@Noldorin \ShowCommand

@barbarabeeton My mother always wanted my siblings and me not to throw away things like old exercise books. On the other hand, I didn't want these things to take up space. When I was 15, my father bought the first scanner, and my mother and I agreed on the compromise of scanning and CD burning.

@UlrichDiez -- And these have all lasted. Have you had any problems with degradation? I think it's still not known how long CDs remain stable and reliable. And it's also unclear how long CD readers will be available. At least the material wasn't consigned to 8-inch floppies; you're probably not old enough to remember them. ("CD burning" sounds terribly ritualistic.)

Ah, it's non-expandable, I suppose?

@user202729 that definition isn't from amsmath

Ah it's amssymb

@user202729 so using it negates the intention of unicode-math to use unicode opentype math fonts. if you can find no other font ok, but..

@Noldorin it should have shown the internal command carrying the body as well

@UlrichDiez Your family sounds terrifyingly organized.

@barbarabeeton Whether the CDs are still good today I don't know. Back then I had burned the material on CDs because hard disks were small and expensive. Nowadays I have all my files, including everything I had on floppy disks and/or CD in the past, on a Raid level 1 system consisting of several 5TB hard disks.

@DavidCarlisle that is only in develop yet. But trying it out, I wonder if the output is right:

\NewDocumentCommand\testA{m}{abc}
\ShowCommand\testA

\NewExpandableDocumentCommand\testB{m}{abc}
\ShowCommand\testB

gives

> \testA=expandable document command:
  #1:m
->abc.
> \testB=expandable document command:
  #1:m
->abc.

shouldn't there be a difference between both? @PhelypeOleinik

@UlrikeFischer testA presumably shouldnt say expandable?

@UlrikeFischer oh I see, they differ if you use say mom as the argument spec so there is non-expandable parsing in the first case

@DavidCarlisle well it looks odd imho. With commands with an optional argument you get \testC=document command:.

@UlrikeFischer yes well if all the arguments are m then it ends up being expandable as it doesn't need to do much so...

@UlrikeFischer except it's protected so it doesn't expand.... hmm

@DavidCarlisle yes they both use \__cmd_start_expandable:nNNNNn, while the version with the optional argument use \__cmd_start:nNNnnn, so one can argue that \testA is a bit expandable, but from the user view it is a bit confusing isn't it?

@UlrikeFischer no I think it's wrong not confusing:

\NewDocumentCommand\testA{m}{abc}
\ShowCommand\testA

\NewExpandableDocumentCommand\testB{m}{abc}
\ShowCommand\testB

\typeout{[\testA{}]}

\typeout{[\testB{}]}

\stop

but it's Ok because:

LaTeX2e <2021-11-15> pre-release-1 (develop 2021-9-2 branch)
L3 programming layer <2021-08-27>
> \testA=expandable document command:
  #1:m
->abc.
> \testB=expandable document command:
  #1:m
->abc.
[\testA {}]
[abc]

@UlrikeFischer @PhelypeOleinik hard to argue testA expanded

@FaheemMitha Stone-age conditions prevail: In my own household, for example, there is only a simple subdivision into materials management, document management (archive and registry) and financial management. In order to know the exact current stock of consumables such as shoe polish, I have to open the shoe cabinet and spontaneously take inventory. So anarchy and chaos ;-))

@UlrichDiez So you don't scan your children's homework, then? :-)

@FaheemMitha No. I do not have children. Apart from that, in Germany children have limited legal capacity from the age of seven. For example, they can establish ownership. If I were to buy them exercise books, I would transfer ownership to them, so the question of how long they keep the crap would be their business, not mine. :-))

@JosephWright what was the policy on fixing typos on LL? PR or direct push?

@Skillmon PR

@JosephWright thx.

Does anyone know what this sentence at the bottom of the siunitx manual means? V 3.0.29, bottom of pg 23?

I can guess what it means. It probably means that zero padding is adding to the degree of stated precision. But the phrase before the colon has me thrown completely.

Specifically, no idea what is meant by `extend'. Can't find a match for that elsewhere in the manual.

(But this doesn't exactly seem suitable for a question on the site, either.)

At any rate, it looks like round-integer-decimal has changed its default from false to true in 3.0, if I'm understanding correctly.

@FaheemMitha I would guess my is a typo for may @JosephWright

@DavidCarlisle Ah, that makes sense.

"more digits (or figures)" presumably translates to the given degree of precision, which is round-precision. So perhaps something like: "The switch roundpad controls whether numbers are padded as necessary to the degree of precision given by round-precision. It defaults to true."

And maybe a word about it replacing round-integer-decimal.

@FaheemMitha This should be: "Rounding may ‘extend’ a short number to more digits (or figures): this is controlled by the switch round-pad, which is true as standard." In case the last digits are rounded to zeros, these zeros either can be omitted or can be append behind the last non-zero digit. It seems: Whether zeros are appended in this case is determined by round-pad=true/false.

@UlrichDiez With "may" it makes sense. Though as mentioned above, I would mention round-precision, because that determines how many zeros are appended.

@FaheemMitha But I don't know what this has to do with "pad". According to my English-German dictionary a "pad" is something like a cushion or upholstery.

@UlrichDiez Yes, that expression derives from the same usage. Pad means add something on. Just like cushioning. In this case, add on zeros.

@FaheemMitha Thus, digit sequences are regarded as already partially filled pads/cushions that have to be further filled to a given level. Nice analogy. Thanks for the clarification. :-)

@UlrichDiez You're welcome.

