TeX, LaTeX and Friends

12:02 AM

@DavidCarlisle I just realised that is probably what is meant. Not that the function gets transformed but that its result does. I'm just being idiotic, I think. 'Function unique up to positive linear transformation' sounded to me like the function is transformed. But that's probably just me being stupid.

David Carlisle

@cfr well it's argualbly a transformed function if you give me the one above that maps marmalade to 3 and I make a function that maps it to 6,

cfr

@DavidCarlisle Yes, I agree. It just wasn't obvious to me. So you can double it but not add to it, right? That would no longer be linear?

David Carlisle

@cfr no you could say marmalade goes to 3f(marmalade)+2f(marmite) and marmite goes to 5f(marmite), which would be multiplying the result vector from f by the matrix [3,2;0,5]

@cfr what you can't do is multiply the result values by each other as that would be quadradic

cfr

@DavidCarlisle So if g is the transformed function, you could have f(marmalade) < f(marmite) but g(marmite) < g(marmalade)?

@DavidCarlisle Because, if so, this cannot be what is meant.

David Carlisle

@cfr yes multiply by -1 for example (unless there are other constraints like you can't do that)

cfr

12:15 AM

@DavidCarlisle Do you need that? What if you multipled the result vector by [1,1000;1000,1]?

cfr

12:31 AM

@DavidCarlisle Thanks. I just realised that if I look up something else in the book referenced, I find what I think the author must mean.

David Carlisle

@cfr :-)

cfr

@DavidCarlisle Now I just need to figure out what I should call it!

David Carlisle

@cfr I call it bed time, bye!

cfr

@DavidCarlisle Nos da ;).

Johannes_B

7:16 AM

I just noticed that you can rate package on CTAN. Why?

Joseph Wright

8:03 AM

@Johannes_B Meant to help people work out which ones are best: not really convinced

Joseph Wright

8:42 AM

@DavidCarlisle See you are in a good mood today

wilx

Or rate down those that are basically broken.

Joseph Wright

@wilx How many end users are going to check?

wilx

@JosephWright No idea. :)

David Carlisle

@JosephWright call a spade a spade

Joseph Wright

@wilx To a first approximation zero :)

@DavidCarlisle I was going to reply but you beat me to it!

@DavidCarlisle In other news, hopefully I can now do an expl3 release

David Carlisle

8:45 AM

@JosephWright I might do hyperref

wilx

I have just noticed the existence of cals packages. It seems to handle tables nicely.

Ousseynou Diagne

Hello everyone, I am glad to join the chat. I teach Physics and Chemistry and it would be great to any of you to point out to me resources or packages to get inspiration to draw Physics figures.

Joseph Wright

@OusseynouDiagne Depends on the sort of figures you mean!

@OusseynouDiagne There are several GUIs that good reviews, in particular Geobra, or one can use TikZ/PSTricks/Asymptote/...

Ousseynou Diagne

For example gravitation field

Gravitational field: Exemple drawing Force

ebo

9:52 AM

A dummy question, I'm lacking relevant keywords for finding a solution by myself: one can define a new length by multiplying an existing one by factor like this: \setlength{\mylength}{2\baselineskip}.

My question is: how to multiply by a factor that is the sum of two terms? (e.g. \setlength{\mylength}{{1.5+0.5}\baselineskip})

(I'm using this in the context of a tikzpicture. Of course, 1.5 and 0.5 would be there replaced by variables - i.e. I cannot add them manually)

wilx

@ebo Using \dimexpr?

6

Q: How to calculate a new length?

What I want to do is calculate the actual size between the edge of the paper, and my header. \evenmarginsize and \oddmarginsize do not do this. \settowidth{\msize}{((\paperwidth - \textwidth)/2)} Doesn't seem to work, how can this be done?

^ Maybe this helps.

ebo

Seems to be the answer indeed! I couldn't find correct english keywords to reach existing questions... Thank you.

David Carlisle

@egreg (@WillRobertson, @JosephWright) looking back at how existing answers are affected by the new fontspec/kernel....

4

A: How can I substitute a command when a particular character is not present in a font?

A better solution requires two steps: first you have to undeclare the composition \=y and then redefine ȳ to do the composition. \documentclass{article} \usepackage{fontspec} \usepackage{newunicodechar} \setmonofont{Latin Modern Mono Prop} \UndeclareUTFcomposite{x0233}{\=}{y} \newunicodechar{ȳ...

