TeX, LaTeX and Friends

5:36 AM

@UlrikeFischer I hope you like my review essay!

6:12 AM

@JosephWright ;-) I'm not sure about the mag and zero tests. I assume David added them for a reason and being conservative sounds safer. I'm not sure what you mean regarding the kernel hooks, I need to check what I did for dvips (it was some time ago but I tested which hook dvips needs)

Joseph Wright

@samcarter github.com/josephwright/siunitx/issues/608 :)

@UlrikeFischer On the kernel hook, for dvips, I mean do we need \@kernel@before@begindocument { \__kernel_backend_first_shipout:n { ... } } - wouldn't just \__kernel_backend_first_shipout:n { ... } work

@UlrikeFischer If we do want to test \mag/0pt, I'd put it in common code in l3pdf and issue an error if they are wrong - it's unsupported and likely to give an erroneous PDF

@UlrikeFischer There's a reason I wrote a comment rather than pushing any code :)

Ulrike Fischer

6:50 AM

@JosephWright i will have to check but I think I wanted to be behind some code from the pdfmanagement.

@JosephWright I think one of the presentation classes (powerdot? ) use mag.

Joseph Wright

7:06 AM

@UlrikeFischer Ah, right: we should document that :)

@UlrikeFischer I'll look round

@UlrikeFischer My plan is to push a separate branch commit with a suggestion: as it's quite a modification of your PR, that seems the best way to go initially - we can cherry-pick if the plan looks good

Ulrike Fischer

7:21 AM

@JosephWright I'm happy with everything that gives me a \pdf_set_pagesize: for hyperref to use ...

Joseph Wright

@UlrikeFischer :)

Ulrike Fischer

@JosephWright but I should really check it that was the reason, perhaps I simply was overcautious in delaying.

David Carlisle

7:33 AM

@UlrikeFischer @JosephWright the don't do anything if \mag is set translates to don't break platex By the time I tried to normalize this over the different back ends, the Japanese setup already had a lot of infrastructure around scaling and horizontal and vertical directions and we couldn't get anything to work in a compatible way other than bailing out and not doing anything in graphics and letting the platex macros handle it.

Joseph Wright

@UlrikeFischer Would be good to have a clear idea

@DavidCarlisle Ah, right: I guess I should look at that and see if they are still doing that

@DavidCarlisle There are all sorts of rubblings in the Japanese user community, but I think we will want to find a way to address this. I'll add to my to-do list

@DavidCarlisle We should be checking the engine then I guess, really

@DavidCarlisle If we can't fix the underlying issue, we should at least put the engine wrapper in such that it applies 'up front'

@DavidCarlisle, @UlrikeFischer I'll work on some code later - I think I can provide a plan :)

David Carlisle

@JosephWright yes (like the \stockwidth test) it's clearly "wrong" but allowed me to get signoff on all the drivers (including a change to dvips) and from the platex maintainers, so everything worked same way, but with more central control and expl3 being loaded in the format simplifying timing issues, we don't necessarily need to make the same choices now.

Joseph Wright

@DavidCarlisle Indeed - what we do want I think is clear notes in the code and an idea of what is 'expected' of third parties

LaTeX2e <2022-06-01> patch level 2

David Carlisle

@JosephWright same here

Joseph Wright

@DavidCarlisle :)

@DavidCarlisle I suspect the issue with pLaTeX comes from github.com/texjporg/jsclasses/blob/…: Google Translate gives me:

% In the standard document class, there was an option to specify the number of points in the body, but the number of points was only 10, 11 and 12, and different class option files were read for each. What's more, it wasn't very convenient because the fonts could be a little out of balance except for the standard 10 points. Here, 9 points, 21, 25, 30, 36, using the hand of increasing the font size to make the page smaller and expanding it overall with the | \ mag | primitive of \ TeX,
% 43 points, 12Q, 14Q can be specified.

@DavidCarlisle The way they do it is to set the real paper dimensions, then apply the \mag, but then they reset \papersize to the 'apparent' size. I'm pretty sure I can set up to handle the \mag correctly ...

