« first day (3620 days earlier)      last day (599 days later) » 

3:09 AM
@UlrikeFischer Rolf uploaded a new version containing fixed example ctan.org/pkg/eso-pic
 
 
3 hours later…
6:23 AM
@DavidCarlisle Ah, I see. Now I am wondering how many of the class files in the tree are actually intended for general use.
 
 
3 hours later…
9:42 AM
@UlrikeFischer I'm looking at the backend again; I think I'll start on a zref replacement
 
@JosephWright yeah ;-) but this isn't only backend isn't it?
 
@UlrikeFischer True, but we need some functionality for parts of the backend; life gets tricky with the PDF stuff, as I guess multi-run stuff is LaTeX-only but it has to be in the backend
@UlrikeFischer First I need to understand some of Heiko's ideas
 
10:02 AM
@UlrikeFischer The first things I'm trying to understand are (1) why does Heiko have 'properties' that are set up separately from lists (2) why does he use csname-restricted labels
 
Hi all, quick/hopefully simple question here
```
\begin{tabularx}{\linewidth}[t]{|X|X|p{3.9cm}|}
\textbf{Heading} & & Date \\ \hline
\multicolumn{3}{|<setting goes here>|}{\lipsum[1-10]} \\ \hline
\end{tabularx}
```
Within that multicolumn, how can I specify that I want that paragraph of text in the `multicolumn{3}` to be centered? Other stuff works out mostly, but with the paragraph, X type column simply overflows to the right, and general paragraph setting has its own issues

I'm hoping I just missed something simple on stackoverflow, but I'm a bit stumped here
 
In other news, HD space free: 5Gb!
 
@WynandMarais |p{\dimexpr\linewidth-2\tabcolsep-2\arrayrulewidth}|}{{\centering stuff \par}}\\\hline probably does what you want
@JosephWright out of how much?
 
@DavidCarlisle 250Gb
@DavidCarlisle Will be sorted soon I think!
 
@Dav
@DavidCarlisle That worked, thanks a bunch
 
10:09 AM
@JosephWright bigger disk or deleting half a dozen old texlives?
 
@DavidCarlisle Bigger disk by virtue of new laptop
@DavidCarlisle I think the TL archive is worth having
@DavidCarlisle With all of the video stuff at work, I think I'll want a decent amount of extra space
 
@JosephWright probably easier than changing the disk in an existing one:-) Yes the old ones are good to have but (since it's not work related) I do trim them when disk gets full I have ~75Gb out of 500 free but just go back as far as tl2016
 
@DavidCarlisle One for Skype ;)
 
@JosephWright (1) so that you can split up, which properties are saved and write special label commands. E.g. with \zlabel{sec}\zsavepos{pos} you get different labels in the aux. It would enlarge the aux, and slow down the use, if every label would always store every property:
\zref@newlabel{sec}{\default{1}\page{1}}
\zref@newlabel{pos}{\posx{8799518}\posy{45866949}}
@JosephWright (2) not quite sure what you mean. But there expandable tests, and the stuff is written to the aux, so both should work.
 
@UlrikeFischer I mean why two namespaces? If you want to do things on a per-'name' basis, why not a single global namespace for all 'label-like' things
@UlrikeFischer Heiko sets up so you get { pdf@abspage } -> \pdf@abspage, so you can only have [a-zA-Z@] in the 'property' name
 
10:24 AM
@JosephWright isn't that the same with keys of properties? I thought that it is where he got the idea from (or the other way round, I have forgotten).
 
@UlrikeFischer No: they are just detokenized 'stuff'. Based on the comments, Heiko was using a much older version of l3prop as inspiration, which might be it: the original plan used control sequences as keys and stored in toks.
@UlrikeFischer I'll have to think about this: key is that zref is I think the next 'blocker' for full integration of PDF resource support
 
@JosephWright what is wrong with combining some properties in a named list? In tagpdf I have such a list, which collect the properties I need to save, e.g. for each mc, and in the code I then set the label with \zref@labelbylist{labelname}{listname}.
 
@UlrikeFischer I'm just trying to get a handled on the API, really: to me if felt a bit awkward that you have \zref@newprop* then \zref@addprop, rather than a single creator function for each tracking item
 
@DavidCarlisle you could use our IoT historic images when needed and save disk space.
 
@PauloCereda easier just to ping Joseph or egreg and get them to test something
 
10:36 AM
@JosephWright but if you want to add a property to more than one list, then splitting it makes sense.
 
