« first day (3635 days earlier)      last day (1314 days later) » 

yo'
7:38 AM
@PabloGonzálezL that quite says it all... (Oh and good morning btw)
 
8:17 AM
@PabloGonzálezL 1pt not 0pt in my comment above. (\p@ not \z@)
 
I just saw in your log that you have a local ./beamerbaseboxes.sty, that probably interferes. — Ulrike Fischer 53 secs ago
@UlrikeFischer That's the whole point: going back to 3.51 to get transparent shadows
(unfortunately this new hook stuff in the kernel broke 3.51 :( )
 
@samcarter_is_at_topanswers.xyz Oh I didn't understand this.
 
Is there a way to temporarily get the old behaviour of atbegindocument?
 
@samcarter_is_at_topanswers.xyz I would need to understand the question first. When I compile this without any hacks the shadow looks ok to me. So what is wanted here?
 
@UlrikeFischer Without hacks the shadow goes from black to white and not to transparent (try with a background image)
There has been exactly one version of beamer (3.51) which implemented black-to-transparent shadows, but this did not work in xelatex and for some strange reason people seem to want to use xelatex instead of having nice shadows :)
I'm trying to understand why the shadows of 3.51 are stuck at the with of the first box. Maybe this whole making shadows global discussion?
 
8:34 AM
@samcarter_is_at_topanswers.xyz I don't think that this can be the problem: the first log used latex 2020-02-02, so no hooks involved. I would more suspect some of the improvements to shadings.
 
@UlrikeFischer It is certainly unrelated to the problem, but it would allow to test with a complete beamer 3.51, which I know worked and then narrow down the problem from there
(personally I suspect changes in pgf)
 
@samcarter_is_at_topanswers.xyz how do I change the color of the shadow?
 
@UlrikeFischer I think this might be hard coded
 
@samcarter_is_at_topanswers.xyz and how can I make it more simpler? e.g. only at the bottom of the box?
 
Ok, it works with an older pgf version
works with pgf 3.1.5b
@UlrikeFischer One has to redefine \endbeamerboxesrounded to change the shadow
 
8:59 AM
@samcarter_is_at_topanswers.xyz I would say it is basically the same problem as here tex.stackexchange.com/questions/445248/…. It works again if I use a cnt in the shadings/fadings declaration to make them unique:
\newcount\pgf@shadingboxcnt
\def\endbeamerboxesrounded{%
  \end{minipage}\egroup%
  \bmb@temp=\dp\bmb@box%
  \advance\bmb@temp by.5pt%
  \setbox\bmb@box=\hbox{\raise\bmb@temp\hbox{\box\bmb@box}}%
  \dp\bmb@box=0pt%
  \bmb@boxwidth=\bmb@width%
  \bmb@boxheight=\ht\bmb@box%
  \advance\bmb@boxheight by4bp%
  \advance\bmb@boxheight by\bmb@prevheight%
  \ifbmb@shadow%
    \global\advance \pgf@shadingboxcnt by 1
    \pgfdeclareradialshading{bmb@shadowball}{\pgfpointorigin}
    {%
      color(0bp)=(pgftransparent!50);
 
@UlrikeFischer Interesting!
OOOOOO, and it works with xelatex, too!!!!!!!!
@UlrikeFischer Will you post an answer to tex.stackexchange.com/questions/73960/… ?
 
@samcarter_is_at_topanswers.xyz my fading example in tex.stackexchange.com/a/445321/2388 now fails with pdflatex too, not only with xelatex. In texlive 2019 it still worked. I wonder what changed here.
 
9:22 AM
argggg, it does not work with xelatex (all these driver files lying around in pgf interrefered)
 
@samcarter_is_at_topanswers.xyz I think I found the reason. It is not the new latex but a change in pgf. The old, for pdflatex working version had a typo, and this meant that the mask was always installed, and after the correction it no longer happens if the name already exists:
\ifx \csname pgfsmaks@#1\endcsname \relax \pgf@sys@pdf@install@mask  %working pgf
                   ^^^ typo

\ifx \csname pgfsmask@#1\endcsname \relax \pgf@sys@pdf@install@mask  %failing pgf
 
(it is really inconvenient that one can't simply throw the whole repository of pgf into one texmf tree to test things with different versions)
@UlrikeFischer :) typos can be helpful sometimes
 
@samcarter_is_at_topanswers.xyz can you take the tds.zip from the releases and add them with tlmgr conf auxtrees add?
@samcarter_is_at_topanswers.xyz what doesn't work?
 
