@samcarter_is_at_topanswers.xyz WOOOOOO
^^^ I propose this reaction for all IoT (Island of TeX) issues
@samcarter_is_at_topanswers.xyz motion approved
meeting is adjourned
ooh efficiency
08:50
@DavidCarlisle I see learnlatex.org stats are ticking along nicely
09:03
@JosephWright yes, and who knows there may be thousands of uncounted users using the overleaf option instead:-)
@DavidCarlisle can't one see the requests sent to overleaf?
@JosephWright I did wonder, as term starts, whether any course will tell a bunch of students to teach themselves latex there which could push up the numbers considerably..
@UlrikeFischer no, have no back end access to github pages or overleaf, I'm counting POST requests to latexcgi.xyz
> Heavy metal music attracts sharks. The low, rumbling frequencies of the music mimic the sounds of struggling fish.
yo'
yo'
@PauloCereda Remember that my offer still stands: you could stay for as long as needed in Prague (in case EU--BR travel got difficult suddenly), and European countries don't seem to make out-of-nothing decisions regarding border crossing.
@yo' <3
@yo' Prague is in the route! <3
yo'
yo'
09:13
@PauloCereda should be! And my country is beautiful as well, so leave enough room for that :-)
@UlrikeFischer, @DavidCarlisle Fix sorted: CSLaTeX is odd
@JosephWright :-)
09:31
@samcarter_is_at_topanswers.xyz nationalgeographic.com/animals/2020/03/…
@Rmano :) Interesting! I would have guessed something like 100 m just from the time it took between putting down a food bowl and the cat running into the garden.
@samcarter_is_at_topanswers.xyz Looking at the video seems that Sir Pratchett's joke is not completely a joke ;-)
@DavidCarlisle @PhelypeOleinik I made a new version of auxhook, imho it should work fine. The only problem currently is the use of \g__hook_execute_immediately_clist to detect if I'm "behind" the once-time hook. Do we miss a test here? And is the message actually needed (I copied it from auxhook, but it is perhaps a bit overdoing, imho people using this package should know where there are)?
\begin{filecontents}{auxhook.sty}
\NeedsTeXFormat{LaTeX2e}
\RequirePackage{xparse}
\ProvidesExplPackage {auxhook} {2020/09/03} {v1.7}
  {draft!!!!! Hooks for auxiliary files (HO)}

% to make KOMA happy ...
\let\@beginmainauxhook\relax
\let\@beginpartauxhook\relax

\NewDocumentCommand\AddLineBeginMainAux{m}
  {
    \clist_if_in:NnT \g__hook_execute_immediately_clist {begindocument}
      {
        \if@filesw
         \PackageInfo{auxhook}
           {
             \@backslashchar AddLineBeginMainAux~comes~a~little~late,
09:47
@samcarter_is_at_topanswers.xyz is there a fast way (apart for spreading \iffalse...\fi around) to compile just one frame of a beamer document? Overlays + pgfplots + Tikz makes compiling a bit on the slow side...
@Rmano have a look at \includeonlyframes{⟨frame label list⟩} from beamer
@samcarter_is_at_topanswers.xyz thanks!
@Rmano or have a look at the standalone class/package. This has also tools for compiling selected beamer frames
@Rmano I use the \includeonlylecture feature
@DavidCarlisle, @UlrikeFischer CTAN will be so happy with me :)
@JosephWright put "not my fault" into the announcement text.
09:52
@JosephWright put "cslatex is odd" in the annoucement.
@DavidCarlisle, @UlrikeFischer I see FMi is not so keen, but really it's not that different from collecting up the primitives in ConTeXt for generic use
@JosephWright well politically he is right: people are expecting too much that their odd code will continue to work only because it worked previously. It is not a good idea to remove a primitive without clear communication with us that we will support it.
@UlrikeFischer I can see the issue with older docs though: in the pre-pdfTeX days, \pdfoutput was a useful test for PDF vs DVI modes
@DavidCarlisle, @UlrikeFischer, @PhelypeOleinik Seems I have to track down some issue before a release anyway :) To do with the backend date ...
@JosephWright well if it breaks packages or documents it is not our problem (but imho most do an ifundefined{pdfoutput} first) ;-)
@JosephWright but fixing is imho ok anyway, but one should nevertheless communicate to cslatex why such an emergency fix was needed and that it would be better fixed at their end.
10:12
@samcarter_is_at_topanswers.xyz Yep, that's really handy. You put \includeonlyframes{workingframe} and then label workingframe the frame you're working on. Comment out and you have all the presentation. 1000 thanks!
@Rmano Worth knowing it does the same thing as the beamer core
@UlrikeFischer @JosephWright I don't think you will get changes at the cslatex side:
The CSLaTeX package is obsolete and it will be maintained no more.
This is our decision discussed on the CSTeX list

  groups.google.com/forum/?fromgroups#!forum/cz.comp.cstex

in October 2012.

