TeX, LaTeX and Friends

6:51 AM

I was using the beamer theme 'Frankfurt' on Overleaf (Texlive 2020). On the header navigation line, there appear white circles for each section. How can I remove or hide those circles?

David Carlisle

8:12 AM

@UlrikeFischer why can't it use \year\month\day or even \filemoddate?

Ulrike Fischer

@DavidCarlisle I have no idea. It looks very odd. There is even a specific document option to avoid the error message if shell-escape is not there (and which then use a fix date D:20181028075445).

David Carlisle

@UlrikeFischer I suppose we could suggest they always use \filemoddate{\jobname.aux} ...

samcarter_looks_forward_TUG'21

@raf You could redefine the headline to use \insertsectionnavigation

Ulrike Fischer

@DavidCarlisle we should have some direct support: github.com/latex3/latex3/issues/699. But yes, anything is better than forcing shell-escape for a simple date. I wonder why no one ever complained about this, the package seems not to have many xelatex users.

touhami

8:49 AM

Hello TeX.SE, hello @DavidCarlisle

I need some help. I want to ( run make4ht -f html5+mjcli file.tex "mathjax" ) but i get mjcli not defined

David Carlisle

@touhami so do I what do you want it to do, its not an option I recognise but I don't use make4ht much

touhami

@DavidCarlisle thank you, but i didn't understand what you mean :-) sorry can you please rephrase :-)

David Carlisle

9:07 AM

@touhami I mean, I get that error if I try it but I have no idea what you expect that option to do, is it documented anywhere? Also I don't use make4ht regularly so I'm probbaly not the right person to ping...

touhami

@DavidCarlisle OK. Thank you, it's in make4ht documentation, section (7.6 The mjcli filter and extension)

David Carlisle

Mar 26 '12 at 19:37, by David Carlisle

@Canageek moral of the story: never read the documentation, bad things happen

@touhami but it is presumably this: github.com/michal-h21/mjcli

touhami

@DavidCarlisle Thank you again. I'll tray and back.

Joseph Wright

@UlrikeFischer Well we are going to fix it anyway ...

@PauloCereda I'm working on my TUG2021 slides: the title and picture are in

@DavidCarlisle Looking forward to 'well, it's all in color anyway`? ')

David Carlisle

@JosephWright expecting hostile comments from the floor?

Joseph Wright

9:17 AM

@DavidCarlisle More 'insightful'

Latest siunitx made it to CTAN

David Carlisle

@JosephWright I'll do an interview with Laura Kuenssberg

Joseph Wright

@egreg I see you are on the network hot list in another place ;)

@DavidCarlisle :)

@egreg I was just reading the same question and was thinking of more or less the same answer :) (One of the few things I know from my time in Italy as an infant)

egreg

@JosephWright Pronto!

Joseph Wright

@egreg Sono io

@egreg :)

egreg

@JosephWright So no Italian? Where have you lived?

Joseph Wright

9:29 AM

@egreg My dad worked in Novara when I was very small: my mum and I were only there for about 7 months of the two years he had that job (two summers, so I was ~18 months and ~2.5 years)

@egreg Dad's job was at an 'international' factory, so he had only a few phrases in Italian, and with a little one my mum didn't get out other than to buy milk

egreg

@JosephWright I'm not sure of Novara, but south of it is the paradise of mosquitos, due to rice cultivation.

Joseph Wright

@egreg Er, yes: also lots of fog in the winter (mum and I were only summer visitors because it was super-depressing the rest of the year, according to my dad)

@egreg Po valley is ... not so exciting

egreg

@JosephWright Fog in the Po plains? That sounds strange! ;-) You don't know fog if you haven't been in the Po delta in wintertime.

Faheem Mitha

@UlrichDiez People presumably want to typeset with TeX. Which is of course a broad area. But knowing something about TeX is likely to be helpful no matter what you are doing, at any rate in the long term. That's all I was trying to say.

As I already said, this is colored by my personal experiences. I.e. 25 years of using TeX without really understanding what it was doing.

@UlrichDiez Yes, that's part of how the world works under international capitalism. The system is well understood but it still continues, and things continue to get worse.

@UlrichDiez I don't know anything about that. Do you have a special interest in this? Are you in the medical profession?

Rmano

10:01 AM

@egreg Torino does not qualify? I remember pretty scary street crossings where you had to look for the green light on your side of the street, the other one was invisible. And make you thing about the car's point-of-view (you know, very slow driver over there...)

egreg

@Rmano My school was on the other side of the road; once I had no clue at all where to cross the road: no traffic lights and no zebras, being a very small town. I couldn't see the school building!

Paulo Cereda

10:31 AM

@JosephWright ooh

@JosephWright Sono inglese di Norwich? :)

Faheem Mitha

10:55 AM

@UlrichDiez Just to return to this comment, you wrote:

> ^^J is tokenized as explicit character-token of character code 10 and category code 12(other). Like any character-tokens of category code 11 or 12, this character-token is taken for a directive for insertion of combinations of glyphs into the output-file/.pdf-files. (Which glyphs you get beneath other things depends on the font-encoding and the font in use.)