@UlrikeFischer ^^^ the shadow in the corners/sides is missing with xelatex
@UlrikeFischer probably, but with other packages I can just use git to checkout whatever commit I want without anything else to do.
 
9:43 AM
@samcarter_is_at_topanswers.xyz so the radial shading is missing. One should try to make a pgf only example to check if this shading works in xelatex.
 
quack
 
@PauloCereda good morning!
@UlrikeFischer I'll try to make one later
 
@samcarter_is_at_topanswers.xyz hi <3
 
how many potatoes to compile a document?
 
@wtsnjp おはよう!
@samcarter_is_at_topanswers.xyz I was planning to give a talk on how to compile a TeX document with my Nintendo Wii. I might come back to this wacky topic someday. :)
 
9:56 AM
@PauloCereda I'm so looking forward to this talk :)
 
@samcarter_is_at_topanswers.xyz Really? I will see if I can bring my Wii back to life. :) It's a rooted console, so I have an underlying Linux distro. It sounds fun to write and compile a TeX document with motion controls. :)
 
@samcarter_is_at_topanswers.xyz I forgot, the fix isn't on ctan yet. Try this:
\makeatletter

\def\pgfutil@insertatbegincurrentpagefrombox#1{%
  \edef\pgf@temp{\the\wd\pgfutil@abb}%
  \global\setbox\pgfutil@abb\hbox{%
    \unhbox\pgfutil@abb%
    \hskip\dimexpr2in-2\hoffset-\pgf@temp\relax% changed
    #1%
    \hskip\dimexpr-2in-2\hoffset\relax% new
  }%
}
\makeatother
 
@PauloCereda If there is an overfull box, just use the motion control to squeeze it back into size :)
@UlrikeFischer Thanks, I'll test
corners are back in xelatex!!!!
 
@samcarter_is_at_topanswers.xyz ooh :)
 
@PauloCereda If the overfull box is really bad, play Wii Boxing
 
10:01 AM
@samcarter_is_at_topanswers.xyz LOL
 
I kind of like the concept, if tex is driving you crazy, this will surely calm you down :)
The question is just how many pt the box will shrink per punch
@UlrikeFischer Is this already in the pgf dev version?
 
@samcarter_is_at_topanswers.xyz Imho yes, he closed the issue as fixed: github.com/pgf-tikz/pgf/issues/928
 
@UlrikeFischer Thanks! It indeed works with the dev version
... so when the changes to pgf make it to ctan, I'll try to get your changes to \endbeamerboxesrounded into beamer, would this be OK for you?
 
10:24 AM
@samcarter_is_at_topanswers.xyz sure, but be aware that it means that every block installs a number of new shadings, which increases both the number of tex commands and the xobjects in the pdf (as the hack did it previously too, it is probably not a problem, but imho it should be at least optional).
 
@UlrikeFischer Good, point! I'll make sure to also test with really long document
Maybe simply a new box, in addition to the shadowbox, a new transparentshadow box?
 
10:40 AM
Hello guys !
 
'ello
 
@BambOo Try natbib with the compress option
 
I just updated my miktex and while trying to test some OP MWE, my beamer throws me a `LaTeX Error: Command \AtBeginEnvironment already defined`.

My script is simply
```
\documentclass{beamer}
\usepackage{lipsum}
\begin{document}
\begin{frame}
\lipsum[6]
\end{frame}
\end{document}
```
Is this a new bug ?
 
works fine
check if you have some old files lying around
 
@BambOo you probably didn't updated everything.
 
10:44 AM
@samcarter_is_at_topanswers.xyz, Thanks ^^. I was about to provide another option, but I trust you ;). Are you spying on me ???
 
@BambOo :) no spying necessary
 
@UlrikeFischer I double-checked and my miktek console states every thing is up-to-date at both admin and user level
 
@BambOo can you show your log file?
 