@FaheemMitha Yup

@barbarabeeton I know 8-inch floppy disks. From the military. At the time I did my military service, computer equipment was downright antique in many ways. For example, there was a radar control regiment that maintained a huge multi-story bunker with computers and large marine diesel engines as an emergency power supply, whose computing power was exceeded by my private 80586-laptop back then. :-)

Getting "Package array Warning: Redefining primitive column ! on input line 32."

from \newcolumntype{!}[1]{!{\vrule width #1}}

which is fine, I can use another character

but what does ! do in the first place??

"The exclamation mark ! is important: it tells LaTeX 'please really respect my choices here'." <- this was all I could find on this

@Avi texdoc array to see the array package manual, !{abc} is like @{abc} excapt that it does not suppress the usual tabcolsep spacing.

@DavidCarlisle Thanks!

@Avi but you presumably knew that as you used ! in the replacement so if it were not for the error your definition would be an infinite loop

@Avi that is presumably a description of ! in [!htp] option to figures but it is a really terrible description, actually it means the opposite of that it means "ignore the constraints I have set on figure placement" where is that quoted from?

fr.overleaf.com/learn/latex/Questions/…

ooh overleaf ^^^ @yo' @PauloCereda :-)

@DavidCarlisle did we break something? (someone?)

@yo' no I just asked Avi where chat.stackexchange.com/transcript/message/59127688#59127688 was quoted from (see last couple of lines above here:-) No one harmed though.

@DavidCarlisle that's what I saw, but it wasn't clear if something's wrong :)

@yo' well the description of ! is wrong but I'm not sure the end user would like the correct description (which would mean reading Frank's opus on how the float placement params work:-)

@DavidCarlisle well, I feel like that's it is a sort-of soft request

so why not "please" :-)

@yo' having seen the quote in context I'm not sure I want to offer any rewording. ! doesn't force anything what it does is ignore the numeric constraints on float placements from such as \topnum etc, but you can't really say that unless you have described all the parameters and you don't want to do that at the level of that page. so probably you want to ignore me...

@DavidCarlisle I'm not ignoring you, I'm absorbing what you have to say. (Very seriously now!) It's just that it doesn't mean there'll be any visible reaction ;-)

I'm tempted to add something like "When you use ! you take a lot of responsibility over the float placement and you can get very weird-looking results."

If I can just vent for a moment, amstex.sty, why oh why would you do \let\@xp=\expandafter? (Not to mention then proceeding to smear the resulting cs all over the other AMS classes). Grr. (As an aside, I'm in the midst of trying to see if I can track down a bug in align/align* and possibly other classes which is being a show-stopper for me.)

The short version being that you can't have align/align* inside of tabular or anything that's implemented with \halign.

@DonHosek well, why would you have a displayed math environment within horizontal contents, moreover a numbered one? You can use aligned

@DonHosek you can but you need {} around it (I keep meaning to fix that one day)

@DonHosek amsmath.sty I assume you mean, amstex.sty is just legacy and is what it is

@DonHosek why do you find \@xp problematic, I probably wouldn't have done it but it's just a csname, could as well be \tex_expandafter:D

\documentclass{article}

\usepackage{amsmath}
\begin{document}

\begin{tabular}{p{5cm}|p{5cm}}
 aaa & bbb\\
  {\begin{align}
    x&=x+y\\
    x+y&=z
  \end{align}}
&
  {\begin{align}
    z&=1\\
    w+z&=2
  \end{align}}
\end{tabular}
\end{document}

@yo' Just because you're inside an \halign does not mean that you're in horizontal mode. In my case, I'm putting source code and output side by side.

@DavidCarlisle My mistake I meant amsgen.sty—all the file nesting makes it a challenge to find where all the cryptic abbreviations are defined. I just did a grep for \@xp (and \DN@) in the amsmath directory in hopes of finding/verifying the definition. TeX code can be cryptic enough on its own without adding unnecessary obfuscation on top of it.

@DavidCarlisle Ah, that makes things easier since I can bake that into my existing macro code.

@DonHosek \show is your friend:-)

@DavidCarlisle And it turns out that putting an extra \bgroup…\egroup into my definition does not fix things but having the braces around aligned in the actual code does.

@DavidCarlisle And the difference is that \tex_expandafter: is at least providing a clear meaning to what's happening for contextual clarity. \@xp saves some typing at the expense of making the reader stop and wonder what it's doing there.

@DonHosek yes in code you probably need the magic \ifnum0=`{\fi} halign brace trick amsmath uses that mostly but I keep meaning to check why they don't work here

@DavidCarlisle What's the underlying cause for this?

I think in any event that it's long past time to rewrite the code that does the side-by-side source/output typesetting. It's 20 years old and I barely know what the hell it's doing myself anymore.

@DonHosek align is parsing ahead to grab the enviornment body and it needs to ensure that any & seen don't end the outer cell but it misses a case somewhere

@DavidCarlisle Hi, I write for this question: tex.stackexchange.com/questions/615100/longer-nleftrightarrow

@DavidCarlisle I have read in the recent documentation of mathjax that is possible to use \require{centernot} to have the negation to the center of the symbol. docs.mathjax.org/en/latest/input/tex/extensions/centernot.html

But for me the \require{centernot} not work.

@Sebastiano the main author of mathjax has already posted an answer

@DavidCarlisle Yes, I have seen and I have voted up :-)....and I not knew that Davide Cervone is the main author of mathjax.

I have seen now the profile :-)