@DavidCarlisle that might work too. :)
 
@UlrikeFischer I guess that's what I'm wondering: does that really make sense? It relies on indpendent packages knowing each others property names
 
@JosephWright oh, you mean if I define a property "abspage" and someone else too we will clash? Sure, but how does it relate to lists?
 
@UlrikeFischer I was thinking you'd normally expect 'properties' to belong to exactly one list, so you'd have \zref@newprop{<list>}{<prop>}{<default>}{<value>}
 
@JosephWright no, you can add it too more than one. And e.g. listings uses the "listing" list, while figure uses the "figure" list, where both list share the counter property, but only figure stores say the float type.
 
10:49 AM
@UlrikeFischer OK, if it's a requirement; I'll think about how best to set it up (I think a one-shot is still useful)
 
@JosephWright I'm not sure if it is used much. But imho it is quite flexible, and I never had a problem with the general idea.
 
@UlrikeFischer how is your understanding of Chinese?
 
@DavidCarlisle bug report in chinese ;-)
 
@UlrikeFischer the title means "I think this place is redundant" which doesn't help much
 
@DavidCarlisle I think it is wrong, this is really an key to "addto".
 
10:53 AM
Somebody really wants to use \\ at end of paragraph, I just got downvoted on tex.stackexchange.com/a/82666/1090 :-)
 
@UlrikeFischer Agreed
 
 
2 hours later…
12:59 PM
@UlrikeFischer I'm not that old, personally would prefer Siezen as well, but then again, compared to my peers I'm more or less the only one who'd opt for it, I guess. Oh, and you don't have to decide, I'm just asking for your opinion so I can make a good decision (it seems like I'm the one who has to for some unknown reason).
 
1:11 PM
@UlrikeFischer I'm trying to fix the shipout stuff but the patch that you suggested doesn't work. With \put I get You can't use `\raise' in internal vertical mode.
 
@HenriMenke can you show an example? Then I will test with my patch version.
 
diff --git a/tex/generic/pgf/utilities/pgfutil-latex.def b/tex/generic/pgf/utilities/pgfutil-latex.def
index 191af23e..e9291d10 100644
--- a/tex/generic/pgf/utilities/pgfutil-latex.def
+++ b/tex/generic/pgf/utilities/pgfutil-latex.def
@@ -119,45 +119,16 @@
 }
 \let\pgfutil@abe\pgfutil@empty%

-\RequirePackage{everyshi}
-\EveryShipout{%
-  % Add at begin page stuff
- \setbox\@cclv=\vbox{%
-    \setbox\z@=\hbox{%
-        % the boxes \pgfutil@abe ("every page") and \pgfutil@abb ("current page")
-        % are used to generate pdf objects / dictionaries which are
This is the patch on top of the current PGF master
This is (your) MWE:
\documentclass{book}

\usepackage{tikz}

\begin{document}
\AddToHook{shipout/background}{\put(5cm,-5cm){%
  xxxx%
  \begin{tikzpicture}
  \shade[shading=axis,
       top color=yellow,
       bottom color=blue ]
       (0,0) rectangle (3,3);
   \end{tikzpicture}xxxx
}}

\begin{tikzpicture}
  \shade[shading=axis,
       top color=yellow,
       bottom color=blue ]
       (0,0) rectangle (1,1);
\end{tikzpicture}
\end{document}
 
@HenriMenke compiles without error for me, I need to compare the patches. Is the patched .def somewhere? that would make it easier for me to check the differences.
 
@UlrikeFischer Pushed to the shipout branch on my GitHub: github.com/hmenke/pgf/tree/shipout
 
1:27 PM
@HenriMenke wrong hook, you are using shipout/before instead of shipout/background. Beside this you don't have to go through the dimexpr/tonumber stuff, picture now accepts length so \put(\hoffset-1in,\voffset-1in) should work fine.
 
@UlrikeFischer Does that also work with current LaTeX?
 
@HenriMenke the picture stuff? Ah no, you are right, it requires like the hook stuff the new latex, and your put command it outside. But why are you changing also the old code to use a \put command?
 
@UlrikeFischer It's shorter and I don't want to scatter compatibility stuff everywhere.
 
@HenriMenke but does it work? I don't see a picture environment in everyshi.
 
@UlrikeFischer No, it doesn't work yet. I will have to tweak some stuff.
 