David Carlisle

7:53 AM

@JosephWright sounds good (I didn't have google translate so was relying on my natural abilities at reading Japanse:-)

Joseph Wright

@DavidCarlisle :)

@DavidCarlisle I'm going to do some tests, work out what is actually being put in the dvi, then make sure I can get to the right values from the data available

\documentclass[uplatex,papersize]{kiyou}
\makeatletter
\show\jsc@papersize@special
\showthe\paperwidth
\showthe\mag

@DavidCarlisle ^^^ A place to start

@DavidCarlisle I guess you were writing the code pre-eTeX? We can use e.g. \dimexpr\paperwidth/1000*\mag\relax (or probably the FPU) here

David Carlisle

@JosephWright etex2 came out about the same time as 2e (how I came to write etex.sty) but we didn't use it for another decade, as lots of people ran old engines, and lots of tex sysems eg y&y never supported it

Joseph Wright

@DavidCarlisle Indeed - what I mean is that doing the 'reverse the \mag' work is easy with e-TeX but more tricky without it

David Carlisle

@JosephWright well yes but that's not why it's not there, I did have the division implemented, but I broke platex and took it out....

Joseph Wright

@DavidCarlisle, @UlrikeFischer Think I have a few ideas that deserve a mail - need to do some day-job work, but expect something later (incl. \stock..., etc.)

@DavidCarlisle Hmm OK, clearly going to need a bit of work but I think I can recruit some Twitter followers for that :)

@DavidCarlisle, @UlrikeFischer I notice that there's no consideration of landscape ...

David Carlisle

8:07 AM

@JosephWright landscape just swaps height/width after you know what size you want so doesn't affect this, I think

Joseph Wright

@DavidCarlisle Depends how you think things should look in a PDF: should landscape pages be rotated by the viewer or not

David Carlisle

@JosephWright for (pdf)lscape I see that but for overall document size I don't think rotation comes in to it you are just setting the media box, "landscape" just means width>height

@JosephWright meanwhile: texlive.net/run?%5Cshowthelog

Joseph Wright

@DavidCarlisle Ah, cool - something we should likely ensure then :)

@DavidCarlisle :)

@DavidCarlisle So no more option clashes for us

Rmano

8:25 AM

I think I had a hallucination ... that \documentclass[12pt]{article} spat an unused option warning from yesterday to today morning... I should sleep more. (Now it's all ok).

Joseph Wright

@Rmano If you were loading e.g. siunitx you'd have seen that - a bug we've now fixed in the new option handling

Rmano

@JosephWright Ah, yes, siunitx is in my standard preamble ;-).

David Carlisle

@JosephWright I added a summary of the above to gh

Plergux

8:49 AM

@barbarabeeton I'm not so worried about my typing, I think it's more my style choices that are going to be a stumbling block if anything. 😅

Plergux

9:02 AM

@barbarabeeton And for me as a "visual" person most of the time the diagrams and graphics are for my own benefit rather than the readers to make sure that I am saying the right thing. :þ

Ulrike Fischer

@JosephWright the mediabox of should reflect the real page size, so have a larger width in landscape. It should not try to fake it with some rotation. I just managed to stop dvips to do this. It breaks the orientation of annotations.

samcarter

9:19 AM

@JosephWright Nice! The new output is now consistent with \ang

Talking about this ... is it by design that `angle-symbol-over-decimal` only works with `\ang` and not `\qty`?
\documentclass{article}
\usepackage{siunitx}

\begin{document}

\qty[angle-symbol-over-decimal]{10.7}{\arcsecond}
\ang[angle-symbol-over-decimal]{;;10.7}

\end{document}

Ulrike Fischer

10:29 AM

@daleif regarding nameref and memoir: this looks quite messy. memfixh seems to test for the wrong command (\NR@sectm@m instead of \NRorg@M@sect) and other stuff looks quite odd too. Imho this should cleaned up. Side remark: if you define \NR@nopatch@sectioning nameref will no longer patch the sectioning commands and memoir could setup \@currentlabelname directly.

Joseph Wright

@UlrikeFischer Hmm, so why is that not required to be 'known' to the backend code?

@samcarter Er, broadly: it's really for arc-style angles used in astronomy (it used to be called astroang in v1)

@DavidCarlisle Thanks

Ulrike Fischer

@JosephWright sorry what should be known?