@UlrikeFischer side note, an article class works fine
 
@BambOo Then show the log-file
 
10:49 AM
This is pdfTeX, Version 3.14159265-2.6-1.40.21 (MiKTeX 20.10) (preloaded format=pdflatex 2020.10.13) 13 OCT 2020 12:48
entering extended mode
**./test.tex
(test.tex
LaTeX2e <2020-10-01> patch level 1
L3 programming layer <2020-10-05> xparse <2020-03-03>
(C:\Users\bamboo\AppData\Roaming\MiKTeX\2.9\tex/latex/beamer\beamer.cls
Document Class: beamer 2020/07/22 v3.59 A class for typesetting presentations

(C:\Users\bamboo\AppData\Roaming\MiKTeX\2.9\tex/latex/beamer\beamerbasemodes.s
ty (C:\Users\bamboo\AppData\Roaming\MiKTeX\2.9\tex/latex/etoolbox\etoolbox.sty
 
etoolbox looks too old.
 
Package: etoolbox 2020/10/05 v2.5k e-TeX tools for LaTeX (JAW)
 
That's weird
Ok it works now, I had to remove/re-install etoolbox. Any idea why it didn't show up on the updateable package list ?
 
@BambOo \begin{joke} because you are using miktex and not texlive? \end{joke}
 
11:01 AM
That's what my computer said when trying to download the full texlive : \begin{deep voice} You shall not pass to texlive \end{deep voice}
 
@BambOo speak friend and install
 
hahaha :)
 
11:43 AM
@samcarter_is_at_topanswers.xyz the newest pdf version 2.0. And you can set it with expl3 in a engine dependant way (but doesn't work on the dvips route)
\RequirePackage{l3pdf}
\ExplSyntaxOn
\pdf_version_gset:n{2.0}
\ExplSyntaxOff
 
@UlrikeFischer Thanks! I'll pass this information on
 
12:39 PM
% arara: pdflatex
\documentclass{article}
\usepackage{lipsum}
\usepackage{microtype}

\begin{document}

\lipsum

\end{document}
Small question: The document above seems to fail with a recent TL when the TEXMF environment variable is set. It did not fail a month ago (in the same setup with the environment variable set). Does anyone know why?
 
@TeXnician what do you mean with "TEXMF" set?
 
@UlrikeFischer Well, in order to properly run the TeX Live ConTeXt I have TEXMF=/usr/local/texlive/2020/texmf-dist/ (i.e. set the environment variable to TL's distribution texmf) which always worked great. But after recent updates, the example suddenly fails. Without the TEXMF environment variable it still succeeds.
$ pdflatex example.tex
$ export TEXMF=/usr/local/texlive/2020/texmf-dist/
$ pdflatex example.tex
in a bash shell for example.
 
12:57 PM
@TeXnician I can't believe that that ever worked, it certainly fails for me in tl2019 and tl2020 on windows. I mean you are removing e.g. texmf-var from the search path. Are you sure you don't mean to set TEXMFCNF?
 
@UlrikeFischer Well, it definitely worked a month ago. I did not test with older TL. Probably I'll have to check again which variables ConTeXt accepts, so yes, maybe TEXMFCNF would be the right choice.
 
@TeXnician I can't believe it. texmf is declared as a long list in texmf.cnf:
TEXMF = {$TEXMFAUXTREES$TEXMFCONFIG,$TEXMFVAR,$TEXMFHOME,!!$TEXMFLOCAL,!!$TEXMFSYSCONFIG,!!$TEXMFSYSVAR,!!$TEXMFDIST}
if you shorten this, then you can't find the formats (it directly tried to recreate them for me).
 
@TeXnician You sure you didn't have a trailing : in the export?
like TEXINPUTS=/home/romano/lib/texinputs//:
 
1:33 PM
@UlrikeFischer You can test e.g. with the island's Docker image from September where it works (tag TL2020-2020-09-20-04-11). Thanks for searching. Now my setup definitely looks wrong and I'll look into it again.
@Rmano Yes, I'm sure, I'll have to recheck. Thanks for the pointer.
 
 
2 hours later…
3:39 PM
@PhelypeOleinik -- I just deleted my comment on tex.stackexchange.com/q/566544 so let's not add a chain of irrelevant comments there. (I didn't know about that package though, so thanks for the pointer.)
 
@barbarabeeton I deleted mine as well (and you're welcome :)
 
 
1 hour later…
5:07 PM
If I pass an ordered pair as an argument to a TeX macro, is that an easy way to obtain the first and second coordinates as separate values from that?
 
@FaheemMitha I think we need an example of your input and desire split
 
@JosephWright Ok.
 
5:52 PM
@FaheemMitha \def\zz(#1,#2){x is #1 and y is #2} \zz(11,22)
 
@DavidCarlisle zordered pair. :)
 
6:08 PM
@DavidCarlisle I'll post an example.
 
