@PauloCereda oi 😲

Ping me if you respond this "trivial" question.

@MoneyOrientedProgrammer why in the world you need something like that?

@yo': I just extended D.E.Knuth exercise: "What do you think happens when you type 4 hyphens in a row?". Just for academic purposes. :-)

@MoneyOrientedProgrammer Look at the answer to Exercise 5.1.

@HaraldHanche-Olsen: OK. Thank you.

@MoneyOrientedProgrammer but note the texbook doesn't cover luatex

@DavidCarlisle OK. I don't need luatex. :-)

I am interested in understanding how the TeX parser works.

assetstore.unity.com/packages/tools/gui/texdraw-51426

@MoneyOrientedProgrammer but ligaturing is handled by the font not by the parser (which is why it's different in luatex)

I have no idea how, for example, the top horizontal line of \sqrt can expand when its contents changes in width. The TeX engine modifies the font metadata to vary the top horizontal line?

@DavidCarlisle I assume an explicit kern will suppress ligatures in luatex too? It's just grouping that doesn't?

@HaraldHanche-Olsen yes

@MoneyOrientedProgrammer the top bar is a rule like \hrule drawn by tex it is not part of the glyph from a font.

@MoneyOrientedProgrammer Run the following code in plain TeX, hit return when it stops, and look at the log file afterwards. Note in particular the result of the second \showlists.

@DavidCarlisle OK. How about \frac? There is no a font associated with \frac? It is just a \hrule?

\showboxdepth\maxdimen
\showboxbreadth\maxdimen

$\sqrt{\sin x}\showlists$\showlists
\bye

@MoneyOrientedProgrammer Yes, with vertical spacing defined by the font.

@MoneyOrientedProgrammer well it's a rule drawn by tex yes

@MoneyOrientedProgrammer as it says, the information is in the log, or use \tracingonline=1 so it is on the terminal as well.

@DavidCarlisle the puzzle of the day: where is the duck gone?

\documentclass{article}
\usepackage{l3pdf}
\ExplSyntaxOn
\pdf_uncompress:
\ExplSyntaxOff
\usepackage{graphicx}
\font\fontB={Segoe UI Emoji:mode=node;+colr;+ccmp;+dist;}

\begin{document}
 \saveimageresource width 10pt  {example-image-1x1.pdf}

 \pdfextension save\relax
 \pdfextension setmatrix{1 0 0 1}%
 \useimageresource\lastsavedimageresourceindex
 \pdfextension restore\relax

{\fontB ^^^^2603\fontB ^^^^^^01f986}

\end{document}

@MoneyOrientedProgrammer you missed the point of the example as you didn't show eh second \showlists

@UlrikeFischer yes I saw that in the gh issue. automatic duck erradication seems OK to me

@DavidCarlisle it is not erradicated, it has found a secrete place to hide ;-)

@MoneyOrientedProgrammer V

### math mode entered at line 5
\radical"270370
.\mathop\nolimits
..\mathord
...\fam0 s
..\mathord
...\fam0 i
..\mathord
...\fam0 n
.\mathord
..\fam1 x
### horizontal mode entered at line 5
\hbox(0.0+0.0)x20.0
spacefactor 1000
### vertical mode entered at line 0
prevdepth ignored

! OK.
l.5 $\sqrt{\sin x}\showlists
                            $\showlists
?

### horizontal mode entered at line 5
\hbox(0.0+0.0)x20.0
\mathon
\hbox(9.18924+1.21072)x27.99306
.\hbox(0.39998+9.6)x8.33336, shifted -8.38928

@MoneyOrientedProgrammer see in the first showists its a \radical atom but in the second showlists after the math mode it has been transformed into an hlist (appendix G of the TeXBook) and has an explicit \rule

@MoneyOrientedProgrammer you killed the job after the first \showlists

I did not kill the job. The job killed itself, maybe?

@MoneyOrientedProgrammer yes that's the first now hit return and you will get the second

OK. Thanks. I did not know the question mark means.

@MoneyOrientedProgrammer type h and it tells you:-) But this is the standard tex error prompt that you get for any tex error, the \show... debugging commands act like errors even though they are intentional

how the numbers are generated?

@MoneyOrientedProgrammer well a log file isn't tex input.

@MoneyOrientedProgrammer you are running log output through tex: that isn't intended to ever work. It's not clear why you just removed the leading ....