> Besides this character-token begins a new table-cell. The subsequent \hline cannot be carried out inside a table-cell. This circumstance causes TeX to complain.

I think I follow that. But for LuaTeX, what character would actually be typeset? The answer mentions:

> Instead the new line character is a normal character, the slot position in cmr10 is Ω.

But isn't that a different font than what LuaTeX would be using?

Paulo Cereda

3

@egreg ^^ :)

Sebastiano

Good morning everybody from the East Sicily...

samcarter_looks_forward_TUG'21

4

@PauloCereda ^^^ The bread book?

Rmano

11:14 AM

@PauloCereda we really have "compagno" and "companatico" for both acceptions...

So you can share bread with a duck (I know, you shouldn't) or... ;-P

Paulo Cereda

@samcarter_looks_forward_TUG'21 ooh

@Rmano ooh

egreg

11:55 AM

@PauloCereda You certainly have bread with ducks.

@PauloCereda You know, duck sauce is very good.

Paulo Cereda

@egreg oh no

Faheem Mitha

12:45 PM

@samcarter_looks_forward_TUG'21 That's an old edition, isn't it?

Joseph Wright

@FaheemMitha TLC1

Faheem Mitha

@JosephWright Pardon? Do you mean first edition?

Joseph Wright

@FaheemMitha Yes

Paulo Cereda

@JosephWright Will TLC3 be the hummingbird book?

Joseph Wright

@PauloCereda I've not seen the cover yet

@PauloCereda Ask FMi

Paulo Cereda

12:58 PM

@JosephWright nobody saw it!

@JosephWright no I am scared

samcarter_looks_forward_TUG'21

May 26 '17 at 9:19, by David Carlisle

@PauloCereda I always think of the Companion as the dog-book and forget they changed it to a compass:-)

Faheem Mitha

@samcarter_looks_forward_TUG'21 Shouldn't it be a Lion book?

Are there only two editions of the LaTeX Companion?

Joseph Wright

@PauloCereda Ask him: you never know your luck

@FaheemMitha TLC3 is being written ATM

Faheem Mitha

@PauloCereda Why are you scared?

Paulo Cereda

@samcarter_looks_forward_TUG'21 woof

Ulrike Fischer

1:01 PM

@PauloCereda ask him, perhaps you can help ;-)

Paulo Cereda

@FaheemMitha just the typical irrational fear of TeX-related stuff :)

Faheem Mitha

@PauloCereda Huh?

Paulo Cereda

@UlrikeFischer ooh then I can tell him I am about to break a stuff he wrote in the book even before the book is out :)

@FaheemMitha it's just for fun, do not worry

samcarter_looks_forward_TUG'21

@PauloCereda meow

Paulo Cereda

@samcarter_looks_forward_TUG'21 ooh

Faheem Mitha

1:09 PM

The second edition is from 2004!

Rmano

1:25 PM

Well, there will be Lion Book, it seems... kickstarter.com/projects/preppylion/…

Paulo Cereda

@UlrikeFischer if the cover reads Teh LaTeX companion we know who designed it :)

Joseph Wright

@PauloCereda I think they get a real pro ....

Paulo Cereda

@DavidCarlisle Joseph said you are not a real pro ^^

:)

samcarter_looks_forward_TUG'21

7

^^^ the ultra secret draft of the cover

Faheem Mitha

@Rmano Ooh, sexy.

Skillmon

1:39 PM

@FaheemMitha counter-ooh?

@PauloCereda did I do that right? ^^^

Faheem Mitha

@Skillmon Paulo doesn't have a monopoly.

Ulrich Diez

@FaheemMitha I am totally with you on teaching people a bit of how TeX works. The question is how far to go: I think in many cases basics are enough:

Faheem Mitha

@UlrichDiez Depending on how one defines "basics".

Ulrich Diez

Knuth's analogy to an organism that looks at a .tex input line by line with its eyes, and during pre-processing (transcoding to ASCII/Unicode, removal of trailing paces, appending endlinechar) takes the line as a directive to put characters in its mouth and to chew those characters into tokens, and to swallow the tokens inton the gullet.

In doing so, drawing attention to the difference between "character" and "character token". In the context of tokenizing, also addressing category codes, and in what ways tokens can be classified ((By state: macro/primitive/undefined/\chardef-token/implici…

@FaheemMitha Fonts: You are right: Traditional TeX enguines by default use Computer Modern Fonts. When tunning pdflatex on \documentclass{article}\tracinglostchars=2 \begin{document}\mbox{^^J} \texttt{^^J}\end{document} the cosole output reveals that cmr10 and 10tt are in use.

Faheem Mitha

@UlrichDiez Is this a quote from somewhere?