@PauloCereda \zz{duck soup,roast duck}
 
@DavidCarlisle oh no
 
yo'
6:43 PM
@DavidCarlisle Use of \zz doesn't match its definition.
^^ @PauloCereda you're saved :D
 
@yo' phew :)
 
@yo' hopefully I didn't do the same to the picture mode re-implementation this release....
 
yo'
@DavidCarlisle :)
 
@yo' can you be saved by a fatal error?
 
yo'
@DavidCarlisle oh definitely, unless you've already been expanded. And that isn't the case. FWIW, {duck soup,roast duck} has only been grabbed, but not expanded.
 
6:49 PM
@yo' but I could re-book the table for this evening \zz(duck soup,roast duck)
 
yo'
@DavidCarlisle :-(
 
documentclass[12pt]{article}
\usepackage{xparse}
\begin{document}

% #1 and #2 are ordered pairs
\NewDocumentCommand{\coord}{ m m }
{
#1, #2

% Ordered pairs which correspond to first value of first pair and
% second value of second pair, etc.
(#1[1], #2[2]), (#1[2]. #2[1])
}

\coord{(1, 2)}{(3, 4)}

\end{document}
@JosephWright ^^. Is this worth posting?
So desired result would look like: (1, 2), (3, 4).
Followed by (1, 4), (2, 3)
I can post it if it's not a dupe. Let me know. I didn't find anything, but that doesn't mean it's not there.
 
@FaheemMitha you can do it either way but why have the brace groups, wouldn't \coord(1, 2)(3, 4) be a nicer syntax?
 
@DavidCarlisle That's just the standard syntax for passing arguments to a macro.
I suppose TeX doesn't care, but it makes the rest of us feel more comfortable.
 
7:05 PM
@FaheemMitha not coordinate pairs though eg the standard latex \put(5,10){\line(1,0){10}}
 
@DavidCarlisle I don't follow. Could I have a complete example.
 
@FaheemMitha well tex cares a lot (you need an extra expansion step to strip the braces if you want the braced version), \put{(5,10)} is a syntax error.
@FaheemMitha I don't actually understand what output you want from \coord{(1, 2)}{(3, 4)} what is the etc in your description?
 
@DavidCarlisle I mentioned is after the example code.
 
@FaheemMitha yes but I have no idea how you got that output from that input
 
(1, 2), (3, 4)
(1, 4), (2, 3)
That's the output from \coord{(1, 2)}{(3, 4)}
The first line is just the input.
The second line is the first coord of the first pt, and the second coord of the second point. Followed by the second coord of the first pt, and the first coord of the second point.
 
7:11 PM
\def\coord(#1,#2)(#3,#4){(#1,#2), (#3,#4)\par(#1,#4), (#2,#3)} \coord(1,2)(3,4) I guess.
 
@samcarter_is_at_topanswers.xyz I think you can avoid to create to many shadings by using the dimension in the name so
\pgfdeclareverticalshading{bmb@shadowvert\the\bmb@boxwidth}{\bmb@boxwidth-4bp}
 
@UlrikeFischer Good idea!
 
 
1 hour later…
8:38 PM
This is very high on the list of most bizarre code I've ever seen:
\begin{figure}
    \begin{align*}
         \subfigure[\SI{3}{\centi\metre}]{\includegraphics[width=0.8\textwidth]{xxx.png}}
    \end{align*}
\end{figure}
 
8:52 PM
@samcarter_is_at_topanswers.xyz Almost as bizzare as \def\center{\trivlist \centering\item\relax} :)
 
yo'
9:04 PM
@AlanMunn so you tell me I could have used \item in place of \par in center all along?
 
@AlanMunn That code actually opens to more gruesome horror, like this:
Sep 13 '18 at 11:38, by Phelype Oleinik
      \begin{doublespace}
        \textbf{\small The \@doctype{} committee for \@author{} certifies that this is the approved
        version of the following \MakeLowercase\@doctype:}

        \hbox

        \hbox

        \textbf{\headingsize\@title}

        \hbox

        \hbox
      \end{doublespace}
(the whole thing is inside a center env)
 
@AlanMunn oh, that's also very good :) 10 points for creativity!
 
 
1 hour later…
10:22 PM
@JosephWright if I got it right then you never get ?? with an unknown label, only for unknown attributes. Is that intended?
thinking about it, that is the same with zref-default.
 

« first day (3635 days earlier)      last day (1314 days later) »