@DavidCarlisle My intent is to get the same output of $\sqrt{\sin x}$ but using the very very primitive macros of TeX.

@DavidCarlisle The leading .... were removed because I thought they are not needed.

I thought they were there just for enhancing readability.

@MoneyOrientedProgrammer none of it is tex. There is no command called \mathon for example that is just logging the part of the internal action of the first $ \kern1.7107 is a kern of 1.7107pt and the fact that there is no space between the n and 1 tells you that it is an implicit kern specified by the font metrics not an explicit kern from \kern1.7107pt in the source.

@MoneyOrientedProgrammer \hbox(6.67859+0.0)x19.6597 is an hbox of height 6.68pt depth 0pt and width 19.66pt but you can not specify an hbox like that in source an hbox size is determined by its contents, this is just logging the result

Is there any command to get the primitive macros representing the $\sqrt{\sin x}$ out of the box? So I can see how the tex works behind the scene?

:5429549 \sqrt is \def\sqrt{\radical"270370 } so there is nothing below $\radical"270370 x$ there are no macros there, just tex primitives

OK. TeX primitives are fine too.

@MoneyOrientedProgrammer as above. $\radical"270370 x$

@UlrikeFischer almost makes it seem like an emoji font is an archive full of images:-)

@DavidCarlisle Thank you very much.

@DavidCarlisle ;-). I learned from my tugboat read that there is book written only with emoji, and that it was a good idea to add them to unicode as it made unicode known to the masses.

@UlrikeFischer Which artcile?

@JosephWright Yannis

@UlrikeFischer, @DavidCarlisle Hook code looking good I think

@JosephWright you mean the shipout part or everything?

@egreg, @barbarabeeton When you get a chance, could you look over learnlatex.github.io/en/lesson-10 and think about what might go into learnlatex.github.io/en/more-10

@UlrikeFischer I was just reading the shipout part, but more generally it's all looking good: I think we will bring some proper order to everything

@UlrikeFischer The Kernel Strikes Back

@JosephWright I tried to read and understand the shipout part yesterday evening but was at the end not quite sure if it does what I think it does (and what we will need) ;-). I will have to reread it. But beside this it is very good that this go back into the kernel.

@UlrikeFischer See my code comment ;)

@UlrikeFischer I think the ideas do the same as the current hooks: we might need to make sure the implementation does!

@JosephWright just reading it.

@JosephWright it would be good to get rid of all "Patch"-whatever parts. But seeing all the boxes there: we need also an understanding about the box-levels to avoid that links crossing pages breaks.

@UlrikeFischer Yes, that is true

@UlrikeFischer As I've commented, I would not patch anything, just check after-the-fact for any known issues and warn

@JosephWright yes and try to provide an alternative package if we know that the original will not be adapted.

@UlrikeFischer That too, yes

@JosephWright if we get an updated latexonline with luatex, need a way to access it... could be an internal script like the zip file packaging but then it's not author setable, currently the logic is use pdftex unless fontspec appears in which case use xetex,...

but could check comments but should it be "any comment with luatex" as in % luatex or should I go for the %!TEX program=luatex or whatever the syntax is? (I'm currently thinking of any comment with any of luatex, xetex, pdftex, sets the engine)

@MarcelKrüger I did sent a message to the context list regarding the emoji problem.

@UlrikeFischer 🦆⟶👻

@DavidCarlisle google translate is failing here ...

@UlrikeFischer Oh, thanks. I mailed Hans this morning off-list with a plain TeX example with the generic fontloader too.

@UlrikeFischer Etwas über unsichtbares Emoji-Ghostwriting

@DavidCarlisle oh no

@DavidCarlisle I any comment with an engine name sets the used engine, that could lead to conflicts: E.g. what happens if an initial comment says: "% Can be used with luatex, xetex, or pdftex"?

@MarcelKrüger It only has to work on the learnlatex tutorial pages so in that case just get @PauloCereda to whack the author with a cricket bat until they conform.

@DavidCarlisle ooh a plan

@DavidCarlisle Sounds like a perfectly reasonable plan.

@MarcelKrüger the current test will use xelatex to compile the document if it contains \iffalse dont use fontspec \fi as it contains the string fontspec

@MarcelKrüger but the possible advantage of using %!TEX program=luatex would be that people could copy the file locally and it might work in some editors (personally I'd rather an emacs lisp local variable setting:-) I suppose could implement the simple test but actually use the longer form in the on-site comments.