Joseph Wright

@UlrikeFischer Because there are two ways one might want a landscape page: to show in a viewer as portrait-with-content-rotated, or with the entire page rotated in the view

@UlrikeFischer One approach 'really' swaps the width and height, the other doesn't

@UlrikeFischer I'm surprised that the backend doesn't have to know the landscape setting to interchange \pageheight and \pagewidth if required

Skillmon

@JosephWright woops, used the wrong GH-account on your issue (that's my work-account)

Joseph Wright

@Skillmon Want me to zap?

Ulrike Fischer

10:41 AM

@JosephWright if you set up a landscape layout (for all pages) then the mediabox should reflect that. You simply set width=27cm and height = 20cm. There is no rotation involved, it is only a different layout. What dvips made instead was to setup a portrait mediabox, to rotate the content and to tell the pdf viewer to rotate in the other direction.

That was perhaps needed to be able to print, but today printer driver can handle such layouts and one should make it more complicated than needed in the pdf.

Skillmon

@JosephWright yes, probably better. Should I then readd the comments under my "reall" account?

(which I could do later today when I'm back home)

Joseph Wright

@Skillmon Done

Skillmon

@JosephWright thanks

samcarter

@JosephWright yes, but in astronomy, arcs are basically used like a length unit, so it would be nice to use \qty{10.7}{\arcsecond} instead of having to count empty ; in \arc

Joseph Wright

@samcarter I'll check, but I think v1 and v2 only ever applied this to to arc format as it's basically a 'feature' of arcs

samcarter

10:51 AM

@JosephWright yes, it never worked for \SI, but is there are reason to not make it work with \qty?

Joseph Wright

@samcarter Arc angles have special implementation to allow astronomy-style formatting, which doesn't apply to general \qty ones - I'd have to allow it to apply to all units, there would be an performance impact and some complex coding - really feels borderline to me as this is something that only applies to arcs

samcarter

@JosephWright ah, ok. Thanks for the explanations!

Joseph Wright

@UlrikeFischer Thanks - I'd misunderstood some code in a Japanese class

Ulrike Fischer

@JosephWright perhaps they have some old code too. But you can check this issue which shows why "cheating" with rotation is a problem: github.com/latex3/pdfresources/issues/24

RogUE

12:02 PM

Hi all,
I have a problem with in-text citations using biblatex. I want the in-text citations to show all authors when the number of authors are two or less, otherwise show the first author and use `et. al.
From the biblatex user manual , I assume the options are `mincitenames=2` and `maxcitenames=21, but it doesn't seem to work. There's this [post](https://tex.stackexchange.com/questions/121625/biblatex-modify-maxcitenames-to-work-in-three-cases) that advises to use only `maxcitenames=2`, but that doesn't work either.

My options are
\usepackage[
style=authoryear-comp,
maxcitenames=2, mincitenames=1,
sorting=nyt, terseinits=true,
isbn=false, dashed=false]{biblatex}

This answer shows the output I want.

Ulrike Fischer

@RogUE "that doesn't work" is not sensible a problem description. You should always make a complete example. Typically, if biblatex shows more authors than you want it is due to the uniquelist setting. So check if you have another entry with the same or similar authors.

2

RogUE

@UlrikeFischer That must be it. I have added uniquelist=false and it is working as expected. Thanks @UlrikeFischer

Ulrike Fischer

1:20 PM

2

Nearing Magdeburg ...

samcarter

@UlrikeFischer If you happen to come by a Sparkasse, have a look at their Geldottomaten

AlexG

1:34 PM

@UlrikeFischer And when talking to a local keep in mind that "a" in Magdeburg is pronounced (very) short.

samcarter

^^^ this kerning hurts!

yo'

@samcarter ooh keming!

samcarter

@yo' :)

daleif

3:02 PM

@DavidCarlisle regarding the coloneq issue, what exactly was the haft fix in newtxmath? Currently I have the list in mathtools generate the same symbols as newtxmath but it seemed that also had and error?

ohh, we're missing -: (then strange that :- does not exist)

David Carlisle

@daleif I sent an email to Michael sharp asking about change to coloneq (but not Coloneq) between txfonts and newtx. No reply yet. I think correct version is coloneq is :=Coloneqis ::=` (same as unicode-math) then define qq versions as same thing for legacy compatibility but that is not what newtxmath does for Coloneq just coloneq

mickep

Do -: and :- really mean something?

David Carlisle

@mickep EXCESS

mickep

@DavidCarlisle Ohhh

daleif

@DavidCarlisle but isn't -: completely missing?