@egreg's answer now fails with an undefined command as UndeclareUTFcomposite is defined by xunicode not really fontspec and so doesn't get defined with new tuenc default.

egreg

@DavidCarlisle Well, if you mix everything up… :-)

David Carlisle

We could do it using the new variant encoding setup, but I wonder if fontenc ought to define UndeclareUTFcomposite if only for compatibility (or if @egreg just needs to update the answer to load xunicode)

egreg

10:07 AM

@DavidCarlisle Probably it's better to update the code than loading xunicode.

David Carlisle

@egreg change always breaks something, it's true.

@egreg yes but update fontspec to define UndeclareUTFcomposite or use the new variant? My original reason for searching for that was to test you could do same with the new code declarations, but wasn't expecting to find the old one didn't actually work any more

yo'

@ebo also, if it's in TikZpicture, look into the tikzlibrary calc

egreg

@DavidCarlisle Having \UndeclareUTFcomposite could be helpful, but probably declaring a new encoding is easier.

yo'

@egreg @PauloCereda I had a dream that I browse TeX.SE and there was a question by @DavidCarlisle !

egreg

@yo' Nightmare!

@DavidCarlisle By the way, is there the possibility to undeclare a composite to begin with?

David Carlisle

10:12 AM

@egreg should also be better if there are some fonts that have the composite and some not in the same document, but any idea how much UndeclareUTFcomposite was used, you've promoted it a bit:-) It might be best to have that as a compatibility thing

ebo

@yo' Thanks. I had a look though, but I didn't succeed in implementing a light way to compute my factors...

Paulo Cereda

@yo' ooh

David Carlisle

@ebo \setlength{\mylength}{\dimexpr 1.5\baselineskip+0.5\baselineskip}

egreg

@DavidCarlisle Oh, yes, there is \UndeclareComposite.

David Carlisle

@egreg well not really, tuenc.def is already input into the format so the declaration has been done , so the idea of the new encode is to declare a variant encoding without it, but there isn't really a way to change tu encoding itself unless I missed something in @WillRobertson's code

@egreg yes but that's restricted to use in an encoding setup

@egreg see my 2017 update version here

3

A: LuaLaTeX is not rendering 'ť' character with Andika New Basic font

You can do \documentclass[paper=A6]{scrartcl} \usepackage{fontspec} \defaultfontfeatures{Ligatures=TeX,Scale=MatchLowercase} %\setmainfont[]{Andika New Basic} \setmainfont[]{AndikaNewBasic-R.ttf} \usepackage{newunicodechar} \newunicodechar{ť}{t^^^^030c} \title{Říkanky} \date{} \begin{document...

which would I think be the "new" way to do your answer linked above, but for the end user, just keeping the previous answer working might be good.

@egreg oh so yes you can redefine TU so :

%\UndeclareUTFcomposite{x0233}{\=}{y}
\DeclareUnicodeEncoding{TU}{%
\input{tuenc.def}% load usual definitions
\UndeclareComposite{\=}{y}% This variant has no \=y
}
\newunicodechar{ȳ}{\=y}

egreg

10:30 AM

@DavidCarlisle On the other hand, with XeLaTeX, the ȳ problem doesn't show up any more, because of Harfbuzz. Just an adjustment to \url is needed. It's needed with LuaLaTeX, though.

David Carlisle

@egreg yes. same as the \v t case in the Andika font question, (xe)tex moans about a missing glyph in the log then passes the string to harfbuzz that finds (or rather composes) a glyph anyway

egreg

@DavidCarlisle Yes. I'm updating the answer.

David Carlisle

@egreg thanks. But do you think @WillRobertson should update fontspec...

egreg

@DavidCarlisle Always. ;-)

David Carlisle

@egreg was a serious question for once:-)

egreg

10:36 AM

@DavidCarlisle To add \UndeclareUTFcomposite as an alias for \UndeclareComposite? Possibly so.

David Carlisle

@egreg yes sort of a globally acting variant rather than an alias (and with the same (useless?) first argument that gives the code point.

egreg

@DavidCarlisle Oh, well: tuenc.def doesn't define a \=y composite for ȳ. :(

David Carlisle

@egreg ah that would explain something I was worrying about. I wonder why not (I thought I'd used all the latin composites that were mentioned in inputenc or t1enc or the orgial tuenc:( let me check...