The last two changes was done by Petr Olsak in Oct.2012.
@DavidCarlisle We could likely force the issue but I get what you mean
@DavidCarlisle well then it makes even less sense to add a fix. Norbert/Karl should simply disable the format.
@UlrikeFischer I don't mind if they disable it because it is unmaintained, but if it's disabled now then it comes across as disabled because we broke it. (the above is 00-README from texdoc -l cslatex by the way)
10:17
@UlrikeFischer :)
@DavidCarlisle mh yes. Perhaps they could change the ini?
@UlrikeFischer unlike other ini files that one comes from cslatex/base not from tex-ini-files so would require cslatex changes at ctan so ....
@DavidCarlisle Well, TL could just drop that one and have a modified copy
@JosephWright well yes but there is a rule (not followed all the time, I know) that ctan packages are taken as-is and any changes need to be made upstream at ctan
@UlrikeFischer is it in miktex?
@UlrikeFischer Hm... Maybe a \IfOneTimeHookUsedTF? Can't check now, but I think it's fairly common to check \ifx\AtBeginDocument\@firstofone, so that test would probably be needed.
10:25
@Rmano You're welcome!
@DavidCarlisle just gone shopping, will check later.
10:47
@samcarter_is_at_topanswers.xyz the d20 answer just gets more and more awesome
11:05
@DavidCarlisle there is an entry. In my settings it is excluded and so normally not built, but I don't know/remember if by default or because I disabled it.
@UlrikeFischer ooh shopping
11:21
@Anush Absolutely!
@samcarter_is_at_topanswers.xyz I'm hoping some of the efforts in l3draw and the related requests to the dvipdfmx authors will address XeTeX issues in beamer
@JosephWright That would be awesome!
@JosephWright you mean beamer issues in XeTeX?
@samcarter_is_at_topanswers.xyz Pretty cool that the mathematica people were trying to 3d print a d20 too :)
@samcarter_is_at_topanswers.xyz do you understand the answer well enough to know the answer to my question in the comments?
@Anush if you rolled a 12, wouldn't that be at the top?
11:30
@Skillmonlikestopanswers.xyz I mean the issues when you use beamer with XeLaTeX :)
@samcarter_is_at_topanswers.xyz yes, but if you were viewing the die from the top, it would be the front :)
@samcarter_is_at_topanswers.xyz yes but then you couldn't see it
@JosephWright when in doubt, beamer's fault (or @UlrikeFischer's?)
I use beamer + XeLaTeX and never ran into issues...
so we will have to assume we are looking from the top
11:30
@Skillmonlikestopanswers.xyz Oh, it's not: it's pgf or dvipdfmx or ...
@PauloCereda It comes up with a few special features, notes on second screen, some aspects of hyperlinks - to do with dvipdfmx interactions
@JosephWright no thanks, I keep my things easy peasy lemon squeezy. :)
@JosephWright nah, much easier to blame beamer!
@Anush If I understand correctly, it would be a matter of choosing the right \Angle to select the number which is at the front
@samcarter_is_at_topanswers.xyz yes that sounds right
it really needs a package :)
11:48
@UlrikeFischer I left the hyperref hook one for you:-)
@DavidCarlisle just answered ;-)
@UlrikeFischer yes
@Anush Now the 12 is at the top (and visible)
\begin{document}
\newcommand{\Angle}{250}
\tdplotsetmaincoords{170+10*cos(\Angle)}{\Angle}
\begin{tikzpicture}[tdplot_main_coords,line cap=round,line join=round,
	declare function={PHI=(1+sqrt(5))/2;}]
  \path[tdplot_screen_coords,use as bounding box] (-4,-4) rectangle (4,4);
  % light source
  \path[overlay] ({-sqrt(1/6)},{sqrt(3/6)},{sqrt(2/6)})  coordinate (L);
  % base color
  \colorlet{mypolyhedroncolor}{blue}
  % draw the d20
  \path[scale=2,transform shape] pic{icosahedron};
\end{tikzpicture}
@DavidCarlisle worse than \outer?! /end trying to be funny
@Skillmonlikestopanswers.xyz yes because outer you can largely ignore but mathchoice just gets everywhere: every constructed symbol or square root or phantom or ....
11:51
@samcarter_is_at_topanswers.xyz cool! I will have a play in a little bit
@DavidCarlisle notice how carefully I mention both the new hook system and that there is a -dev version (thinking about it, I could also sneak learnlatex in by mentioning the !TEX syntax)
