Today I discovered there's a duck farm around the corner from my house.

hi there, I'm trying to convert from mathjax to latex and am trying to find an equivalent to the mathjax bbox package, anythoughts for something that I can quickly replace it with?

which renders as

I've finally discovered that Unicode doesn't provide medium weight sans serif italic Greek letters. That means that in physics, certain tensors named with Greek letters (permittivity, permeability, Levi-Civita, etc.) can't really be used according to ISO 80000-2 guidelines. Apparently ISO didn't think about that.

Bold is great for coordinate-free notation but regular is preferred for index notation.

@AndrewMicallef Try searching the main site for "box around equation" and you'll find many solutions.

ok, looking, I was just hoping for something with similar syntax that I can just do a find and replace with. suppose I should have to define that myself huh

@AndrewMicallef There are different ways of doing it in different contexts.

I see, I think \boxed should do the trick (and I just throw away the optional arguments because the defaults seam to give a near enough equivalent)

yup looks good

@AndrewMicallef \boxed (amsmath) ? (ah sorry you just said that:-)

@LaTeXereXeTaL unicode.org/mail-arch/unicode-ml/y2005-m05/0171.html also last part of devblogs.microsoft.com/math-in-office/…

@DavidCarlisle -- Thanks for links. The latter is dated late 2020, so I should still be able to get in touch with Murray. @LaTeXereXeTaL has identified an ISO standard that explicitly requires some sans letters, so maybe there's an improved chance to get them added. By the way, I did send Murray, several years ago, some example papers that used calligraphic and roundhand script with contrasting meanings. I'll nag him again about that. sigh.

@samcarter_looks_forward_TUG'21 -- I think you might enjoy this: youtube.com/watch?v=56aEudL2Hr8 The perpetrators were interviewed this morning on the radio, and it was hilarious.

@DavidCarlisle Thank you for the links. Now I have to learn about the "Private Use Area."

@barbarabeeton Yes indeed. In physics, there are quite a number of tensors that are labeled with Greek letters. Among them are permittivity, permeability, and the Levi-Civita tensor (aka the Levi-Civita tensor). The existing boldface sans serif italic letters are perfect for coordinate-free notation but a normal weight character is called for with tensorial index notation.

I seem to have found them here on page 41: ctan.mirror.rafal.ca/fonts/stix2-otf/STIXTwoMath-Regular.pdf

Now to figure out how to use them.

@PauloCereda causing trouble again: github.com/latex3/latex3/issues/879

Hello!

@DavidCarlisle It's documented. Unless you run in verbose mode, arara will not be in interactive mode.

@TeXnician I shall refer you to @UlrikeFischer for making the scurrilous suggestion that people read the documentation! Although here people probably don't associate \show with "fatal error in non interactive mode"

@DavidCarlisle Well, what does TeX do in a non-interactive shell, i.e. when there is no input stream?

@TeXnician it does an emergency stop but I mean whatever tex actually does, if users think of \show\foo as \typeout{\meaning\foo} then you get a shock when you try it...

@DavidCarlisle Ah, well, that's of course right. But I somehow got caught by Frank's “analysis”.

@TeXnician I think your description "It's a misunderstanding of how arara works and what it tries to do. Running this document is expected to yield whatever TeX yields" is very much from the point of view of an arara maintainer. I don't think even experienced tex users would describe it that way. Most people (including I guess me and Frank) would expect that to mean arara: pdflatex was the same as running pdflatex document

... but there is no easy way to run pdflatex manually with no input stream available so the emergency stop on input request is something that is normally only achievable by explicitly requesting batchmode

@DavidCarlisle You are absolutely right that this is a quirk of arara and of course documentation is more or less something the least will read. On the other hand, batchmode would require to run TeX with some arguments which we actually don't do if the user does not ask for it. So it's more or less a technicality.

I just wanted to clarify things as Frank speculated about intentions and what arara is “meant” to do without having used it. And as JouleV started to use it the issue was a pretty good fit to try to explain what happens :)

@TeXnician yes sure (and I think the second message you posted explains it well) Of course running arara in cygwin is too painful to try too often:-)

Good morning, everyone!

Quick question, what's the best way of creating a multi-page code listing using the listings package for a code file (read with \lstinputlisting) that doesn't fit on a single page in such a way that it doesn't interfere with preceding shorter, floating listings?

If I let the longer listing float, it doesn't get split across pages anymore -- not entirely unexpected --; when I don't let it float it gets interrupted by floating listings, which is really confusing.