Ulrich Diez

1:44 PM

In the .pdf I get omega (Ω)-symbols as CMR-fonts have omega in slot 10. Iirc until 2016 xelatex and lualatex also defaulted to Computer Modern, but nowadays these engines default to Latin Modern Fonts. In case of T1-encoding LaTeX switches to cm-super.

Running LuaLaTeX or XeLaTeX on the same yields that lmmono10-regular.otf and lmroman10-regular.otf are in use. Additionally I get Missing character: There is no (U+000A) in font [lmroman10-regular]:+tlig;! and Missing character: There is no (U+000A) in font [lmmono10-regular]:! because unlike CMR these fonts don't have a character in slot 10(dec)=000A(hex).

Faheem Mitha

@UlrichDiez Ah. So there is no glyph corresponding to that character? Dammit, I forgot to look at the log.

Ulrich Diez

@FaheemMitha In my normal professional life, I am not active in the medical field. But here in Germany I am subject to compulsory military service and am assigned to the medical corps. Thus I'm interested in the stories and history of colleagues around the world. :-)

@FaheemMitha The .log reveals this only if you do \tracinglostchars=2 or the like.

Skillmon

@UlrichDiez uuh, new features :)

David Carlisle

@UlrichDiez or wait until the next latex release:-)

Phelype Oleinik

@Skillmon Yeah, bleeding-edge 1998 technology ;-)

Ulrich Diez

1:50 PM

@FaheemMitha That analogy is in the TeXbook. E.g., in chapter 7: How TeX reads what you type.

@Skillmon ????? \tracinglostchars is described in the TeXbook, Chapter 24 Summary of Vertical Mode and Chapter 27 Recovery from Errors. What new features are you referring to?

David Carlisle

@UlrichDiez tracinglostchars=1 is TeXBook, tracinglostchars=2 is e-tex so "new-ish" tracinglostchars=3 (error) is very new (texlive 2022 sources)

Ulrike Fischer

@DavidCarlisle tracinglostchars=4 is then with sound and color ;-)

David Carlisle

@UlrikeFischer that calls mktexotf to call fontforge and designs an extended font with the character added on the fly

2

Ulrich Diez

@DavidCarlisle I didn't realize this yet. Thank you. The TeXBook just says that the value of \tracinglostchar needs to positive for causing TeX to record each time a character has been dropped as it does not appear in the current font.

Ulrike Fischer

@DavidCarlisle @MarcelKrüger's next project?

Faheem Mitha

2:03 PM

@UlrichDiez Oh.

David Carlisle

@UlrikeFischer he's had over a minute by now, it's probably done.

Faheem Mitha

@UlrichDiez LuaTex doesn't return an error, though.

Ulrich Diez

@UlrikeFischer And what will be "with sound and color and bright"? :-)

Skillmon

@UlrichDiez what @DavidCarlisle said :)

David Carlisle

@samcarter_looks_forward_TUG'21 looks like a good book, I'd buy that.

Faheem Mitha

2:06 PM

@UlrichDiez I see it:

> Missing character: There is no (U+000A) in font TeXGyreSchola:mode=node;script=latn;language=dflt;+tlig;!
{/var/lib/texmf/fonts/map/pdftex/updmap/pdftex.map}] (./char.aux))

Wow, that's a mouthful.

Ulrich Diez

@FaheemMitha Yes. I don't get an error. It's a note on the console and in the .log. Otherwise one would get plenty of error-messages when applying \nullfont.:-) Afaik TikZ does....

Faheem Mitha

Actually, I see the same error (or warning) message without \tracinglostchars=2.

So apparently it's not necessary after all.

David Carlisle

@FaheemMitha the second line is unrelated just logging file input at the same place so it's only one line and it couldn't use a lot less to tell you which character you tried and which font did not have it.

Faheem Mitha

I suppose dflt is default. Not sure what tlig is, though.

Skillmon

@FaheemMitha wait, dflt is no double precision float?

David Carlisle

2:10 PM

@FaheemMitha latex sets it to 1 so you get some messag ein teh log (but not the terminal) by default

Faheem Mitha

Ah, perhaps tiling?

@Skillmon Language = double precision float?

Perhaps so. I don't know.

@DavidCarlisle OK, I see.

Ulrich Diez

@FaheemMitha With XeLaTeX I don't get the phrase "(U-000A)" after the linebreak. I just get the linebreak. Copy-pasting console-output to hex-editor seems to show that a carriage-return, not a linefeed is written. :-) Probably due to 10, denoting to linefeed-character, being the value of \newlinechar....

David Carlisle

@FaheemMitha dflt means default

Skillmon

@DavidCarlisle :(

David Carlisle

@FaheemMitha tlig is tex-ligaures (so --- = em dash etc)

Faheem Mitha

2:13 PM

@UlrichDiez So does it actually produce a newline then? It's hard to tell. And if it does, then why does it complain about a missing glyph?

