« first day (3853 days earlier)      last day (1087 days later) » 

9:13 AM
how do you do "\def\mylist{$\infty$/null, "bar","baz"}" ?
the $\infty$/null part is not allowed
sorry I meant \def\mylist{{$\infty$/null,"bar","baz"}}
hmm.. scrap that it seems to be how I am using it
here is a M(not)WE example
\documentclass{beamer}
\usepackage{tikz}
\usetikzlibrary{arrows.meta}


\makeatletter
\newcommand*{\slideinframe}{\number\beamer@slideinframe}
\makeatother

\def\mylist{{$\infty$/null,"bar","baz"}}

\begin{document}



\begin{frame}
\frametitle{Demonstration}
\begin{tikzpicture}
\begin{scope}[every node/.style={fill=blue, circle,thick,draw, minimum size=1.3cm}]
    \node[label={[xshift=0em, yshift=-1em] \pgfmathparse{\mylist[\slideinframe-1]}\pgfmathresult}]  (A) at (0,5) {A};
    \node[label={[xshift=0em, yshift=-1em] \pgfmathparse{\mylist[\slideinframe-1]}\pgfmathresult}] (B) at (5,5) {B};
 
9:54 AM
(problem solved)
 
10:06 AM
New headset has arrived :)
 
@JosephWright is it nice? which one is it?
 
10:26 AM
@JosephWright you will have to tell me how good it is. I'm pondering about a wireless one too, my headset is not bad, but I constantly have to disentangle the cables.
 
@UlrikeFischer I'm on a meeting right now :) Will report back
 
 
1 hour later…
11:51 AM
Quack :)
@JosephWright ooh
Friends, out of curiosity, is anyone experimenting problems with font cache in LuaTeX?
Steps to reproduce:

1. I have a project which loads e.g, EBGaramond from a local `fonts/` folder via `fontspec`. It works.
2. I then have another project in other location which also loads EBGaramond from a local `fonts/` folder (not related to the first one).
3. LuaTeX fails. I have to delete the `texmf-var/luatex-cache` folder.
</opt/paulo/applications/texlive/2021/texmf-dist/fonts/opentype/public/ebgaramo
nd/EBGaramond-Italic.otf>
! error:  (file ) (type 0): cannot find file ''
!  ==> Fatal error occurred, no output PDF file produced!
It used to work a couple of days before.
 
12:12 PM
@PauloCereda why do you have the font in more than one location?
 
@PauloCereda Is any of these fonts the one referenced in the LuaTeX output?
 
@UlrikeFischer For some projects, I have fonts in a local folder and loaded via fontspec.
@MarcelKrüger Yes, EBGaramond is one of them, but it only exists in my local project, not in the system fonts.
@MarcelKrüger Actually, no, they aren't referenced in there. This one is from other location...
 
@PauloCereda How do you load the font in TeX?
 
@MarcelKrüger vvv
\setmainfont{EBGaramond}[
    Path = fonts/,
    Extension=.otf,
    UprightFont=*-Regular,
    UprightFeatures={SmallCapsFont=*-SC}
]
@MarcelKrüger I know that if I say that "it used to work before" isn't helpful... :)
@MarcelKrüger @UlrikeFischer I can come up with a MWE if it helps.
 
12:37 PM
@PauloCereda Given that there wasn't a luaotfload release in the last four months, "it used to work" "a couple of days before" can be translated to me being innocent. That seems like a very helpful comment. :P
 
@MarcelKrüger ooh :)
 
@PauloCereda That would be great, especially since it might explain how you end up loading the italic version.
 
@MarcelKrüger Let me see if I can write something. :)
 
@MarcelKrüger we could blame Joseph and the new siunitx ;-)
3
 
@UlrikeFischer Sounds like a great plan.
 
12:53 PM
@MarcelKrüger Will write a detailed report later. Thing seems more intricate than I originally thought. :)
@UlrikeFischer ^^ would you like to be CC'ed?
@MarcelKrüger for now, is there a quick way to invalidate the luatex-cache? I can delete the folder. :)
 
@PauloCereda does your document use some ebgaramond fonts which are not mentioned in your setup? E.g. the italic Marcel already asked about? Does it help if you add a ItalicFont declaration?
 