@chsk you can stop any floats floating past by using \clearpage at a suitable point

@chsk you just offended the Australians:-)

@chsk or of course you can split the listing in to two and float them both, perhaps using \ContinuedFloat from float package

@DavidCarlisle Thanks! That worked. Now it's still getting interrupted by subsequent floating objects though, is there a way to tell LaTeX to not place any floats until after the listing has been processed fully?

@DavidCarlisle Oh, that's a great idea, too. I'll look into that.

@DavidCarlisle I can live with that. ;)

@chsk after a \clearpage there should be no floating objects, that's what the "clear" refers to.

Aye --- but I now have (basically) \clearpage\lstinputlisting[...]{...} some text \begin{figure}[tp] ... \end{figure}, and the figure gets placed at the top of the page that still has part of the listing. I suppose I could use p instead.

@chsk oh you mean you have a [t] float immediately after that floats back? simplest is simply don't do that, don't use [t] , or use \suppressfloats[t] immediately after the non-floationg float

\suppressfloats[t] is what I was looking for, thanks! :) That way if more text gets added in between, the t float can still be at the top of a page.

@DavidCarlisle ooh

Different question while I'm here, when you have tables, figures and listings in a document, is there a standard and/or recommended order for \listoftables, \listoffigures and \lstlistoflistings?

@chsk no idea:-)

Fair enough! :)

@PauloCereda @TeXnician tried to protect your honour. I'd have simply closed the report with "the example uses Arara it's all Paulo's fault" which would have been simpler

@DavidCarlisle oh no

Thanks for your help, @DavidCarlisle!

(And to any Australians that might be present: no offense. ;) )

@DavidCarlisle does suppressfloats works now?? github.com/latex3/latex2e/issues/144

@UlrikeFischer define "now". It works when it works, and when it doesn't work it doesn't work.

@DavidCarlisle which mostly means it doesn't work when I want it to work ;-)

@UlrikeFischer but I think it will work here as the \clearpage before the listing will mean theer are no deferre floats to complicate thie issue

@DavidCarlisle thie= teh

@UlrikeFischer and theer =there and deferre= defer but I was close

arrrrggggggggggggggggggggg my 1992 self broke my 1993 self and confused me in 2021 :(

@DavidCarlisle You need a DeLorean :p

@DavidCarlisle ?

I thought I'd try regression-test instead of test2e.tex but that has \let\@@@end\@@end and that;s worse:

! TeX capacity exceeded, sorry [input stack size=5000].
\bm@command ->
              \mathchar 30017 \begingroup \long \def \begingroup \long \def ...
l.20     $}

!  ==> Fatal error occurred, no output PDF file produced!

@DavidCarlisle why does bm care about end?

@UlrikeFischer because it seemed like a good idea at the time? I needed a non-expandable marker token to insert at the end that was unlikely to get there by accident and \@@end seemed like a good choice.....

@DavidCarlisle ah, so \@@zzz would do to?

@UlrikeFischer has to be defined and not expandable

@DavidCarlisle isn't \@@zzz always defined in your packages?

@UlrikeFischer yes but it's got to be ifx testsable and the most natural definition \let\@@zzz\@@end would have same problem

@UlrikeFischer I just pushed a change as test now passes as I locally reset @@end inside the test.

@DavidCarlisle yes saw it. Has expl3 a token that you could instead of \@@end?

@UlrikeFischer I could probably use \protected\def\zzz{} in these modern times (not sure actually if \protected is enough or if it really has to be unexpandable, but I'm not planning on changing it:-)

Does someone know a MathML renderer which supports table alignments? (maligngroup/malignmark) Both Firefox and MathJax seem to ignore them...

@MarcelKrüger mathplayer did, more or less nothing else, they were a pig to implement, I supported some simple cases in one xslt mathml to tex version.

@MarcelKrüger they are not in mathml-core so likely to get deprecated in mathml4

@MarcelKrüger oh word might

@DavidCarlisle Is that a more general rule? Are all(/most/many) of the features which aren't in core likely to be deprecated in mathml4?

@MarcelKrüger no. Some things eg all the elementary math layouts and long division etc and mfenced will be in mathml4, but to get mathml4 published we'll need to claim two implementations and it may be hard to polyfill malign (most of the rest already have javascript polyfills in either mathml-refresh or mathjax)

@MarcelKrüger malign looked like a possibly reasonable idea in but it means that you have to carry so much extra alignment information it really complicates the mapping from mtable to css or tex or anything table model so it has a high cost to implement and as noone really understands it it is hard toi find documents using it so there is never a great push to get anything implemented. the current idea in mathml4 is that you will be able to use mtable to get the alignment

@DavidCarlisle Is the second implementation going to be Word or is the plan to make two Javascript implementations?