@egreg, inputenc utf8enc has no declarations using \= nor Will's original tuenc in last years fontspec.

egreg

@DavidCarlisle It should.

David Carlisle

@egreg agreed...

egreg

10:50 AM

@DavidCarlisle On the other hand, if I do \DeclareTextComposite{\=}{TU}{y}{"0233} and then \newunicodechar{ȳ}{\=y}, I get the combining accent, not the proper glyph.

David Carlisle

@egreg you do? why?

egreg

@DavidCarlisle Let me make a small example. Wait a minute.

@DavidCarlisle It's quite strange: if I try ā, then I always get the composite accent (also with euenc). Not with, say, é.

David Carlisle

@egreg I was economical with the truth about inputenc having no \= I was looking for \= instead of its more friendly name \@tabacckludge= but it is missing

U+0232 \=Y
U+0233 \=y
U+1E20 \=G
U+1E21 \=g

Paulo Cereda

@DavidCarlisle you were wrong?! /gasp

David Carlisle

@PauloCereda ?

Paulo Cereda

11:02 AM

@DavidCarlisle you are naughty :)

David Carlisle

@PauloCereda British political quotation:-)

Paulo Cereda

@DavidCarlisle so I see. :)

egreg

@DavidCarlisle Can you try the following? I always get “a with combining macron” and “é” (not combined)

\documentclass{article}
\usepackage[
  %euenc
]{fontspec}
\setmainfont{Latin Modern Roman}
\begin{document}

ā é

\=a \'e

\end{document}

David Carlisle

added showoutput and with luatex:

....\TU/LatinModernRoman(0)/m/n/10 ā
....\glue(\spaceskip) 3.33 plus 1.665 minus 1.11
....\TU/LatinModernRoman(0)/m/n/10 é
....\penalty 10000
....\glue(\parfillskip) 0.0 plus 1.0fil
....\glue(\rightskip) 0.0
...\glue(\parskip) 0.0 plus 1.0
...\glue(\baselineskip) 4.91
...\hbox(6.98+0.11)x345.0, glue set 317.22998fil, direction TLT
....\localpar
.....\localinterlinepenalty=0
.....\localbrokenpenalty=0
.....\localleftbox=null
.....\localrightbox=null
....\hbox(0.0+0.0)x15.0, direction TLT
....\TU/LatinModernRoman(0)/m/n/10 ā

so both pairs using precombined glyph slots?

@egerg same with xetex, no combining accent in the log

@egreg have you updated today (to get the latest fontenc)

egreg

@DavidCarlisle Yes

David Carlisle

11:18 AM

@egreg er can you add \listfiles and \showoutput and send me the full log is the usual response to an OP at this point. (Along with, too much code for chat, please ask as a question on the main site)

@yo' ^^

@egreg a full log by mail would be good:-)

egreg

@DavidCarlisle Maybe it's just a fault of the previewer. Don't worry.

yo'

@DavidCarlisle pushy

David Carlisle

@egreg you should use windows (then you could use xpdf like me:-)

egreg

11:32 AM

@DavidCarlisle I can use xpdf. :P

Ulrike Fischer

@DavidCarlisle @egreg: Imho an encoding-depending \Undeclare ist better than some general one. But it would be good if one could do it without have to reload e.g. \input{tuenc.def}. So something like \UndeclareCompositeForEnc{TU} ...

egreg

@UlrikeFischer Something like \DeclareEncoding[base=TU,name=TUx]{...}?

David Carlisle

@egreg don't tell Will:-) one version he had a specific interface to specifying a base encoding, but it wasn't clearly the right features so for this first version we went for the simplest thing of a generic code argument with \input{tuenc.def} if you needed that.

@UlrikeFischer but the idea in the new encoding code is that you should define a new encoding rather than change an existing one (which is why tuenc.def never has TU literally always via macro so when you declare TUx and input tuenc.def it defines a new TUx encoding

egreg

@DavidCarlisle With the proposed syntax one can maybe do cascading definitions.