1:36 PM
@HenriMenke well you could create a second file. In hopatch I made a "hopatch-2020-XX.sty" which loads the old frozen code. That can then be easily dropped at some time.
 
@UlrikeFischer I went with the big branches after all anyway.
@UlrikeFischer The preview stuff should also be fixed then.
 
@HenriMenke well then you could simplify the \put command after all ;-).
@HenriMenke I don't see that you removed the preview patch?
 
@UlrikeFischer It's inside the branch.
@UlrikeFischer I've hidden whitespace changes. That's what the ?w=1 in the URL does.
 
1:52 PM
@HenriMenke Ah, wrong braces matched ;-). Looks ok. Do your test suite works with it?
 
@UlrikeFischer The test suite only runs on LaTeX release versions. I will have to test locally, but then again, manual doesn't use a lot of shipout trickery.
@UlrikeFischer Also, I'm not too scared of errors but of wrong output.
@UlrikeFischer Just had a look at github.com/pgf-tikz/pgf/issues/927 again. Why does this raise an error at all actually? Do the headers perform \scantokens? Because if \lhead stored the picture in a macro there should be no problem, because at that point $ has catcode 3..
 
@HenriMenke well you can then blame us ;-). But in earnest: I think the new state is much more stable then the old one with atbegshi, everyshi and who else fighting over shipout.
 
Re: #927 I think I understand now. The catcode check fires, even though the catcodes are already frozen.
That's a very tricky one.
Maybe just remove catcode checks altogether and let the users deal with cryptic error messages instead.
 
@HenriMenke yes.
 
@UlrikeFischer I'll make a release later (today or tomorrow)
 
2:07 PM
@HenriMenke well you could make a more specific test, according to the comment the reason was an active $, or you could make it a warning.
@HenriMenke nice, thank you.
 
3:00 PM
Hello. I hope everyone is well. I'm just dropping in between my classes.
 
@LaTeXereXeTaL Thanks! You too!
 
Students await. I'll return later.
 
user image
4
don't look at the comma :)
 
ooh teh comma
@Marijn congrats!
 
now I can finally write tag wiki usage guidance without supervision :D
 
3:26 PM
How on earth did we not know about this: gitduck.com @PauloCereda
6
 
@AlanMunn OH MY
 
@Marijn Did you know that @PauloCereda sends his credit card number to every trusted user?
 
@egreg oh no
 
@PauloCereda That's an ancient tradition!
@PauloCereda Since @Marijn has become trusted with a palindrome he's certainly qualified.
 
@egreg that's true :)
 
4:18 PM
Ouch... Syntax highlighting is not in its best moment... tex.stackexchange.com/a/564540/38080
 
@Rmano oopsie
 
@Rmano You can report bugs in as answers to the meta question meta.stackexchange.com/questions/353983/…
 
@AlanMunn Ah, thanks, will do.
@AlanMunn Although they seems to tell people to directly report on highlight.js repo, which will be overwhelmed in no time ;-) (unless they are a couple of hundred of developers, that is...)
 
4:36 PM
@Rmano I would report using the meta question anyway. It does seem to be a bit of a blame game.
 
4:51 PM
@PauloCereda QUAAAAAAAAAAAAACCCCCCCCCCCCCCCCCKKKKKKKKKKKKKKKKKKKK
 
@JosephWright ooh
 
@MartinScharrer Hi Martin! How are you?
 
@PauloCereda Ok, and how are you?
 
@MartinScharrer fine, I guess... :)
 
5:02 PM
@MartinScharrer Oooh, hello
@UlrikeFischer Starting on an approach as ltref.dtx: not sure if its lt or l3 just yet!
 
I pushed an update of filehookto CTAN yesterday to make it work with LaTeX 2020/10/01
 
@JosephWright writing to the aux requires probably lt. Do you already have a plan how to encode the property/value pairs there?
 
@UlrikeFischer Yes, vaguely: I'm thinking \SomeSafeCsname{key}{value} or similar, so the keys can be entirely arbitrary
 
About highlighting, on highlight.js site: «At best I'd consider our Latex support to be more of a "placeholder" than a real grammar... If someone would like to pitch in here to help I'd be more than happy to answer any HLJS specific questions about our grammars, etc.» See github.com/highlightjs/highlight.js/issues/…
 
@JosephWright do you think arbitrary keys are needed?
 