@MarcelKrüger oh, overlap ;-). If it is fixed soon we could add it to the release I want to make because of the tool-bug anyway.

@DavidCarlisle Go for the TeXworks/TeXshop statements I think

@DavidCarlisle We can then explain about the commets in 'more text'

@JosephWright will add it one evening, should work with overleaf already, less so with latexonline

@DavidCarlisle will luatex be available soon?

@DavidCarlisle Sounds good

@AlwaysConfused That's really 'What's the difference between a word processor and typesetting application'

@JosephWright don't cloud my don't do that message :-)

@UlrikeFischer the islandoftex people made a pull request....

@DavidCarlisle ooh

@PauloCereda actually not a PR in the github sense just an invite to the OP to pull a fork back:-)

@DavidCarlisle oh :)

@JosephWright Thank you. Can you refer some such applications that fix page numbers on text? as well as how does the content and index page is created? is there any software for that?

@AlwaysConfused Well, you've come to the chat for (La)TeX ...

@AlwaysConfused what do you mean by "fix"? Do you want to have some text always on page 2 or do you want references and table of contents to adjust to changes?

@AlwaysConfused There is no application that will do this. Either stuff fits on a page or it doesn't.

@JosephWright which packages or classes are not compatible with array??

@AlwaysConfused And at least in LaTeX if you use proper methods (\label and \ref etc) to refer to things, rather than hard coding references, then changes will be reflected automatically.

@AlwaysConfused If you want all the text to stay on the same page, stick to one page documents. :)

@UlrikeFischer For example tex.stackexchange.com/questions/37814/…

@UlrikeFischer Comes up from time to time with siunitx as I load array

@JosephWright that package has an option to support array: \DeclareOption{array}{\def\use@@pkg@@array@@variant{1}}

@UlrikeFischer Was just the first thing that I remembered seeing: I've not kept the emails, I suspect, but I know there are others

@AlanMunn Wow that is amazing

@AlwaysConfused PRetty standard LaTeX ...

@AlwaysConfused Word can do this too, although it's a lot clunkier. Don't ask me how, though. I haven't really used Word for many years.

@AlanMunn It's not so bad, although it's not 100% smooth

quack

@PauloCereda Quack

@JosephWright ooh a duck

@PauloCereda extra dinner

@DavidCarlisle oh no

@AlanMunn did you see in tugboat the linguistic article?

@UlrikeFischer I guess I'm in the group of "linguists [who] have inherited Saussure’s disdain of the written word" :) Well maybe not disdain, but it's not something I'm much interested in.

@JosephWright I have lost track of the number of horror stories I've heard of word documents – cross references in particular – breaking once they get too big, as a thesis often does. But perhaps this is a thing of the past. I haven't heard the complaint for a while. (But then nobody in the math department would use word for a PhD thesis!)

Should this kind of idea be patented? patents.justia.com/patent/7377785

@HaraldHanche-Olsen ER, yes: I had a lot of 'fun' with a thesis from a student, where we had EndNote to fight with (I had to hack it's internal labelling inside Word), then at the point of printing some of the cross-refs simply refused to work. A very stressful afternoon for me, as she'd had to leave me to print out for her ...

@AlanMunn I find it interesting. So many communication goes through writing in the one way or the other nowadays and there are obvious differences to speaking. For example the ... Joseph just used.

@MoneyOrientedProgrammer After a cursory reading, I'd say no. But of course, the patent has already been granted (in 2008 if I read that correctly), so now the question is how aggressive (and successful) the patent holder will be defending it. As far as I know, trying to defeat a granted patent is a million dollar expense, with no guarantee of success.

@UlrikeFischer Yes, that's why I wouldn't really admit to 'disdain'. But since we have all been illiterate in our lives, but still had a fully functioning language, I'm interested in uncovering that aspect of language and how we learn it. But I'm also one of those linguists who thinks that the properties that language has don't have much to do with communication.

@DavidCarlisle -- Please take a look at the last couple of comments under this question, in particular the mention of ntheorem and amsmath. The "recommendation" appears only in the catalog entry for amsmath, not in any of the amsmath` documentation. I guess something needs to be fixed, but I'm not sure what or how. Help!