David Carlisle

3:10 PM

@daleif :- you mean, yes that is not in Unicode

daleif

@DavidCarlisle no, you mentioned -: (U+2239) in github.com/wspr/unicode-math/issues/590, that does not seem to be implemented in newtxmath (it was just for completeness)

Perhaps I should get ready to go hear some German noise....

mickep

@DavidCarlisle And :: means "two colons". :)

David Carlisle

@daleif ah Ok yes, but despite my bold comment above I have no idea what -: is used for, so it being missing isn't much of a loss

daleif

And so basically switch \Coloneqq and \Coloneq such that qq indicates two colons. That would make more sense. Hmm, then we probably need and option to restore. Is there any way to list the macro name mappings that unicode-math uses?

Ahh, no that is not the naming convention, capital letter means :: the qq are just legacy

Are ::- in unicode?

David Carlisle

@daleif there is the tex file in unicode-math but I use github.com/w3c/xml-entities/blob/gh-pages/unicode.xml look for <mathlatex set="unicode-math">\dotsminusdots</mathlatex>

@daleif no

@daleif exactly

daleif

3:23 PM

Thanks. So there is dashcolon, but no colondash. How unsymmetric. I wonder if dashcolon is used in logic?

Might be an idea then to list: these are defined in unicode and named as such. These are additional symbols with no unicode equivalent ...

mickep

3:53 PM

@daleif The example here access.uoa.gr/mathbraille/index.php/en/math-symbols2/… looks strange to me.

(I sometimes wonder where all these odd symbols come from. I never see them in the articles I read.)

Then all of a sudden some binary symbol is used as an index... tex.stackexchange.com/q/648571/52406. Well well.

Alan Munn

4:51 PM

Oddly enough, TeX didn't make it on the list of languages used by those "Learning to Code". :) survey.stackoverflow.co/2022/…

Joseph Wright

5:28 PM

@AlanMunn :)

mickep

@AlanMunn But lua is there.

Joseph Wright

@mickep So it is, if you hunt

barbara beeton

6:40 PM

@Plergux -- Well I always value visual examples to aid clarity and understanding. Both prose and visual backup are necessary, really. Can lead to instant recognition of a problem if something is out of kilter.

samcarter

@AlanMunn That's clearly the selection bias from showing the banner only on SO and not on the other sites of the network. So of course a language with its own site will be under represented :P

Joseph Wright

@samcarter x:)

David Carlisle

@JosephWright ! LaTeX3 Error: No property given in definition of key 'toplevel'. ?

Joseph Wright

@DavidCarlisle ???

David Carlisle

@JosephWright you want a mwe? I thought my question was clear:-)

\documentclass{article}

\usepackage[noprint,hyperref=false,multicol]{doc}
\newcommand\tst[1]{\typeout{\string#1: \meaning#1}}
\begin{document}

\makeatletter
\typeout{===}
\tst\ifdoc@noprint
\tst\ifdoc@hyperref
\tst\ifdoc@multicol
\tst\ifdoc@debugshow
\tst\doc@idxtype

\DeclareKeys[doc]{toplevel=true,idxtype=abc,debugshow}
\typeout{===}
\tst\ifdoc@noprint
\tst\ifdoc@hyperref
\tst\ifdoc@multicol
\tst\ifdoc@debugshow
\tst\doc@idxtype

\end{document}

barbara beeton

6:47 PM

@yo' -- That's not kerning; that's a collision.

David Carlisle

@JosephWright ah you fixed it

@JosephWright clearly explaining to a duck helps debugging

Plergux

@barbarabeeton Very true. And if you can make a point with a little bit of humour it helps all the better. :þ ualberta.ca/arts/faculty-news/2019/october/…

mickep

@DavidCarlisle/@barbarabeeton: There is in unicode lowercase greek sansserif bold and lowercase greek sansserif bold italic, but not lowercase greek sansserif or lowercase greek sansserif italic. Do you happen to know why?