@DavidCarlisle tex-ligatures?

Does TeX only look in the font glyphs when it thinks that something is a printable character?

David Carlisle

@FaheemMitha as I sais classic tex ligatures that are not normally in a standard open type font

Faheem Mitha

Because of course not everything is a printable character.

@DavidCarlisle I don't actually know what that is. Ligatures are those things that tie letters together, right?

Like ff or th.

David Carlisle

@FaheemMitha Ulrich mean that the U+00A0 was not printed to the log not that it wasn't typeset to the pdf. typesetting to the PDF causes a missing character warning message as putting a character 10 in a message causes a linebreak in teh mesage as \newlinechar = 10

@FaheemMitha well like ff th as a ligature is more rare, but the tex ones are the bhhhones that are only in tex fomts, as I say like --- being em dash or ' ' being ”

Faheem Mitha

@DavidCarlisle OK.

@DavidCarlisle I'm having difficulty parsing that. What message are we talking about here?

David Carlisle

@FaheemMitha the last comment from Ulrich

Ulrich Diez

2:18 PM

@FaheemMitha I am talking about the message on the screen not the pdf-file:

Seems the message is intended to contain the linefeed-character which is missing in the font. Thus a linefeed-character is included into that message.

When it comes to writing that message to screen, the linefeed-character (erroneously?) is replaced with a carriage-return-character by the writing-routine because the value of \newlinechar is ten while the writing-routine replaces those characters by carriage-return whose number equals the value of \newlinechar. :-)

Faheem Mitha

@UlrichDiez Oh, I didn't see the message on screen. Is that with \tracinglostchars=1 or 2?

The message on the screen and in the log look the same to me.

Ulrich Diez

@FaheemMitha I did \tracinglostchars=2. And this is with XeLaTeX. With LuaLaTeX I didn't check what character is actually written to the console as LuaLaTeX also tells you the number - the (U+000A)-thingie.

Faheem Mitha

They both say:

> Missing character: There is no
(U+000A) in font TeXGyreSchola:mode=node;script=latn;language=dflt;+tlig;!
{/var/lib/texmf/fonts/map/pdftex/updmap/pdftex.map}] (./char.aux))

@UlrichDiez Oh, you're talking about XeLaTeX? I didn't try that. One sec.

> Missing character: There is no
in font TeX Gyre Schola/OT:script=latn;language=dflt;mapping=tex-text;!
[1

That's the XeLaTeX log message. Again, the screen and log messages are the same, except there are a couple of linebreaks after the 1 in the log message.

Is that a bug, then?

David Carlisle

@FaheemMitha so Ulrich's point is that the character after "There is no" is actually 13 not 10 the fact that the message after the linebreak is dropped looks like a bug in xetex

Ulrich Diez

@FaheemMitha Yes. With the message there is a linebreak behind the word "no". If not the linefeed-character was missing, but the character A was missing, then there would not be a linebreak. Instead A would be displayed. Because the intention is to display the missing character in question. Seems the linebreak comes into being because TeX's routine for writing things to screen replaces the linefeed by carriage-return due to \newlinechar having the value 10. :-)

Faheem Mitha

2:25 PM

@DavidCarlisle How does he know it's 13?

Ulrich Diez

@FaheemMitha I did copy-paste of the console-output to a hex-editor...

Faheem Mitha

@UlrichDiez Wow.

David Carlisle

@FaheemMitha well he did say that how he looked a couple of minutes ago:-)

Faheem Mitha

So will someone report it, then? And LuaTeX's output is correct?

Ulrich Diez

@DavidCarlisle If my observation is correct - is it a bug? To me it seems the \newlinechar-mechanism works correctly. :-)

David Carlisle

2:27 PM

@UlrichDiez I think the interaction with \newlinechar is expected, the message after the linebreak being dropped is a bug though.

Faheem Mitha

@DavidCarlisle I must have missed that.

David Carlisle

@UlrichDiez yes \newlinechar is fine but th emessage vanishing is less fine I think.

Faheem Mitha

So XeLaTeX should also report (U+000A) then?

Ulrich Diez

@FaheemMitha Before reporting anything the matter needs examination. :-) I might have made a mistake somewhere...

Faheem Mitha

Or not?

So, a final question about font glyphs before I go for dinner. Does TeX simply look up 10 in a font table or similar to see if there is a glyph that matches 10? Or is it more complicated?

David Carlisle

2:29 PM

@FaheemMitha @UlrichDiez oh no my recollection of history is fault. It's all as expected. The (U+000A) message is a luatex thing anyway.

Faheem Mitha

@DavidCarlisle Not a very helpful error message, regardless.

David Carlisle

@FaheemMitha why is it not helpful. Admitedly it is not optimised for the case of someone typing a linebreak but I don't think that is a problem since it never happens anyway and what would you do instead?