@PhelypeOleinik I'm not completly sure if both tests have identical behaviour in all cases (didn't check) and the same "target" (I really want to know if the hook has been already used), but yes, I think the test is quite common.
12:11
@DavidCarlisle I know :(
@PhelypeOleinik actually the standard test is \ifx\@begindocumenthook\@undefined ;-)
12:31
@UlrikeFischer @barbarabeeton twitter.com/MrTimDunn/status/1301490066917908480
4
@Skillmonlikestopanswers.xyz 🐇 ❤️ 🥕
@samcarter_is_at_topanswers.xyz ooh
@PhelypeOleinik Turns out I can't count :)
@UlrikeFischer I should probably fix achemso in this regard too
@JosephWright do you do typesetting? Most things looked like definitions.
@UlrikeFischer \maketitle
12:47
@JosephWright you are adding it to \document, this is probably late enough, but using the new hooks should be better.
@UlrikeFischer Oh, yes, so I am: now I remember
@UlrikeFischer You remind me I still think \hook_log:n isn't quite right: doesn't interrupt the run with a prompt
@JosephWright why should hook_log stop the run? I would expect a message in log/terminal from a command with this name.
@UlrikeFischer \show does, \tl_show:n does, ...
@JosephWright sure, \hook_show:n should (if it existed), but why log?
@UlrikeFischer Well Frank's used it to implement \ShowHook ...
12:57
@JosephWright log I would expect to not stop, like \wlog but \ShowHook I would expect to stop
@DavidCarlisle Indeed
@DavidCarlisle Keeps catching
@JosephWright despite the fact that99% of users seem to run in scrollmode so it makes no difference:-)
@DavidCarlisle I think \ShowHook is for us mainly
@JosephWright ah, you mean \ShowHook should stop, yes probably, at least optionally it would be nice. But this would need an internal \hook_show:n and not a change to \hook_log:n imho.
@UlrikeFischer Fair
@UlrikeFischer I guess I should just do that
@DavidCarlisle, @UlrikeFischer Fix for CSLaTeX off to CTAN
13:02
@JosephWright I don't quite remember why Frank or @PhelypeOleinik implemented it like this, I think there was a discussion somewhere.
13:20
Hi mr. koala!
Oh no he's gone
is there a version without hspace of
\begin{equation}\label{eq:prob1}
\Pr(W > 0) \leq Np \hspace{0.5cm}\text{and}\hspace{0.5cm} \Pr(W = 0) \leq \frac{1}{Np}
\end{equation}
I just want the right spacing around the word "and"
@Anush \quad?
@PauloCereda Oh, \quad? Is that the sound of a duck who got the cookie stuck in the throat?
@Anush \text{ and }
@mickep ooh a conundrum
@DavidCarlisle bah :)
13:32
@DavidCarlisle that's not enough space sadly
@PauloCereda that does it. Thanks
@DavidCarlisle ^^ oh no
@Anush shrug, that's the font specified inter-word space
@DavidCarlisle ooh technical terms
yes but in display math it seems to look better with a larger space
It helps make it clear the "and" isn't part of either statement
13:53
TeXShop auto-completed xa to \alpha, xb to \beta, etc. This wont' work if I type $xa. The preceding $ prevents the auto-completion. Did anyone else get annoyed by this and is there a workaround?
Would such a question be on-topic on the main site? I assume not.
@Szabolcs I don't use autocomplete much with TeXShop, I find it more annoying than useful, but you can edit the autocompletion file itself so maybe you can add $xa etc for things you use a lot?
ok, that's kind of obvious
:-)
I should have thought of it myself
In fact, looking at the auto-completion file, I realize that it is possible to complete from \xa instead of xa, which is good enough for me.
@Szabolcs ooh success
14:04
I am failing to visually parse that. I see a bird with a stick in the beak :D
@Szabolcs ooh I better try Unicode :)
🦆
So I got it right, but the stick is your oar?
@DavidCarlisle ^^ I think the guy who wrote this table is not a Linux fan...
@Szabolcs ooh it could be a cricket bat
Mallard on iOS, wood duck on Android?
Seems so. :)
14:16
@Szabolcs why don't you input as $ xa?
@UlrikeFischer Nice spot on that \rm question - I guess the bug was from ~1993
Because I'm fussy and I hate to see that space in the source.