David Carlisle

@mickep ~~ITS NOT MY FAULT~~ I know:-)

mickep

@DavidCarlisle Hehe

David Carlisle

6:57 PM

@mickep hang on, link coming up....

mickep

(Not that I long after them, given how the existing sans alphabets often look.)

barbara beeton

@daleif -- eqq in other contexts means a "two-line equal"; as in \eqqplus \pluseqq, approxeqq, etc. And the uppercase letter (first seen in DEK naming) doubles that component of a symbol. I'd prefer to stick to that interpretation. But it does leave a problem with "dash-colon" and "colon-dash", for which I don't know a meaning.

mickep

@barbarabeeton There is an example on access.uoa.gr/mathbraille/index.php/en/math-symbols2/…. Something something with arithmetic sequences. But I don't know if it is "for real".

David Carlisle

@mickep I couldn't find what I was looking for but unicode.org/mail-arch/unicode-ml/y2005-m05/0171.html lists.w3.org/Archives/Public/www-math/2021Jun/0023.html

@mickep we have been pleading for this for 20 years....

barbara beeton

@Plergux -- Oh, that's lovely!!! I've got quite a list of people who need to see it!

David Carlisle

7:05 PM

I blame @barbarabeeton

mickep

@DavidCarlisle Funny coincidence: I browsed Courant-Hilbert the other day and noticed the sans serif Gamma. Well, it seems difficult. And still a bit funny that the bold versions are there.

barbara beeton

@mickep -- I really don't know. They were trying hard to restrict the "math alphanumerics" to two plane-1 blocks, so they left out things that weren't explicitly requested, and those weren't on the STIX list. But they also left out things that were explicitly requested, including callligraphic separate from script. I no longer have access to the paper documentation I used to compile the STIX list, but the origins of what's in the list are in an accessible table.

mickep

@barbarabeeton Yes, stixtwo is one of the few that has both calligraphic and script variants. Cambria and Lucida also have them. I wonder if there were more.

barbara beeton

@mickep -- Thanks. I wasn't aware of that, and am not enough of a mathematician to know independently what it might mean. (I had to take it on faith that what the other STIPub organizations requested were legit. That was my remit, and there are a couple of items that I really do know are questionable, but I had to include them anyhow.)

@DavidCarlisle -- Well, Murray Sargent, I believe, is still the UTC rep from Microsoft, so twist his arm.

mickep

@barbarabeeton It is added as sstyXX (of course different numbers in different fonts, and also different version added as special in different fonts... _)

barbara beeton

7:18 PM

@mickep -- I was able to find a published document where script and calligraphic were both used, with clearly defined different meanings. It still took years for Unicode to accept the concept.

mickep

It is not that sans greek is really needed, but given that they add all kind of emojis and so makes me wonder a bit...

@barbarabeeton Hehe, amazing.

barbara beeton

@mickep -- There was a document promulgated by the U.S. National Bureau of Standards (or its successor, NIST) on proper scientific typography that specified a sans cap Theta for something. So one letter, UTC included the whole alphabet. Unfortunately, that NIST document has been superseded, and I can't find the original, which I did use for the STIX compilation.

mickep

@barbarabeeton Thanks, always interesting to hear about how things went.

barbara beeton

@mickep -- Cambria and Lucida were both significantly influenced by the STIX collection. (I was supposed to consult on the development of Cambria, but unfortunately, became ill and was out of commission as far as that project was concerned for the duration.)

Plergux

@barbarabeeton I made quite a few of these small comics when they shared the data with me :þ

Joseph Wright

8:04 PM

@DavidCarlisle Er, you know you are trying to set keys in the defining phase ...

@DavidCarlisle Did you mean \SetKeys?

David Carlisle

8:57 PM

@JosephWright yes, but I noticed as soon as I pasted it here:-)

Joseph Wright

@DavidCarlisle :)

0

Q: Using \Makelowercase in includegraphics

I feel like this should be very simple to do but even using multiple other SE solutions I have not been able to figure it out. I would like to pass an argument like Evans1963 to a command that I have defined: \newcommand{\placefigure}[1]{ \begin{wrapfigure}{O}{0.47\textwidth} \centeri...

Looks like a job for ... expl3 ...

David Carlisle

9:14 PM

@JosephWright \lowercase{includegraphics[width=0.45\textwidth]{#1_results}} :-)

David Carlisle

9:39 PM

@JosephWright grr I must have misread it, I thought it showed passing tests but ! LaTeX Error: File multicol.sty' not found.` I guess I should use mulicol=false...

Transcript for

$TeX - LaTeX$ TeX, LaTeX and Friends