@UlrikeFischer I actually had the italic part, for some reason I omitted here, sorry. But I recall it failed yesterday with another font. Will write to you too. :)
 
@PauloCereda The name cache could be regenerated with luaotfload-tool -fu, but my guess is that your issue is with the font cache. Then removing the files is probably the best approach.
 
1:13 PM
@MarcelKrüger which variable is luatex using for the cache location? Paulo could perhaps set it by project and so keep the caches apart?
 
@UlrikeFischer Good idea. TEXMFCACHE(?)
 
@PauloCereda you could perhaps set texmfcache, see
% First writable tree here is used by Lua(La)TeX for the font cache.
% LuaLaTeX uses the value here, while ConTeXt uses the same variable but
% from texmfcnf.lua; therefore the two values should be kept in sync.
% (As should everything else in texmf.cnf <-> texmfcnf.lua.)
TEXMFCACHE = $TEXMFSYSVAR;$TEXMFVAR
 
2:00 PM
@MarcelKrüger @UlrikeFischer thank you!
 
Whoever thought it was a good idea to make bash syntax space sensitive should be [insert your favourite horrific end method here].
2
 
2:16 PM
@AlanMunn a fanatic Word user
 
@AlanMunn don't cast the first stone...
\documentclass{article}
\usepackage{tikz}
\tikzset{declare function={
        addok(\x,\y)=\x+\y;
        addfail(\x, \y)=\x+\y;
}}
\begin{document}
    \pgfmathparse{addok(1,2)}\pgfmathresult
    \pgfmathparse{addfail(1,2)}\pgfmathresult
\end{document}
;-)
 
@PauloCereda ?
 
@AlanMunn horrific end method? :)
 
@PauloCereda Ah, I'm not sure that counts as a method.
 
@AlanMunn ooh()
 
2:20 PM
@Rmano Well at least this throws an error. :)
 
user image
2
@JosephWright after releasing siunitx 3.0.1
 
 
1 hour later…
3:30 PM
@PauloCereda but presumably before dinner
 
Bavarian schools are mean: after more than a year mostly spend in lockdown, one of their final exam questions started with something like "Four families visit an amusement park ...." -- do the pupils still know what amusements parks are?
 
4:11 PM
@DavidCarlisle oh no
@samcarter_looks_forward_TUG'21 oh no
 
Does anyone know if, when collecting (importing) potentially large quantity of data from a file, does it make more sense, either in terms of speed or resource allocation, to import the data into a \def or into a \toks list? (I realize the answer is neither...instead process it as you import it. But let's assume that is not possible at the moment.)
 
well you could do it and use the pdftex timers to test. I would guess the toks register is faster (and the largest time likely to be opening and reading the file rather than tex expansion time differences)
 
@DavidCarlisle Thanks. That was my gut feeling as well, on speed.
@DavidCarlisle But I didn't know if one took up more memory or handles or whatever, than the other.
 
@StevenB.Segletes no the token list storage is more or less the same when there was only 256 registers you might have had to worry about using up one (unless you can use a scratch register) but now there are 32000 that's not really an issue.
 
@DavidCarlisle Thanks again for your valuable insights.
 
4:51 PM
The semester is finally over and I'm free to learn more about LaTeX and friends over the summer. It will be hard to beat last summer's learning.
 
@LaTeXereXeTaL ooh learning
 
@JosephWright for me, the new version of siunitx doesn't show Ohm in an old document of mine correctly :(
 
@Skillmon Ohm My!!
 
@Skillmon ohm no
@StevenB.Segletes ooh
 
@PauloCereda You beat me to it by about 5 seconds, as I scrounged around the internet for a unicode Ω.
 
5:01 PM
@StevenB.Segletes ΩΩH
3
:)
 
@JosephWright and putting \RequirePackage{siunitx}[=v2] before \documentclass results in an error:
! LaTeX Error: File `siunitx-abbreviations.cfg' not found.
 
@Skillmon Fixed in v3.0.1
 
@StevenB.Segletes <3
 
@Skillmon Could do with an example
 
@JosephWright good :)
 
5:02 PM
@JosephWright ooh
 
@JosephWright that might take a while (great font setup is great...)
 