@JosephWright it has its use of that I checked the gnatsbug list some months ago ;-)
is it [t] as in \begin{column}[t]{0.5\textwidth} to have the text start at the top of column?
in beamer that is.. let me check the manual
14:37
@Anush To align the columns at the top, yes
@PauloCereda That Microsoft one could be a vulture or a pigeon, but it is definitely not a duck.
@Szabolcs agreed
@samcarter_is_at_topanswers.xyz thanks. Aligning a tcolorbox and text from itemize seems not so simple. There is t and T and also it seems you can use them in both columns and column in any combination. I will make a MWE :)
Friends, a quick Q: do you recommend any packages for customizing/styling chapters and sections? I will use memoir for this project...
15:06
@PauloCereda \def\section{\@startsection {section}{1}{\z@}{-3.5ex \@plus -1ex \@minus -.2ex}{2.3ex \@plus.2ex}{\normalfont\Large\bfseries}}
@DavidCarlisle I pushed in a branch a new auxhook version. There is one breaking change: I disabled the "internal" hook commands \@beginmainauxhook, they are not used anywhere (only in a test in the koma classes) and imho if someone wants an internal command they can use the LaTeX hooks. But perhaps you could take a look and comment ...
@PauloCereda memoir has its own lingo for this.
15:26
@DavidCarlisle ah :)
@egreg Thanks, I will take a closer look. :)
@DavidCarlisle thanks too :D
@UlrikeFischer They should have. The hook is added to the clist at the point it's used (it only actually becomes an one-time hook at the point it's used), which should have the same effect as the old \let\AtBeginDocument\@firstofone.
@UlrikeFischer auxhook misled me :-)
@JosephWright That version-check code will be hated in no time ;-)
@PhelypeOleinik :)
@UlrikeFischer @JosephWright Only a bunch of \iow_term:ns were added. I don't recall we discussing interaction for that. Though I agree that \ShowHook should be \hook_show:n and start interactive mode... Speaking of which, I should finish \sys_interact:...
15:49
@PauloCereda If you're using memoir you should us its built in functionality for formatting most things. Certainly chapters/sections and contents lists.
@DavidCarlisle Didn't you forget \let\section\zzz ?
Anonymous
@cis Thank you very much @cis I will look at it tonight or tomorrow very closely. Very busy day today. Greetings to you!
@AlanMunn Thought so, thanks Alan. :)
16:14
@JosephWright I get an error here with the newest backend files:
\RequirePackage{l3pdf}
\ExplSyntaxOn
\pdf_uncompress:
\ExplSyntaxOff
%
\documentclass[11pt]{report}
\ExplSyntaxOn
\pdf_object_new:nn {blblb} {dict}
\ExplSyntaxOff
\begin{document}
test
\end{document}
! LaTeX Error: Missing \begin{document}.
ooh die Räder vom Bus drehen sich rundherum
@JosephWright I pushed the fix, hope that is okay.
Anonymous
17:01
@cis Is this error to be expected:
Anonymous
17:47
@JosephWright I found another error and corrected it (but I wonder if the annotation_last and link_last command really need to be defined with x-type).
Anonymous
I don't seem to be able to compile, because of this error.
18:08
@VincentMiaEdieVerheyen that is a quite common effect of an error ;-) You should correct it. Is the code somewhere?
18:21
@UlrikeFischer l3build ctan ...
@DavidCarlisle which one? the fast or the slow one?
@UlrikeFischer slow at moment, but not as slow as it used to be:-)
@DavidCarlisle ;-) I wonder did I remove the tds.zip from the new package? (not that it would matter much).
@DavidCarlisle looking at the remaining packages. Could one integrate the dvipscol package into color/graphics-def?
18:40
@UlrikeFischer one of us did:-) ctan.org/incoming
@DavidCarlisle that's one of the nice thing of teamwork- you don't have to think at everything alone ;-)
@UlrikeFischer (@JosephWright) that optimisation (not using the stack at the top level) could be done for everything not just dvips I guess, what does the l3color version do?
@UlrikeFischer and if it goes wrong you are there to take the blame
@UlrikeFischer it is avoiding filling the color stack but ...
\documentclass{article}