5:09 PM
@UlrikeFischer Not entirely arbitrary, but the logic should I think be the same as keys or prop or whatever: take the text, detokenize, use that
@UlrikeFischer I think numbers for example might be wanted
 
@JosephWright ah yes, that would probably be nice.
 
@UlrikeFischer I'll work on it for a few days then likely send you something to look over
 
@JosephWright ok
 
5:26 PM
@Rmano I think we should be worried that the maintainer wrote Latex instead of LaTeX...
2
 
@PauloCereda ouch :-(
 
5:48 PM
@PauloCereda Oh dear
 
6:07 PM
@UlrikeFischer Release 3.1.6 github.com/pgf-tikz/pgf/releases/tag/3.1.6
 
@HenriMenke woo
 
@PauloCereda This time with OpenPGP signed files.
 
@HenriMenke ooh
 
@PauloCereda Have you taken the five for Nix yet?
 
@HenriMenke I haven't got the time, but I am reading about it.
 
6:12 PM
You also have to thank @AlexG and @muzimuzhiZ for PGF/TikZ 3.1.6. They did most of the work.
 
@HenriMenke you could check you are in the output routine (eg \label is \@gobble or we could provide a more specific flag?) and skip the check there
 
@DavidCarlisle Sounds good but I don't want to pull LaTeX specific stuff into the TikZ implementation.
 
@HenriMenke ah true, but there's no good general test that you are processing already tokenized text so the catcode test will always be a bit suspect. You can only catch "common cases" and what is common is format specific. perhaps make it a warning not an error?
 
@DavidCarlisle I wanted to remove it entirely, in particular because people do not read error messages, so I don't expect anything less of warnings. The number of overfull boxes in my colleagues documents is frightening.
 
@HenriMenke OK with me, I think not reporting an error in some cases is better than the current situation that in some cases errors get reported in non-error situations.
 
6:37 PM
Welp. I can't put it off anymore, I need to pick a citation package for my tehsis
 
@Canageek \begin{thebibliography} I did this all myself nothing done earlier by other people matters.\end{thebibliography}
 
6:53 PM
@DavidCarlisle Yeah, this chapter is likely going to have over 20 citations and be the lowest number
(I did invent the technique myself, but we found similar things and cite some historical stuff in the papers)
Also I need to cite where I published this stuff
 
@Canageek I'd say pick a journal style and use whatever is convenient; are you doing per-chapter citations? (Sounds like it)
 
I have 119 citations in one of my other papers though, which will be the basis of Ch2
@JosephWright Yeah, I'm thinking RSC style, but that means I need to pick which rsc supporting package is the best.
But it means I can ignore how journals do it an actually cite books the way I want
(Ie the author of the chapter instead of the editor of the whole book)
 
@Canageek You can, but my advice is usually to follow an established style
@Canageek 'X, Y, "Chapter title" in Z, "Book title", ...'?
 
@JosephWright The style doesn't list guidelines for that, and ACS style SAYS to do it my way, but editors typically ignore it
 
@Canageek Ah, the joys of the editorial offices
 
6:57 PM
@JosephWright It will be whatever the ACS guidelines are, but with RSC name order, as the RSC instructions don't actually say how to do it and the ACS one make sense
 
@Canageek I think that other than journals, most editorial offices let you do whatever you like :)
 
It probably says something about the two groups that the RSC citation template is something like 12 pages and the ACS one is something like 120 pages
(The German one is apparently super short, but they just expect everyone to know their style already and follow it perfectly)
 
@Canageek :)
@Canageek Angewandte?
@Canageek For the RSC styles I write, I went with e.g. 'P. M. Henry, in Handbook Of Organopalladium Chemistry for Organic Synthesis, ed. E.-I. Negishi, Wiley Interscience, New York, 2002, vol. 2, ch. V.3.1.1, pp. 2119–2140.' plus a switch to turn on/off chapter title
(That one is actually in my thesis ...)
 