David Carlisle

@egreg exactly, but we were out of time to work out exactly what that is supposed to mean

Joseph Wright

12:01 PM

@DavidCarlisle 'Update the docs'

Paulo Cereda

@DavidCarlisle: you could apply for the GuIT board of directors

Ulrike Fischer

@DavidCarlisle Hm. Yes I think it is cleaner. But some cascading interface as suggested by @egreg would be usefull in the end. If you declare a new encoding how easy is it to switch to it for all text fonts in a document?

David Carlisle

@PauloCereda Sono sicuro che sarebbe onorato di avere me.

Paulo Cereda

@DavidCarlisle :)

wrong smiley :)

David Carlisle

@UlrikeFischer easiest would be to redefine TU as above,

@UlrikeFischer but one reason for sticking to a very basic interface this time round is to see what use cases come up

@JosephWright fontspec ones?

Ulrike Fischer

12:10 PM

@DavidCarlisle Yes. It is interesting to ponder about the implications.

David Carlisle

12:40 PM

@JosephWright one for you:-) tex.stackexchange.com/questions/350868/…

Joseph Wright

@DavidCarlisle See my comment

@DavidCarlisle Ones using xunicode features

David Carlisle

@JosephWright oh OK, yes especially where they were implicitly documented as fontspec features.

@JosephWright seems reasonable enough:-)

Christian Hupfer

12:58 PM

Is there a long .bib example file in the TeX distribution, i.e. longer that the usual biblio.bib where the usual LaTeX manual or The Graphics Companion is 'cited' I just need it for producing more than one page of bibliography

Johannes_B

1:15 PM

@JosephWright Same here. Decision between standard, KOMA and memoir defines at least some package choices.

@ChristianHupfer biblatex-examples.bib But since years are given with dates, classical BibTeX has no years.

@ChristianHupfer Make the page smaller :-p

Christian Hupfer

@Johannes_B Nope

Johannes_B

@DavidCarlisle The question in this comment left me speechless.

David Carlisle

$ grep '\.bib' `kpsewhich -all ls-R` | wc -l
672

@ChristianHupfer which would you like?:-)

Christian Hupfer

@DavidCarlisle Thanks, I was lazy. I've updated my answer to the complaining O.P.

baxx

how to find an error from the error messages?

LaTeX Warning: Unused global option(s):
[19pt].

(./main.aux

LaTeX Warning: Label `subsec:label' multiply defined.

LaTeX Warning: Label `subsec:label' multiply defined.

LaTeX Warning: Label `subsec:label' multiply defined.

! Extra }, or forgotten \endgroup.
l.77 }

?

I can't seem to find that anywhere in the doc and I've commented out everything since it last compiled >.<

David Carlisle

1:21 PM

@baxx first error is 19pt on first line, last is a stray brace on line 77 and between those you have a duplicate label that is line 77 of the aux file so start off by deleting the aux file and regenerating it it may fix itself

baxx

oh it's working now... urgh - ok that means stray on line 77 that's good to know.

think it must have cleaned that aux up or something then because i didn't really do much, thanks

Johannes_B

@baxx Are you creating subsections by clicking something in your editor? Change what is inside the generated \label or don't use the clicky button at all.

baxx

no just typing them in

\subsection{} \subsubsection{} etc

David Carlisle

@baxx line 77 of the aux was just } but scrambled aux files are usually caused by an erro rin a caption of section heading

Will Robertson

@UlrikeFischer I haven't tested such things, but it should be fine to write, for example, \defaultfontfeatures{NFSSEncoding=TheGreatUlrikeEncoding}... of course, at that point, perhaps it's better to write \def\encodingdefault{TheGreatUlrikeEncoding}, but I don't think fontspec would pick up on that; it should, I think!

(BTW, I don't like the keyval NFSSEncoding but, as David said before, the easiest approach was to keep things very bare bones to start out with.)

@JosephWright — sorry for the dumb question, but after Travis fails, what's your workflow for fixing a particular TLG file? Is there a way to "log in" to the virtual machine that Travis is running on and, you know, look at the outputs and poke around?

@DavidCarlisle I think Chris would have kittens if we proposed having a self-modifying encoding

egreg

1:33 PM

@ChristianHupfer Look for beebe: for instance texbook3.bib

Paulo Cereda

@WillRobertson: Am I safe to write my thesis now because of fontspec? :)

Christian Hupfer

@egreg Thanks, I will give it a try

barbara beeton

@ChristianHupfer -- nelson beebe's bibtex bibliography collection has some long ones. the one from "texniques" might be a reasonable size: ftp.math.utah.edu/pub/tex/bib/texnique.html

Christian Hupfer

2:39 PM

@barbarabeeton I've added your link to my list for future usage. Thank you very much.