@FaheemMitha luatex at the last release made it more helpful by adding the U+000A part so it is always visible even when the character that is missing is not printable

Ulrich Diez

@DavidCarlisle The difference between LuaLaTeX and XeLaTeX is that the further within the message also delivers the number of the missing character - the "(U-000A)"-thingie. But that's not the point I am observing: Both with XeLaTeX and LuaLaTeX the message seems to be intended to also display the missing character. As the missing character is a linefeed, a linefeed should be displayed. Seems due to \newlinechar this linefeed is replaced by carriage-return. I don't know if this is a bug.

David Carlisle

@UlrichDiez I think it's a feature (there is no easy way to mention a raw linebreak character whether 10 or 13 in the log)

Ulrich Diez

Displaying non-printable characters on console is a funny business in any case.

Paulo Cereda

2:34 PM

@Skillmon ooh <3

David Carlisle

@UlrichDiez exactly

Paulo Cereda

@FaheemMitha oh no

@Skillmon er... counter-uuh? :)

@samcarter_looks_forward_TUG'21 ooh <3

Ulrich Diez

@DavidCarlisle I think it might be a bug: The newlinechar-mechanism is applied with printable characters also:

\documentclass{article}\tracinglostchars=2 \begin{document}\nullfont\mbox{ABCDEFGHIJ}\newlinechar=`\D\mbox{ABCDEFGHIJ}\end{document}

In the second case in the message D is replaced by a linebreak.

Skillmon

2:50 PM

@DavidCarlisle could use ^^ notation.

@PauloCereda uuh, new game!

Joseph Wright

@UlrikeFischer Hopefully I can look over the xform stuff shortly: should I be looking to move it to the core l3backend?

Ulrich Diez

@JosephWright Probably you replied to the wrong person - I have no idea what xform stuff is about... :-)

Skillmon

@UlrichDiez I guess that should've been an @UlrikeFischer

All those weird German names!

Joseph Wright

@UlrichDiez Mumbles

Ulrike Fischer

@JosephWright I'm just integrating this into the pdfresources and then I guess I will make an update. We should at some time start to move the extra backend code, but this parts depend on the management code, so it will need to be looked over if one can move it without danger.

David Carlisle

3:15 PM

@UlrichDiez yes that's what I expect. If you say character zzz should print as a newline, the it prints as a newline.

Ulrich Diez

@Skillmon :-))) Let's be glad that my name is not Johann Gambolputty....

@DavidCarlisle Why is the newlinechar-mechanism applied to a message not written due to a \write-command?

David Carlisle

@UlrichDiez well it has to do everything else mapping teh internal character encoding to the system encoding so it's not totally unreasonable that it does this as well.

Skillmon

@UlrichDiez I read somewhere that there is a 33 character limit on the names on stackexchange...

Ulrich Diez

@Skillmon So we should be pretty safe.... :-)

Joseph Wright

@UlrikeFischer Right yes

Faheem Mitha

3:29 PM

@DavidCarlisle Because it shows a blank where the reader would naturally expect something else? Or apparently a line break, but most will not know the difference.

@DavidCarlisle That's clearly much better.

samcarter_looks_forward_TUG'21

@Skillmon 30 (including spaces etc)

Faheem Mitha

@UlrichDiez Isn't "(U-000A)" a standin for the missing character? Where does it look like it's trying to display the missing character?

David Carlisle

@FaheemMitha yes but this is not much different than if you type chinese using latin modern and get told that they are missing characters, if you use an ascii based font to view the log the message is not really understandable as the character is missing from the mesage saying that it is missing. It still basically has the information you need, somewhere.

@FaheemMitha it follows the character. If it was trying to say x was missing the form would be x (U+0078) it is just that if x is not a printable character the message output is a bit arbitrary

@FaheemMitha for some people and some uses it is better. Probably the vast majority of people just find it confusing, and would not recognise U+000A as line feed, or U+0078 as x.

Ulrich Diez

@FaheemMitha E.g., \documentclass{article}\tracinglostchars=2\begin{document}\nullfont\mbox{A}\end{document} with LaTeX and XeLaTeX yields: Missing character: There is no A in font nullfont!, with LuaLaTeX yields Missing character: There is no A (U+0041) in font nullfont! In any case the missing character is displayed unless the \newlinechar-mechanism applies to it. (You can have fun displaying non-printable characters.)

Faheem Mitha

@UlrichDiez OK. Hang on, let me try that code.

@DavidCarlisle Well, I prefer to see something rather than see nothing. Even if the something is hard to understand. But that's just me.

@UlrichDiez That works for me. I didn't know \nullfont was a thing. That could be handy for testing.