@barbarabeeton that's weird

@DavidCarlisle -- any ideas? Does the suggestion to omit the reference to ntheorem from the catalog entry have merit? (I know essentially nothing about using ntheorem in any context.)

@barbarabeeton ntheorem changing trivlist as described in egreg's answer sounds like something a package shouldn't be doing

@DavidCarlisle -- Okay. Sounds reasonable. Should someone (you?) ask CTAN (Gerd?) to change the catalog entry? Obviously, I'm not going to suggest mentioning ntheorem in the amsmath documentation, even negatively.

@JosephWright Apart changing \mathrm{d} into d, you mean?

nearly ;)

@UlrikeFischer failed

@DavidCarlisle there are simply not enough nasty people in the world. Nobody takes 2 points from me ;-(

@UlrikeFischer ^

;-)))) yay

@UlrikeFischer now question is do you want to keep the 2-7 cycle or should I undo the downvote:-)

@DavidCarlisle it can stay.

I should probably have picked a worse answer (I just picked your first one:-)

@DavidCarlisle do I have worse answers?

@UlrikeFischer Much worse:-)

(any with ducks in, for example_

@DavidCarlisle something with ducks

@DavidCarlisle I tried to correct the typos first

@UlrikeFischer you should learn to type without typos.

@DavidCarlisle ooh

@DavidCarlisle -- speak for yourself.

@UlrikeFischer Nice: A palindrome in base 10 and in base 100 and 1000 at the same time! (And in base 1000000, but that doesn't really count.)

@HaraldHanche-Olsen and more importantly with the comma in the right place ;-)

@UlrikeFischer that, too – if only the comma itself were symmetric!

@HaraldHanche-Olsen yes that is a pity. With a german dot it would be perfect.

@UlrikeFischer or with an half-space... ;-)

@Rmano Yes, few things are more symmetric than empty space.

@UlrikeFischer Hans answered that my proposed patch "Sounds ok for now", so I guess/hope that will be fixed in the next ConTeXt upload.

@MarcelKrüger what is the patch? Can we apply it manually to test with out suite?

@UlrikeFischer

diff a/tex/context/base/mkiv/font-ocl.lua b/tex/context/base/mkiv/font-ocl.lua
--- a/tex/context/base/mkiv/font-ocl.lua
+++ b/tex/context/base/mkiv/font-ocl.lua
@@ -278,10 +278,11 @@ local function initializeoverlay(tfmdata,kind,value)
                         local s = #colorlist
                         local goback = w ~= 0 and leftcommand[w] or nil -- needs checking: are widths the same
                         local t = {
+                            start,
                             not u and actualb or { "pdf", "page", (getactualtext(tounicode(u))) },

I pushed it into a fontloader branch.

@MarcelKrüger looks good, I think we should simply apply it and update. The main question is if we should rename the fontloader to avoid confusion - probably yes, one could add a patched or change the date

@UlrikeFischer Yes, especially in case that this leads to issues it's probably best if it is separate. What is currently the default fontloader? The dated one or reference?

@MarcelKrüger a normal user will get the dated one, but one can request the reference folder through options. I will think about it, aber now I have to make dinner first.

@UlrikeFischer Enjoy your meal

@MarcelKrüger I will upload a new version, just building. I renamed the fontloader and added a comment to the other files.

@UlrikeFischer Thank you.

@MarcelKrüger perhaps it will be only tomorrow - I managed to forget an itemize and the docu didn't build (this mix of latex and context is driving me mad sometimes) ;-)

@UlrikeFischer I'm not quite sure which day "tomorrow" refers to currently, but I would be okay with either option.

@MarcelKrüger I meant today tomorrow ;-)

Is @mbork around here? Wanted to ask about the findings at tex.stackexchange.com/q/208889/9075. - My findings are that examplep > cnltx-example > latexdemo > showexpl. I am in contact with the authors of latexdemo and showexpl; and started a conversation with the examplep and cnltx-example. Feeling a bit bad to ask them in parallel about a feature. I am aware that LaTeX development takes time.

More background: github.com/cgnieder/cnltx/issues/4. - I think, my real issue is that the newest version of examplep (from mid 2005) is not uploaded to CTAN (details at tex.stackexchange.com/q/193102/9075) and that I cannot get LaTeX code highlighting to work (with the newest version from the author's homepage). l - The author of examplep does not reply and I am unable to quickly understand the code of the package. :p