Joseph Wright

@WillRobertson Travis-CI works on a 'yes/no' basis for tests. If they fail, you have to check 'locally' what's up. That's fine where there is an issue you can reproduce, but where it's down to the difference in system set up it's harder. That's where a VM comes in: I have Ubuntu 12 LTS available for that

@WillRobertson Once the tests are set up most of the time you are fine: I very rarely need to check the L3 ones on my VM

David Carlisle

3:02 PM

@egreg $ svn commit -m "macron on YyGg" utf8ienc.dtx ltoutenc.dtx changes.txt so this works with pdftex

\documentclass{article}

\usepackage[utf8]{inputenc}
\begin{document}

[Ȳ][ȳ][Ḡ][ḡ]

\end{document}

egreg

@DavidCarlisle :)

David Carlisle

@egreg I guess we'll plan for a release closer to TL2017 (or if we need a 2017/01/01 PL2)

egreg

@DavidCarlisle Probably not so urgent an issue

David Carlisle

@egreg no actually more useful is the tulmvtt fd file I added yesterday (eg typesetting the hyperref manual with luatex needs that) but let things settle down for a week or two and see if we need a patch release I hope

Joseph Wright

3:27 PM

@DavidCarlisle I'd favour holding until we have a bigger set of 'stuff' unless there is anything significantly broken

Johannes_B

It seems reading a manual, however small it might be, seems to fix stuff.

CarLaTeX

@egreg I didn't realize the difference because I added a view command after the arara one in TeXstudio: tex.stackexchange.com/q/350888/101651 I haven't had time to learn another editor yet...

David Carlisle

@JosephWright yes agreed possibly aim for a new full release just before tl2017

egreg

3:46 PM

@CarLaTeX Added an answer

CarLaTeX

@egreg Thank you for your detailed answer! Now I've understood :)

Maïeul

4:06 PM

hi, I am looking for older version of MacTeX (I would like to get MacTeX 2015)

where are the archives?

I found here ftp.belnet.be/ftp.tex.ac.uk/ctan/tex-archive/systems/mac/mactex

but not obvious

Maïeul

4:22 PM

Oh

I wanted to write "Later" and and wrote "LaTeX"

…

egreg

@DavidCarlisle %% AMS-LaTeX Created with the Wolfram Language seems a bit pretentious. It should be PseudoLaTeX rubbish code created by incompetent people at Wolfram

@Maïeul There are archived versions of TeX Live (including MacTeX) at ftp.ams.com (anonymous ftp)

Maïeul

@egreg thanks

David Carlisle

4:40 PM

@egreg I don't know, compared to yesterday's funnelweb, it seems quite good.

egreg

@DavidCarlisle I particularly like r{}^{\wedge}2

CarLaTeX

5:40 PM

@Maïeul You are becoming LaTeX addicted!

Stefan Kottwitz

5:54 PM

@CarLaTeX So he joined a huge group.

Johannes_B

6:14 PM

@StefanKottwitz -ingoption? i never use texcount.

CarLaTeX

@StefanKottwitz Of course :):):) ... P.S. = I'm just reading your "LaTeX Beginner's Guide"!

Stefan Kottwitz

@Johannes_B LC? -inc

@CarLaTeX Nice to hear! I like the LaTeX Cookbook more. But I think about writing another one. :-)

CarLaTeX

6:40 PM

@StefanKottwitz I started with the beginner book, after that I'll read something more complex :):):)

Johannes_B

@StefanKottwitz Yes.

Stefan Kottwitz

@CarLaTeX The cookbook is a collection of useful recipes that are simple enough to use with a great result

egreg

6:56 PM

@CarLaTeX You surely know what to read, it's even in Italian. ;-)

CarLaTeX

7:25 PM

@egreg "Appunti di programmazione in LaTeX e TeX"? :)

David Carlisle

8:01 PM

@egreg the colortbl manual passed through google translate?

Joseph Wright

After a couple of false starts, expl3 update ready to go to CTAN

CarLaTeX

@DavidCarlisle :):):):):)

David Carlisle

8:21 PM

@CarLaTeX it's a much better read than your suggestion.