@Skillmon Use the en.wikipedia.org/wiki/Om symbol instead :)
 
@JosephWright got one (that behaves differently, but still wrong):
\documentclass[]{article}

\usepackage{helvet}
\renewcommand\familydefault{\sfdefault}
\usepackage{mathastext}
\usepackage{siunitx}

\begin{document}
$R = \SI{1}{\kilo\ohm}$
\end{document}
 
yo'
I wonder, would it be illegal for a company or a non-profit to buy an ISO standard and put it on the internet?
 
@yo' yes
 
5:05 PM
@samcarter_looks_forward_TUG'21 I'm sure my chef would be delighted, but afraid my students wouldn't understand what I mean.
 
yo'
@DavidCarlisle Too bad. (Currently pissed off at twitter.com/PDFAssociation/status/1394907074249826304)
 
@yo' be nice, we're members:-)
 
yo'
@DavidCarlisle could we as members push for all of PDF2.0 to be free?
 
@DavidCarlisle ducks too?
 
@PauloCereda Maybe I just should have said "voltage", since Ω i (oh my) has units of volts.
 
5:07 PM
@yo' no, getting free access to iso standards is hard, At one point I was the nominated editor for an ISO draft and I couldn't even get access to the previous draft.
 
@StevenB.Segletes ooh that would work too. :)
 
@PauloCereda equals Ω³s
 
@DavidCarlisle OH MY
@Skillmon rabbits are good at equations
 
yo'
@DavidCarlisle ISO is ridiculous with this. I don't mean things like 27001, who needs to care about this has the money to buy it. But for things like PDF, Unicode etc. it's non-sense.
 
\documentclass{article}
\usepackage{helvet}
\renewcommand\familydefault{\sfdefault}
\usepackage{mathastext}
\begin{document}
$\mathrm{\Omega}$
\end{document}
@Skillmon ^^^
 
@yo' need or not need is not really the point. You need bread too and have to pay money for it.
 
@JosephWright oh no, sans serif!
 
@Skillmon 'It's not my fault'
 
yo'
@UlrikeFischer I get your point. But OTOH, if gov't, instutitions etc. make it official (e.g. PDF/A or PDF/X for theses), they should pay for making it so that it's free.
 
yo'
If a gov't forces you to comply with a standard that's not free, it's simply wrong.
s/false[/false]/ but that's doesn't answer your question :)
 
@JosephWright I can look later, but I was very impressed this morning to see a bug report of someone using siunitx in X-columns ;-)
 
@UlrikeFischer ooh duck-related bugs. :)
 
@yo' I bet gov't has a printed copy on display in a dark cellar, locked into a filing cabinet, without stairs, which you could consult :)
 
yo'
@samcarter_looks_forward_TUG'21 likely.
 
5:17 PM
@UlrikeFischer Thanks
@UlrikeFischer I know ... but it is fixed ;)
 
yo'
@JosephWright For the second one, have a kv for that and default it to $\,$ might work, or do you see some issues with that?
 
@yo' you have to buy lots of DIN norms in germany. And you normally don't need the norm to create pdf/A, you only need a tool which can do it, and perhaps a validator, and verapdf is free.
 
@JosephWright ooh
 
@yo' Nothing wrong per se: it's mainly a question of what is 'right' in the sense 'should using text mode even try to do math like spacing' and 'should some fonts get fixed'
 
5:34 PM
@JosephWright what is the input if \texttimes is used? And how is \texttimes setup? can I change is to \duck?
 
@UlrikeFischer YES PLEASE
 
@yo' yes but even things like ISO mathml which apart from the cover page is identical to W3C mathml will cost you CHF 198 I didn't even get a copy and I edited the thing.
 
@yo' -- Probably. It wasn't even possible for a member of an ISO working group to get a free copy of a relevant standard unless it was the one currently being worked on.
 
yo'
@barbarabeeton yeah, I know. Sorry, it was just a rant
 
@yo' there are copies to be found eg google for pdf 2 fdis but of course that doesn't really help if you want to cite them in any official way
 
5:41 PM
@yo' /duck hug
 
@yo' -- But believe me, it's shared.
 
yo'
@barbarabeeton thanks, that's appreciated :)
 
@UlrikeFischer It's automatically set if exponent-product is \times - I convert it to \ifmmode \times \else \texttimes \fi in the begin-document hook
 
@UlrikeFischer -- Last time I checked, bread, even really good bread, was more affordable than most ISO standards. At least Unicode has the decency to put the reference charts up for general access. Unlike ISO 10646.
 