@MarcelKrüger but then use new semantic intent attributes (not fully specced out yet) to specify "this row is teh expression) ax^2+bx+c" even if there are table alignment cells at intermediate points

@DavidCarlisle That answers the big question I had about that: How would this ever work for accessibility...

@MarcelKrüger you only need two implementations of each feature. But anway too soon to say yet. For mathml3 we included Word, my xslt to tex, various browsers. c

@MarcelKrüger that was the motivation for malign that accessibility would be improved by keeping the expressions in one piece but withthe alignment points specified but ot didn't really work. I think just using mtable but with semantic attibutes will work out better, let me see if I can find a public draft...

@MarcelKrüger mathml-refresh.github.io/mathml/docs/intent#tables

@DavidCarlisle Thank you, that looks interesting.

@MarcelKrüger how can I catch the mathml output in your code, so that I can e.g. write it to the pdf?

@UlrikeFischer I haven't found a nice interface yet, but I thought that the easiest way if it get's written directly into a PDF object and then just the object number is returned. Then it could just be a register which can be queried after the formula. Or do you have a better idea?

@UlrikeFischer Also from where should it be accessed? Lua code? TeX code?

@MarcelKrüger actually I asked because I want to test some interfaces ;-). If it is a stream object, it probably can't be used as actualtext too (not sure, need to check the docu even if David will complain).

@UlrikeFischer Do we want MathML as ActualText?

@MarcelKrüger no idea, that is a rather open question. Probably not, some normal text will be better. So for now I think simply an object reference would be ok.

@MarcelKrüger I would prefer to write the filespec dictionary in tex, the interfaces are all there.

@UlrikeFischer Including the /EmbeddedFile stream or only the Filespec?

@MarcelKrüger I think the luacode should write the /EmbeddedFile stream and return its obj reference and I then write the FileSpec with the AFRelationship.

@UlrikeFischer Ok, I'll create the interface.

@UlrikeFischer In the latest version you have \luamml_get_last_mathml_stream:e which provides the object number for the last seen formula. (The argument is for any additions to the embeddedfile directory, /Subtype is set automatically)

@MarcelKrüger just trying. Can you add a build.lua?

@UlrikeFischer Should be there now.

@MarcelKrüger thanks, that makes testing easier (btw: the build.lua is copied and installed too).

@UlrikeFischer Fixed...

Hi. How am I supposed to set TEXDOCS? I have in my local TEXMF tree an experimental version of circuitikz (and its manual), but texdoc insists on not seeing it, so I should be doing something wrong.

⌂2.60 [romano:~] 3s % TEXDOCS=/home/romano/texmf: texdoc -l circuitikz
 1 /home/romano/texlive2021/texmf-dist/doc/latex/circuitikz/circuitikzmanual.pdf
   = Package documentation
Enter number of file to view, RET to view 1, anything else to skip: ^C
⌂2.55 [romano:~] 3s 130 % kpsewhere circuitikz.sty
/home/romano/texmf/tex/latex/circuitikz/circuitikz.sty
/home/romano/texlive2021/texmf-dist/tex/latex/circuitikz/circuitikz.sty
[romano:~] % ls -l /home/romano/texmf/doc/latex/circuitikz/circuitikzmanual.pdf

@Rmano do you know how to find out what to put in the link to link to a specific section in a pdf with texdoc?

In my case, I want to link to the l3regex section of interface3...

@Skillmon I just opened the side bar with the outline and copied the link from there. It's something as ...pdfurl#xy where xy is variable.

@Rmano in which viewer? I don't see any specific link in evince (perhaps I can mess with Zathura...)

@Rmano texdoc.org/serve/interface3/0#page%2055 works :)

Both in the internal firefox one and okular. It's this: texdoc.org/serve/interface3/0#chapter.8

@Rmano that's better :) Perhaps I should use pdf.js more often :)

Ok, found something: tex.stackexchange.com/questions/46003/…

The problem is that I do not have ls-lR in my TEXMFHOME (it would be really awkward to have to use it...), but yet texdoc is not able to find it...

@MarcelKrüger some symbols need some improvement and I will have to ponder about interfaces but beside this the ngpdf.com editor recognize the attachement:

@UlrikeFischer Without unicode-math every family needs a mapping table to translate it to unicode-math. I included the tables for the cmmi and cmsy fonts which cover many of the more common symbols, but since it's in many parts hard to automate my motivation ran out after that, but they are easy to add. It looks like cmex tables are needed to print that properly.

@MarcelKrüger yes, I got that, I only tried to push the system to the limits ;-). And at least with luatex unicode-math is better anyway.

@MarcelKrüger one for you? :-? tex.stackexchange.com/questions/594414/…