CarLaTeX

@DavidCarlisle I don't know if @egreg agrees... :)

David Carlisle

@CarLaTeX try reading this, I don't think @egreg's that great a writer... translate.googleusercontent.com/…

@CarLaTeX it must be the original italian text at fault as I trust google translate completely

CarLaTeX

@DavidCarlisle Nooo, you can't trust something that translates Henry Gregory!

David Carlisle

@CarLaTeX OOh I didn't spot that, we should call him that from now on.

CarLaTeX

@DavidCarlisle It even sounds good!

egreg

8:36 PM

@CarLaTeX It doesn't fit with my usual abbreviation for “procedura di eliminazione di Gauss”, that is, obviously, EG.

CarLaTeX

@egreg I've read about that procedure on your fan page on FB!

egreg

@CarLaTeX :)

CarLaTeX

@egreg I'm writing another question about arara, get ready to answer!

egreg

@CarLaTeX Ask Professor von Duck, he should know!

5

CarLaTeX

@egreg He's out with his friends at the moment :):):)

Joseph Wright

8:47 PM

@DavidCarlisle L3 update received by CTAN: now we find out who is using deprecated functions!

yo'

@JosephWright not until people tlmgr --update or get new vanilla TL...

Joseph Wright

@yo' When I say 'now' I mean 'in the coming days' not 'this evening' ;)

yo'

@JosephWright :-) I wonder how many of them I actually use

Joseph Wright

@yo' I mailed everyone I could after a sweep over TeX Live

David Carlisle

9:03 PM

@JosephWright usually turns out to be @WillRobertson :-)

Joseph Wright

@DavidCarlisle Or me ...

yo'

the feeling of guilt when you are forced to use libreoffice for a presentation...

David Carlisle

@JosephWright I am free of blame

@yo' you mean that "I should be using PowerPoint" feeling ?

yo'

@DavidCarlisle hey!

Joseph Wright

@DavidCarlisle Keynote, surely

egreg

9:10 PM

@CarLaTeX Not an expert in tikzexternalize, sorry. Surely you need shell escape at the first run, because something might have changed, but for the subsequent runs I don't know whether TikZ is happy with the already present files. Note that you don't need to run frontespizio each time, but just if the frontispiece data has changed.

CarLaTeX

@egreg I know I need to run forntespizio only when I change it

@egreg It's only an example

egreg

@CarLaTeX If you add a short example with externalization, maybe we can test it.

CarLaTeX

@egreg OK, I'll add it

yo'

darn it, I give up. I'll send the guy only a PDF file. If he's not happy, 'tis not my fault.

egreg

@JosephWright Did you publish the list of now removed functions?

Joseph Wright

9:14 PM

@egreg Yes, they are in the LaTeX-L/CTAN text

@egreg l3obsolete as usual covers them

@egreg As always, I did a TeX Live sweep and have held off until the 'core' updates were done (fontspec, unicode-math primarily)

egreg

@JosephWright It seems I have already marked my answers where \c_job_name_tl was used, warning to change it into \c_sys_jobname_str

Joseph Wright

@egreg Indeed: the biggest set of alterations I think were the sys-related ones

CarLaTeX

9:41 PM

@egreg mwe added: tex.stackexchange.com/q/350943/101651, se/quando puoi, grazie, ciao!

David Carlisle

@JosephWright er....

! Package array Error:  Illegal pream-token (\__siunitx_table_collect_begin:NnS
{table-format=7.0}): `c' used.

Joseph Wright

@DavidCarlisle ?

David Carlisle

@JosephWright ah I see OP error:-)

@JosephWright sorry made the mistake of running code from the main site expecting it to do something useful.

@JosephWright no still same, since you are hereL

\documentclass[sigconf]{acmart}
\usepackage{booktabs,makecell,tabularx}
\renewcommand\theadfont{\small}
\usepackage{siunitx}
\usepackage{adjustbox}
\usepackage{array,booktabs}
\usepackage{graphicx}
\usepackage{epstopdf}

%\newcolumntype{C}[1]{>{\centering\arraybackslash}p{#1}}
\newcolumntype{L}[1]{>{\raggedright\arraybackslash}X}

\begin{document}
\begin{table}
    \centering
    \setlength{\tabcolsep}{5pt}
    \small
    \begin{tabularx}{\columnwidth}{@{}c L S[table-format=7.0]*{2}{S[table-format=2.0]}@{}}

Joseph Wright

@DavidCarlisle L is defined taking one argument ...