David Carlisle

3:44 PM

@UlrichDiez @FaheemMitha I get the Missing character message in xetex, texlive 2021:

\documentclass{article}
\tracinglostchars=2
\begin{document}
\nullfont
\mbox{A}
\mbox{\char10}
\newlinechar=`\X
\mbox{X}
\end{document}

pdftex
Missing character: There is no A in font nullfont!
Missing character: There is no
 in font nullfont!
Missing character: There is no
 in font nullfont!


luatex
Missing character: There is no A (U+0041) in font nullfont!
Missing character: There is no
 (U+000A) in font nullfont!
Missing character: There is no
 (U+0058) in font nullfont!


xetex
Missing character: There is no A ("41) in font nullfont!

@FaheemMitha see the three forms above ^^ but even in the pdftex case you do "see something" it is just that the character being printed is shown as a linebreak (as it is specified to do that by the document)

Faheem Mitha

@DavidCarlisle I guess ("A) is the hexadecimal number corresponding to A?

And similarly for ("58). Though I'm not sure how the same syntax would work for both.

Since 58 is decimal, and A clearly isn't.

David Carlisle

@FaheemMitha yes that is used pervasively in tex. as in \mathcode"8000 etc.

Ulrich Diez

@DavidCarlisle The character being printed as a linebreak is due to the \newlinechar-mechanism.

David Carlisle

@UlrichDiez yes that;s what I meant by " (as it is specified to do that by the document)"

@FaheemMitha ? "58 is decimal 88

Ulrich Diez

@FaheemMitha "A is a hexadecimal number. The hexadecimal digit A denotes the number ten which is the number of the linefeed-character both in ASCII and in Unicode.

David Carlisle

3:50 PM

@FaheemMitha as I said a few minutes ago not everyone would read U+000A as character 10, linefeed

Paulo Cereda

@DavidCarlisle ooh

Ulrich Diez

@FaheemMitha "58 = 58(hex) = [ 5*(16)^1 + 8*(16)^0 ] (dec) = 88(dec)

David Carlisle

@PauloCereda "6F "6F "68

Paulo Cereda

@DavidCarlisle ooh

Ulrich Diez

@FaheemMitha \nullfont is a font without glyphs. E.g., TikZ uses it to ensure that typing characters within tikzpicture-environments yields visible output only in specific situations. But you can still mess around with control sequences for changing the font, for placing boxes and inserting glue...

David Carlisle

3:58 PM

@PauloCereda 10986

@PauloCereda can you make 11111 before tea time?

Paulo Cereda

@DavidCarlisle ooh a challenge

Faheem Mitha

@DavidCarlisle @UlrichDiez Sorry for the nonsense I was talking. Yes, I see that those are all hexadecimal numbers.

@UlrichDiez I see.

David Carlisle

@UlrichDiez or you could say "tikz uses it to try to hide bugs" :-)

Faheem Mitha

@DavidCarlisle So this is (to be clear) a new feature not present in the version that I was running?

David Carlisle

@FaheemMitha I just ran a standard texlive 2021 version

Faheem Mitha

4:09 PM

@DavidCarlisle I get

> Missing character: There is no A in font nullfont!
Missing character: There is no
in font nullfont!
Missing character: There is no
in font nullfont!

So clearly there is been an update.

xelatex --version
XeTeX 3.14159265-2.6-0.999992 (TeX Live 2020/Debian)
kpathsea version 6.3.2
Copyright 2020 SIL International, Jonathan Kew and Khaled Hosny.
There is NO warranty.  Redistribution of this software is
covered by the terms of both the XeTeX copyright and
the Lesser GNU General Public License.
For more information about these matters, see the file
named COPYING and the XeTeX source.
Primary author of XeTeX: Jonathan Kew.
Compiled with ICU version 63.1; using 63.1
Compiled with zlib version 1.2.11; using 1.2.11

Apparently now out of date.

David Carlisle

@FaheemMitha you are running texlive2020 so yes texlive 2021 has updates.

Faheem Mitha

@DavidCarlisle OK. Didn't quite register 2021 was out.

David Carlisle

@FaheemMitha it's been available as test since about Feb and was general release since April I think. If you try to use tlmgr on your system it presumably tells you that your local repository is too old and you can not update

Faheem Mitha

@DavidCarlisle Still using Debian binary packages. Sorry. :-(

I don't often use emoticons. Just to indicate that I'm sorry.

David Carlisle

@FaheemMitha ah in that case 2020 might be best available (you are lucky it's not still 2015:-)

Faheem Mitha

4:14 PM

@DavidCarlisle You give too little credit to Norbert.

But yes, 2021 is not available yet.

So just to go back to the original ^^J for a moment. Since LuaTeX can't find a glyph for a document containing just ^^J, does that mean the document would be blank?

David Carlisle

@FaheemMitha someone was asking on site yesterday with a debian tl2015, if you choose one of the stable LTS releases you get a stable system that doesn't update.....

@FaheemMitha if that was all it is you probably get no pdf at all if you have a page number or something else that allows a page to be shipped out then you get that. basically it's as if it wasn't there.

Faheem Mitha

@DavidCarlisle Ah yes. I backported from unstable. Which is easy, apart from Biber.