\usepackage{color}

\newcount\zz
\begin{document}

\loop
\advance\zz1
\typeout{\the\zz}
\ifodd\zz\color{red}\else\color{blue}\fi
x
\ifnum\zz<10000
\repeat

\end{document}
18:59
@DavidCarlisle you mean one doesn't need it (anymore)?
@UlrikeFischer well I stuck 10000 colors on the stack without exceeding it, just looking at the dvips source code, I think that error isn't possible any more and it enlarges the stack, but need to check a bit more
@DavidCarlisle well then we could slice it out of oberdiek and declare it as obsolete.
how do you align equations that you reveal one transition at a time? As in
\documentclass{beamer}

\begin{document}
\begin{frame}{Frame Title}
\begin{itemize}[<+->]
\item[] Define r.v.\@ $Y_r = \sum_{j, f_j>0}I_{r,j}$. That is the number of elements in the stream that have at least $r$ zeros in their tail.
\item[] Let $T$ be the final value of $z$
\[
Y_r >0 \text{ iff } T \geq r
\]
\item[] \[
Y_r=0 \text{ iff } T \leq r-1
\]
\item[] \[
\mathbb{E}(I_{r,j}) = \Pr(zeros(h(j)) \geq r) = \Pr(2^r \text{ divides } h(j))
\]
\end{itemize}
\end{frame}
\end{document}
@UlrikeFischer That was to force the presence of a space
@DavidCarlisle Optimisation?
@PauloCereda -- Pack warm clothes. (@yo') And be ready to be shown a beautiful city. Ask if you can see the head of the dragon that was allegedly slain by St. George. (It's actually a crocodile; I'm sure of it; too chunky for an alligator or caiman or gharial.)
cis
cis
19:13
@VincentMiaEdieVerheyen Where does the error occur? If you compile the "New Version" from here https://tex.stackexchange.com/a/560909/46023 ?
This code runs on my system when I translate it into a New Document.
@JosephWright see texdoc dvipscol if etex tells you that you are at the top level group nesting then there is no need to do an \aftergroup restore or push the current color on the stack, the package claims this avoids running out of stack space in dvips but looking at the source I don't see the error (even at the start of the texlive svn in 2006)
@barbarabeeton ooh a crocodragon :)
cis
cis
@VincentMiaEdieVerheyen Try to delete the auxiliary files.
\documentclass{beamer}

\begin{document}
\begin{frame}{Frame Title}
Define r.v.\@ $Y_r = \sum_{j, f_j>0}I_{r,j}$. That is the number of elements in the stream that have at least $r$ zeros in their tail.

\pause
Let $T$ be the final value of $z$
\begin{align*}
\visible<.->{Y_r >0 \text{ iff }} &\visible<.->{T \geq r}\\
\visible<+(1)->{Y_r=0 \text{ iff }} &\visible<.(1)->{T \leq r-1}
\end{align*}
\pause
\[
\mathbb{E}(I_{r,j}) = \Pr(zeros(h(j)) \geq r) = \Pr(2^r \text{ divides } h(j))
\]
\end{frame}
@Anush Like this?
ooh that looks clever
thank you
19:20
@Anush You could move the 3rd equation into the align as well, but I did not how at which point to align it ...
@Anush \mathrm{zeros} or \operatorname{zeros}
@JosephWright ah, I thought you had done it to be able to define the pdftex and luatex version in one go. And the space was gone in luatex ;-) but I added it again.
cis
cis
@VincentMiaEdieVerheyen It works with original content as well at mine, in the difference to example-images.
\def\ImageScale{0.075} % <----- !
\def\ImageLeft{\UseImage{cmHgI}}
\def\ImageRight{\UseImage{cmHgI}}
@DavidCarlisle thanks
@samcarter_is_at_topanswers.xyz the trick is to understand . and +
19:40
@Anush . stands for the current overlay, + creates a new one
how is visible different from onslide?
@Anush \visible is like \onslide+ : the content is not show before, but still present
@DavidCarlisle I see Manfred handled it.
19:58
@UlrikeFischer yep
@UlrikeFischer found it!! if (strlen(p) + csp > cend) error("! out of color stack space") ;
@DavidCarlisle and what are the variables?
@UlrikeFischer you need to get dvips from ftp.cs.stanford.edu/pub/dvips/dvips576.tar.gz it's not in any version in texlive svn
@UlrikeFischer that version is dated 1997
@DavidCarlisle god, how did you find it?
@UlrikeFischer it seems like yesterday...
@DavidCarlisle and yesterday the problem was resolved?
20:09
@UlrikeFischer not sure if it's worth having at all, could just park the dtx in ho-tex/old-source are there really documents using it (@JosephWright^^^)
@DavidCarlisle we could try. If someone comes along and complains that it is gone we will know...
Hi everyone
Is there a TikZ bug or a document class bug here?
1
Q: Tikz: Different positioning behavior when using scrbook or book