@JosephWright Yeah, I've not looked it up but heard about it from people
 
@Canageek The started from those guides, but most of the styles I've set up are based on empirical evidence ;)
@Canageek it is very short: I have read it
 
7:01 PM
@JosephWright I went with "J. Lefebvre and D. B. Leznoff, in Macromolecules Containing Metal and Metal-Like Elements: Metal-Coordination Polymers, ed. A. S. Abd-El-Aziz, C. E. Carraher, C. U. Pittman and M. Zeldin, John Wiley & Sons, Inc., Hoboken, NJ, USA, 2005, vol. 5, ch. 7, pp. 155–208."
WHich I think is your package?
@JosephWright You wrote most of the packages, which one would you recommend
 
@Canageek For ease-of-use, the biblatex version, particularly with per-chapter citations: running BibTeX on each .aux in a 'classical' set up is a PITA
@Canageek Also, you can more easily customise the biblatex one in-document
 
(Also, yes, per chapter citations, otherwise the list will be way to long. My papers have 118, 79, 8, and 13 citations
(Spot which two are chemistry and which two are 3D printing)
So "\usepackage[backend=biber,style=chem-rsc,doi=true,articletitle=true,pageranges=true,biblabel=dot,eprint=true,urldate=iso,date=iso,url=true]{biblatex}" would work then
Well, a few edits are needed to that, since it is from a document where I was making a list of citations for my group, not a paper
ALSO NO ONE TELLING ME I CAN'T PUT DOIs IN MY CITATIONS
 
@Canageek I'm not sure what you mean by 'way too long', but it's your call (I assume: I know some US places can be odd, but here in the UK and I assume in Canada it's pretty flexible)
 
@JosephWright Oh, just a pain to find things in. But I'm not sure yet, with LaTeX it shouldn't be too hard to flip between the two as I go, should it?
 
7:20 PM
@Canageek True, particularly if you go with biblatex
 
@JosephWright and no reason not to use biber, right?
 
@Canageek Only speed; Biber is slower, but with per-chapter it's much easier
 
8:11 PM
@JosephWright Is there an easy way to put the citation in the main text?
"For reproduction of material from all other RSC journals and books:
Reproduced from Ref. XX with permission from The Royal Society of Chemistry. "
Which means I need the Ref. XX in my text
\textcite adds the author
 
@Canageek Ref.~\cite{...} should do it
 
@JosephWright That will put in a superscript number
 
@Canageek Not unless you've altered the setup
 
@JosephWright Parts of this chapter are adapted from Ref. [1] Closer. How do I make the [] go away?
 
@Canageek Usual business, MWE needed
\documentclass{article}
\usepackage[backend=biber,style=chem-rsc,doi=true,articletitle=true,pageranges=true,biblabel=dot,eprint=true,urldate=iso,date=iso,url=true]{biblatex}
\addbibresource{Literature.bib}
\begin{document}
Some text \cite{Wright2016} and \autocite{Wright2016}.
\printbibliography
\end{document}
Works for me
 
8:25 PM
@JosephWright Complied your code (with a .bib file) got:
 
@Canageek Right, yes: have to remind myself how the brackets work :)
 
Ah, you were responding to the previous point, yeah, I got my citation packages mixed up
(The other setup I use has \cite give superscripts)
 
@Canageek ah, yes, the standard set up has
\DeclareCiteCommand{\cite}[\mkbibbrackets]
  {\usebibmacro{prenote}}
  {\usebibmacro{citeindex}%
   \usebibmacro{cite}}
  {\multicitedelim}
  {\usebibmacro{postnote}}
@Canageek So you'll want to re-declare without the optional bracket-adding
\documentclass{article}
\usepackage[backend=biber,style=chem-rsc,doi=true,articletitle=true,pageranges=true,biblabel=dot,eprint=true,urldate=iso,date=iso,url=true]{biblatex}
\addbibresource{Literature.bib}
\DeclareCiteCommand{\cite}
  {\usebibmacro{cite:init}%
   \usebibmacro{prenote}}
  {\usebibmacro{citeindex}%
   \usebibmacro{cite:comp}}
  {}
  {\usebibmacro{cite:dump}%
   \usebibmacro{postnote}}

\begin{document}
Some text \cite{Wright2016} and \autocite{Wright2016}.
\printbibliography
\end{document}
 
@JosephWright Thanks, my way was SUPER not working
@JosephWright PERFECT, thank you
 
@Canageek No probs
 
9:02 PM
user image
5
Ambigram!!!
Well, modulo the comma....
(goodnight!)
 
 
3 hours later…
11:43 PM
Welp, that is a new error to me; "syntax error: found "Quandt", expected end of entry ("}" or ")") (skipping to next "@")"
In the biblatex subsystem
Oh for the love of, there was an extra "
 

« first day (3620 days earlier)      last day (599 days later) »