I do run stable. But stable has 2018, not 2015.

But 2021 is not currently available in any version of Debian.

@DavidCarlisle I guess there's a page number by default, since it's LaTeX.

@UlrichDiez Do you have a medical background, or does the German military just randomly assign people to the medical corps? :-)

Ulrich Diez

4:32 PM

@DavidCarlisle Reconsidering things I come to the conclusion that this is not a bug:

TeXbook, Chapter 24: Summary of Vertical Mode says about \newlinechar:

"\newlinechar (character that starts a new output line)"

In the TeXbook I did not find any statement clarifying/restricting how the output where the new output line belongs to is to come into being.
So the \newlinechar-mechanism is not restricted to output due to \write but also applies in situations where output to file/console comes into being in other ways, e.g., due to TeX writing whatsoever things to .log-file or console.

barbara beeton

5:36 PM

@PauloCereda c-- But duck sauce is good with egg rolls, and not a duck in sight!

Paulo Cereda

@barbarabeeton ooh

barbara beeton

5:47 PM

@UlrichDiez -- This is getting closer to "with sound and fury".

Marcel Krüger

6:10 PM

@DavidCarlisle That sounds more like \tracinglostchars=-1: Not just nor warning about it, but actively hiding the missing character... The more realistic option would probably be to use Noto's variable font to automatically instantiate a font instance close to the current font and using that as a fallback, kind of like Acrobat Reader uses MM fonts to replace missing fonts. In practice this gets complicated though since the numeric weight/width/etc. values mean different things in different fonts.

Ulrich Diez

@FaheemMitha Medical background? Not really. :-) At the time when I was drafted into the basic military service exactly one day after receiving my high school diploma, many new positions had to be filled with combat nurses.

This job seemed far more appealing to me than to enjoy myself, e.g., in some home-guard-infantry-battalion, training how to shoot at tanks with pistols. So in the interviews, I drew focus on my experiences with the youth departments of the volunteer fire fighters and the Red Cross. That's all. :-)

Faheem Mitha

@UlrichDiez Ooh, sneaky.

Well, as long as you enjoy medical stuff, I guess it worked out.

Certainly better than training how to kill people, IMO.

"youth departments of the volunteer fire fighters and the Red Cross"? Impressive initiative.

Marcel Krüger

@FaheemMitha @DavidCarlisle In LuaTeX that message could be replaced by a version which uses the fontspec name instead of the internal name if we want. That might be more user friendly in some cases. Personally I prefer the current message, but then again I often write .fd files for my OpenType fonts instead of using fontspec so I'm not really a typical user.

Faheem Mitha

@MarcelKrüger You have your own OpenType fonts?

Marcel Krüger

@FaheemMitha I was referring to fonts I use, not fonts I created. (I sometimes do that too to test stuff, but if you ever saw glyphs "designed" by me you will know that noone wants to use them in real documents(myself included))

Faheem Mitha

6:22 PM

@MarcelKrüger Oh. But if you didn't create them, why do you need to write .fd files for them (whatever those are)?

Ulrich Diez

@barbarabeeton Fury? I like fur better. For example, I once was allowed to pet an anteater in the zoo. That was very cuddly. :-)

Paulo Cereda

@UlrichDiez ooh anteaters are awesome

Ulrich Diez

@PauloCereda Yes. Anteaters, aardvarks, tapirs. And ducklings. :-)

Paulo Cereda

@UlrichDiez ooh :)

David Carlisle

@FaheemMitha fd files are tex files describing fonts not fonts. latex uses them for all font descriptions since the latex2e font system was introduced pdftex almost always access fonts via an fd file lualatex less often as most people use fontspec

Marcel Krüger

6:29 PM

@FaheemMitha Most font packages either don't support LuaLaTeX at all or just call fontspec, so if you don't want to use fontspec (mostly for speed and control reasons) you have to write them yourself.

Faheem Mitha

@DavidCarlisle "describing fonts not fonts"?

Paulo Cereda

@FaheemMitha font metrics?

@Rmano ciao signore squirrel! :)

Faheem Mitha

@DavidCarlisle @MarcelKrüger So not all fonts have .fd files for LuaTex, so you need to DIY. Got it.

@DavidCarlisle Sorry, parsed that sentence poorly.

Marcel Krüger

@FaheemMitha They describe font families. So for example which font to use for bold/italic/bolditalic/etc parts, which optical sizes to use, etc.

Ulrike Fischer

@MarcelKrüger how much does it make it faster?

Faheem Mitha

6:34 PM

@MarcelKrüger I see. I'm not familiar with fontspec, so comparisons with it mean nothing to me.

Ulrich Diez

@FaheemMitha You say it. :-))

Rmano

@PauloCereda ciao signor papero!

Paulo Cereda

@Rmano <3

Faheem Mitha

Duck in Italian is papero? Weird.