schrottulkwhen embedding my tikz images into different types of document-classes, I experience some very annoying differences. I have provided the following MWE: \documentclass[ ]{book} %]{scrbook} \usepackage{tikz} \begin{document} \begin{tikzpicture} \draw [rounded corner...

Anonymous
@cis Yes the error is produced when I compile the 'New version:' (snippet of code most to the bottom) of your answer tex.stackexchange.com/questions/560883/…
Anonymous
Removed all auxiliary files and started with a fresh file name before compiling with 'LaTeX' in Texshop.
@AndréC According to the log file, the book class loads 10pt, the scrbook class 11pt font.
Anonymous
20:23
Anonymous
@cis Here is the .log file: we.tl/t-UCFqIdv8PN
@TeXnician True, but the TikZ code builds two identical (-4,0) rectangle (4,-3) boxes there is no reason for them to have different sizes when printed.
@AndréC I did not print them (only did on-screen comparison) but are you sure that the printer respected all margins and did shrink neither when printing?
@AndréC Try
\documentclass[
fontsize=10pt
%]{book}
]{scrbook}
\usepackage{tikz}

\newsavebox{\mybox}

\begin{document}

\savebox\mybox{\begin{tikzpicture}
    \draw [rounded corners, fill=green!10] (-4,0) node [anchor=north west]{xXxXxXxXxXxXx} rectangle (4,-3);
    \node[draw] (n1)  { node 1};
    \node[below of=n1, draw] (n2)  { node 1};
    \node[below right of=n2, draw] (n3)  { node 1};
\end{tikzpicture}}
\the\wd\mybox

\end{document}
You should not be seeing any difference between this and if deleting the option and uncommenting book.
20:38
@TeXnician You are right.
@DavidCarlisle -- But I think you'll find \quad recommended in the TeXbook for cases like this.
@PauloCereda -- The one on the upper left loos like a mutant dodo.
20:58
@barbarabeeton yes but @PauloCereda had already suggested that:-)
@barbarabeeton the linux one is a duck sandwich I assume.
cis
cis
21:16
All I see is "Runaway Argument".

So delete the auxfiles
and maybe comment in \usepackage{mwe} % Dummy Images
90% of the log is texcomb, so maybe get rid of \usepackage{texcomb}
and use \textcircled{c}
22:04
@JosephWright I played around with the new specials of dvipdfmx for transparency and I think I found a bug. If someone else (e.g.pgf or the pdfresources) add also an ExtGState value, it combines them correctly but the numbering of the names gets confused:
\RequirePackage{l3pdf}
\ExplSyntaxOn
\pdf_uncompress:
\ExplSyntaxOff
\documentclass{article}
\usepackage{tikz}
\begin{document}
blblbllb
\tikz[opacity=0.3]\fill[red](0,0)rectangle(1,1);
\special{pdf:bxgstate << /CA 0.2 /ca 0.2 >>}
   Transparent\newpage text.
\special{pdf:exgstate}%
\end{document}
and then the transparency after a page breaks fails.
 
2 hours later…
23:43
Is there a way to prevent TeX from stripping braces in certain contexts? Suppose I have a macro \def\test#1\endtest{... do something with #1 ...}, then expanding this as \test{foo}{bar}\endtest will preserve the braces but \test{foo}\endtest will strip them. Is there a way to avoid this?
23:58
@HenriMenke If it is for typesetting (or some context where an \@empty won't harm) then:
\def\empty{}
\def\test{\testaux\empty}
\def\testaux#1\endtest{... do something with \detokenize{#1} ...}

\tt\test{abc}\endtest

\bye
@HenriMenke Otherwise, an extra helper and a couple of \expandafter are enough to remove the \@empty:
\def\empty{}
\def\test{\testaux\empty}
\def\testaux#1\endtest{\expandafter\textauxi\expandafter{#1}}
\def\textauxi#1{... do something with \detokenize{#1} ...}

\tt\test{abc}\endtest

\bye

« first day (3595 days earlier)      last day (1633 days later) »