@JosephWright perhaps you can make is configurable. Then people can add \, if they want. By default I would probably simply use \texttimes.
 
yo'
5:47 PM
@UlrikeFischer I believe there was a strong push for e.g. civil engineering standards to be for free here. IIRC it went up to the Constitutional Court here.
 
@UlrikeFischer Well yes - see the comments about dodgy spacing
 
@yo' well someone will have to pay. People work to do this and they want to eat too. If you make them free, who should give the money? The taxpayers? Industry? Donors?
 
@UlrikeFischer In a sense it already is configurable, that's the thing: if the user sets exponent-mode = \ifmmode \times \else \, \texttimes \, \fi, that will be left alone by the auto-stuff
 
yo'
@UlrikeFischer For legally-binding standards, the law-maker.
 
@UlrikeFischer I'd be surprised if selling standards covers a reasonable fraction of ISO costs and of the ones I have seen the people on the technical working groups don't get any money from working on standards, they cover the costs themselves or their employers cover it more or less the same as working on open standards
 
5:56 PM
@yo' "law-maker"...You mean "the taxpayer".
 
yo'
@StevenB.Segletes yes.
 
@yo' actually I was wrong in the case of MathML, the electronic version is free, the charge is for paper version.
 
yo'
Note that in the end we really pay for everything. The question is how do you distribute it. It's a difference if a global consortium pays for it and if a local company does.
 
6:23 PM
@DavidCarlisle the din writes that over 60% of the costs are covered through "eigene Erträge", so such sellings, din.de/de/din-und-seine-partner/din-e-v/finanzierung
 
@UlrikeFischer quoting documentation again, where will that lead?
 
@JosephWright point for you, but siunitx v2 was working for my document (which has a more complicated font setup, I'll try creating an MWE that is "better")
 
@UlrikeFischer actually though that is a fairly balanced page arguing that paying for standards isn't necessarily evil (@yo') din.de/en/din-and-our-partners/din-e-v/financing
 
7:11 PM
@Skillmon Well yes, I have changed all of the font stuff - I'll see if there is something I can auto-detect here
 
@barbarabeeton ooh bread
@DavidCarlisle disaster of nightmarish proportions and whatnot? :)
 
7:38 PM
@PauloCereda or dinner as it's known
 
@DavidCarlisle dinner needs documentation too, known as cookbook. I could quote some interesting receipts ...
 
@Skillmon One reason for providing \usepackage{siunitx}[=v2] is that I know that the font changes are non-trivial, and I couldn't be sure of supporting every existing use case
 
8:33 PM
@PauloCereda -- I keep telling you, bread is not good food for ducks! Unprocessed grain, or fresh greens.
 
yo'
8:45 PM
@barbarabeeton indeed!
 
9:20 PM
@barbarabeeton but does it affect the taste?
 
9:31 PM
@JosephWright well, font stuff is really nothing I'm too good at. But I'll try some other things myself. Don't expect me to be fast though. This has low priority for me (I'll just compile with TL19 for the time being)
Oh, or TL20.
 
@DavidCarlisle Karl put me on the secret list ;-)
 
9:47 PM
@UlrikeFischer against your will?
@UlrikeFischer is that so you can reply on hint? :-)
 
@DavidCarlisle yes, that seems to be his plan ;-)
 
@UlrikeFischer i'd wondered about forwarding that to team list, not sure how many of us are on the secret list that is a secret
 
@DavidCarlisle I think the question is a bit wrong, the question is not so much what he should provide but what to do with additional features: the backend code can always create a no-op or ignore an argument, but it is a bit of a problem to support features that the other engines don't have.
 
10:03 PM
@UlrikeFischer yes in that one and the other one Frank answered the answer is similar, in that ot depends whether he wants to forget all about pages or to implement them as delayed nodes that do something dynamically as the viewer paginates on the fly
 
@DavidCarlisle yes, I saw that too. I wonder what he would do with watermarks.
 
@UlrikeFischer or even simple things like table of contents, how do they get implemented if page numbers are not known (or don't exist?)
 
@DavidCarlisle little ducks linking to the section?
 
@UlrikeFischer only for restaurant menus
 

« first day (3853 days earlier)      last day (1087 days later) »