David Carlisle

@FaheemMitha I thought you were using luatex these days? (or xetex)

Faheem Mitha

6:38 PM

@DavidCarlisle I am using LuaTeX. But I don't know anything about fontspec.

Except you have to add it to your file, otherwise things don't work.

Marcel Krüger

@UlrikeFischer I don't have current numbers on that, I can do some measurements later. On my old system (which had rather slow font lookups due to a slow disk and many fonts) it could save 10% on short files with many fonts.

Rmano

@FaheemMitha yep. I think "anatra" is more correct, but I think most people use "papero". Oh, well, if you are speaking of food is "anatra all'arancia", thinking well...

Schubladenzieher

'Ente' sounds good:D

touhami

6:54 PM

@DavidCarlisle now i get : cannot find module 'mathjax-full/js/input/tex/AllPackages.js'
do i need to install mathjax?

Ulrich Diez

@FaheemMitha Don't get too martial ideas about it. :-)) On the other hand, back then some of the village festivals have pushed me to my limits. :-)

Faheem Mitha

@UlrichDiez You mean you didn't spend your time rescuing beautiful maidens from burning buildings?

Ulrich Diez

7:16 PM

@FaheemMitha No. Approaching fire-breathing creatures that can fly by themselves is dangerous...

Faheem Mitha

@UlrichDiez Hmm.

Ulrich Diez

@FaheemMitha I know it. I have many sisters. :-)

Paulo Cereda

@Schubladenzieher quack

Faheem Mitha

@UlrichDiez LOL

Paulo Cereda

Martha Jones is back!

@JosephWright ^^

Joseph Wright

7:25 PM

@PauloCereda ???

Paulo Cereda

@JosephWright the companion no ones remembers :)

David Carlisle

8:08 PM

@touhami you don't need to install any mathjax related code for the standard makefoht mathjax option, but I don't know what the options you were using are for, i would guess though from mjcli name that is for using mathjax to convert to htm+css or svg on your machine and serve the result rather than the usual way of serving tex or mathml and having mathjax javascript convert in the browser, so in that case you would need a local mathjax

@FaheemMitha then you are presumably using fontspec to change fonts (\setmainfont and friends)

touhami

@DavidCarlisle thank you again. So i need to look for a guide in how to install mathjax i local.

Werner

9:59 PM

@DavidCarlisle: Here's a weird interaction between cellspace, siunitx and tabularx:

\documentclass{article}

\usepackage{siunitx}
\usepackage{cellspace}
\usepackage{tabularx}

\begin{document}

\noindent
\begin{tabularx}{\linewidth}{c X}
  one & two
\end{tabularx}

\end{document}

A different load order for the packages does solve the problem, but there are some load orders different to the above that still fail.

To phrase the question differently: Whaddup whiddat?

Is this something that requires your or someone else's attention?

David Carlisle

@Werner let me look although presumably it isn't tabularx at fault...

@Werner (@JosephWright) changing to \usepackage{siunitx}[=v2] it works again with a warning...

Werner

@DavidCarlisle It never is tabularx's fault. Never.

@DavidCarlisle What magic is this? \usepackage{siunitx}[=v2]?

David Carlisle

@Werner it's the standard latex rollback argument (if a package chooses to implement it) so it gives you siunitx version 2 in this case

Werner

@DavidCarlisle M'kay.

David Carlisle

10:15 PM

@Werner found one error in cellspace that causes the warning about endpbox, but not tracked down why X is an illegal tokem

Werner

@DavidCarlisle Yeah.

David Carlisle

@Werner been a long day, probably easier to ask @JosephWright what changed at siunitx v3 (apart from "everything")

Werner

@DavidCarlisle :)

David Carlisle

@Werner look who arrived:-)

Werner

He's been tagged.

David Carlisle

10:22 PM

@Werner s/tagged/blamed/ ?

Werner

Right on time.

Joseph Wright

@DavidCarlisle Oh, I see

@DavidCarlisle, @Werner github.com/josephwright/siunitx/issues/512

David Carlisle

29 mins ago, by David Carlisle

@Werner let me look although presumably it isn't tabularx at fault...

@JosephWright actually I was concentrating on looking on cellspace before thinking as it's different with v2 I possibly ought to look there (or ping you:-)

Werner

@DavidCarlisle Ha ha! Never doubted it.

Stefan Kottwitz

@PauloCereda The doctor?

Werner

10:33 PM

@DavidCarlisle You should write to the cellspace author. You know the details about the issue there related to \@endpbox.

Joseph Wright

@DavidCarlisle Fix in hand

David Carlisle

@Werner just a missing \par here (with par added) line 37 \CheckCommand*{\@endpbox}{\@finalstrut \@arstrutbox \par\color@endgroup \egroup \hfil}% (I'll be offline for a bit)

@JosephWright thanks

Transcript for

$TeX - LaTeX$ TeX, LaTeX and Friends