David Carlisle

L[1] @JosephWright :((( :)

@JosephWright which was what I saw above when said it was OP error then didn't actually change it before trying again. Time to stop I suspect.

Joseph Wright

9:53 PM

@DavidCarlisle :)

CarLaTeX

@samcarter Thank you!

egreg

@CarLaTeX First of all, give up with latin9 and use UTF-8. Always and only.

CarLaTeX

@egreg OK, at once!

CarLaTeX

10:16 PM

@egreg Thank you for the answer! In the meantime I changed latin9 into utf8 in my document but I get an error on ` \EUR{100} ` of eurosym package: ! Package inputenc Error: Keyboard character used is undefined

David Carlisle

10:33 PM

@CarLaTeX which character number did the error say? (and how did you define \EUR)

@CarLaTeX did you re-encode the document to utf8, as well as changing the declaration?

CarLaTeX

@DavidCarlisle \EUR is defined by eurosym package

David Carlisle

@CarLaTeX doesn't really need any external package for euro but I'll look what it does

CarLaTeX

@DavidCarlisle How can I re-encode the document?

egreg

@CarLaTeX It depends on what editor you use.

CarLaTeX

@DavidCarlisle I'm trying to build a mwe that reproduce the error but it works!

David Carlisle

10:36 PM

@CarLaTeX most text editors will save the file in utf8. don't just change the encoding declaration, you need to tell inputenc which encoding the file is saved in. if it is latin9 just telling inputenc that it is utf8 doesn't work.

CarLaTeX

@egreg I'm using TeXstudio

David Carlisle

@CarLaTeX what character do you have before \EUR in the error document

egreg

@CarLaTeX There should be some pop-up menu for reencoding. Unfortunately, TeXStudio refuses to run on my machine.

CarLaTeX

@DavidCarlisle "bond of " with a blank

@egreg I'll search...

egreg

@CarLaTeX If you look at the bottom of the window, you should see a pop-up menu with the encodings.

David Carlisle

10:40 PM

@CarLaTeX shouldn't error on EUR then (but if the specified encoding doesn't match the file encoding then any error is possble:-)

egreg

@CarLaTeX Make a back-up copy, of course, before doing these experiments.

CarLaTeX

@egreg Don't worry I have n backups of my files...

egreg

@CarLaTeX For n → ∞

CarLaTeX

@egreg Hahaha! I've found the pop-up menu with the encodings, I've changed it in utf8 and now the \EUR works, I have other errors but much less

David Carlisle

@CarLaTeX if there are any characters that are supported in [latin9]{inputenc} but not [utf8]{inputenc} can you let me know. That should not happen.

CarLaTeX

10:50 PM

@DavidCarlisle I have two errors at the \printbibliography line, maybe there is something wrong in my bib file...

David Carlisle

@CarLaTeX that would need to be utf8 (or better, ascii with latex markup so \'a etc)

CarLaTeX

@DavidCarlisle My bibliography is in English, hence, in theory, I should not have accented letters...

David Carlisle

@CarLaTeX sensible language

CarLaTeX

@DavidCarlisle Hahahahahahaha

CarLaTeX

11:27 PM

@egreg @DavidCarlisle I don't know what action mended things but 1) I've changed "Università" in my frontespizio into "Universit`{a}"; 2) I've deleted all the temporary files; 3) I deleted all the images created by tikzexternalize so that the compilation re-created them... and now all works perfectly, I can go to sleep calm! Thank you!

egreg

@CarLaTeX If you don't use inputenc, but only \usepackage[T1]{fontenc}, some accented characters print the corresponding glyph, but it's luck. Try straße and you'll get “straSSe”. When you use accented characters in a LaTeX document, always load inputenc with the relevant option.

CarLaTeX

11:48 PM

@egreg I had \usepackage[latin9]{inputenc} and now I have \usepackage[utf8]{inputenc}, as you told me. P.S. = un'assistenza così quelli di Word se la sognano! Good night!

Transcript for

$TeX - LaTeX$ TeX, LaTeX and Friends