@FaheemMitha You mentioned UNC. Do you mean University of North Carolina?

@LaTeXereXeTaL Indeed I do. Specifically, UNC-CH.

@PauloCereda the same here, no way that my machine is taking so little....

@JosephWright unicode.org/versions/Unicode14.0.0

@DavidCarlisle I'd seen

@DavidCarlisle gotta love X1FAC3. You should have added "beer" in "other keyword" ;-P

@JosephWright I looked yesterday but it was still saying draft, but it woul dhave been very early US time:-)

@Rmano how have you managed to discuss all your pregnant male friends all years without that emoji?

@DavidCarlisle yes, a real pain

@DavidCarlisle well at least one can now also discuss "pregnant woman with short hairs"

@yo' @PauloCereda in MDPI journal classes, there is an hidden \begin{paracol} inside the \begin{document} (no comment). Before the bibliography the author is expected to issue a lone \end{paracol} --- that confuses quite a bit the Overleaf editor (see screenshot below). is there a way --- some magic comment --- to tell the syntax highlighter that all is fine (with a very low value for "fine", I understand).

@Rmano Yep, let me dig up the Help page on that

@Rmano overleaf.com/learn/how-to/…

@yo' Wow, thanks, works perfectly.

@Rmano :)

you're welcome!

@Rmano apart from the magic comments to help the syntax highlighter a "portable" solution might be to put \def\zzz{\end{paracol} in the preamble and use \zzz

@DavidCarlisle maybe \def\whataabadideaputtingthatinthebegindocument{...} to give hints to the journal... ;-)

@Rmano Well that's a daft approach from the journal!

@JosephWright that's just one... look at tex.stackexchange.com/a/606032/38080 and marvel...

I think that what they want to achieve is not trivial, but the solution is suboptimal. Varying margin and marginpars would have been so much better.

@DavidCarlisle oh wow it seems I had the same idea and forgot tex.stackexchange.com/questions/586475/… ;-P

@Rmano there is no tex question that's not been asked before

@PabloGonzálezL have you ever tried flatpak? I must admit that's an interesting concept, way better than snap. I don't have it in my main workstation, but I have another machine with a couple of flatpaks installed and it's quite impressive!

@MarcelKrüger I've updated unicode-data to 14.0.0 but I'm not sure where emjoi-data.txt comes from or if it needs to be replaced: could you check?

@JosephWright I wonder when we can start using script and calligraphic as separate alphabets... w3c.github.io/xml-entities/script.html

@DavidCarlisle Hmm, good question

@JosephWright unicode.org/Public/14.0.0/ucd/emoji/emoji-data.txt

@DavidCarlisle Great

@DavidCarlisle, @MarcelKrüger I think we are good to go now

@JosephWright strangely enough I had a tab open in the UCD this morning:-)

@DavidCarlisle Odd that

@DavidCarlisle I see that data file is a bit naughty :(

@JosephWright github.com/w3c/xml-entities/pull/5

@DavidCarlisle Show-off

@JosephWright What's special about that one?

@MarcelKrüger It's not ASCII

@JosephWright having the emoji inline?

@DavidCarlisle Yes: I'm surprised that was allowed

@JosephWright unicode consortium can't really be blamed for thinking that unicode is text

@DavidCarlisle Except one can't see the data without the fonts ;)

@DavidCarlisle Every other file is pure ASCII, so it stands out

@DavidCarlisle Especially since it's officially only in the comments.

@JosephWright true but emoji would be dull without the pictures

@JosephWright Not included in unicode-data, but e.g. NormalizationTest.txt isn't ASCII either.

@MarcelKrüger If we don't use it, I don't care :)

looks better with the emoji ^ :-)

@JosephWright It's used in lua-uni-algos...

@DavidCarlisle your lunch menu?

@MarcelKrüger Why's in not in unicode-data then?

@samcarter I find Japanese post offices a bit crunchy

@JosephWright because the format there was set in stone in pre-history, they daren't even add a comment with the version number (it has no comments at all)

@DavidCarlisle Neither does UnicodeData.txt: we simply hard-code that into the README or loader or whatever

@JosephWright sorry I answered the wrong question (I had same thought that we/you should include it if it's being used for luatex)

@JosephWright It's basically a list of testcases for normalization, so it's not really useful for anything else and would just be 2.5MB of pointless data (since it's only needed during testing). Therefore I considered it to be a part of the test suite of that package.

@MarcelKrüger Ah, right: that's OK then

@MarcelKrüger You happy with my update? We can get this out today I hope

@JosephWright Looks good. Thanks :)

@MarcelKrüger Cool - to CTAN

@FaheemMitha Just wondered. I was an undergrad there '85-'89. Small world.

@PauloCereda The truth is that I tried it a few years ago, but it didn't work as I expected, of course I haven't tried it since.

@Rmano This is my output for systemd-analyze blame

@Rmano I don't know ubuntu, but in fedora there are always daemons/services left that I don't use and they just slow down the startup.

@LaTeXereXeTaL Ah. That's a bit before my time. But it's not like a super-tiny place, or anything.

@PabloGonzálezL it's been improved greatly. :)

@FaheemMitha Indeed.

@LaTeXereXeTaL Have you been back since?

@PabloGonzálezL In my home system, which is the slowest one (and that I have to reformat one of this days), I have

⌂4.48 [romano:~] % systemd-analyze blame | head -20
1min 27.121s vboxdrv.service
     50.138s udisks2.service
     40.185s man-db.service
     35.287s plymouth-quit-wait.service
     32.009s docker.service
     19.361s snapd.service
     16.969s containerd.service
     14.748s dev-sda3.device
     14.167s systemd-journal-flush.service
     12.565s accounts-daemon.service
     10.807s networkd-dispatcher.service
     10.388s NetworkManager-wait-online.service
      8.762s dev-loop11.device
      8.614s dev-loop15.device

...and then a lot of things I do not know what they are doing.

I don't think that ` vboxdrv.service` is a problem, it should run well in background...

^^^ Stupid question, but these code blocks from the recent mails on the texlive list are so gorgeous. Is this a gmail thing or are there any mail programs which format them like this?

@samcarter the OP posted it as html mail using highlightjs.org/usage

@DavidCarlisle Thanks!

@samcarter doesn't answer your question of whetehr any mail client is making that easy though:-)

@DavidCarlisle no, but it gives me a search term to put into duckduckgo :)

@DavidCarlisle oh no HTML mail

@PauloCereda <b>dinner</b>

@DavidCarlisle <marquee>oh no</marquee>

@PauloCereda <blink>

@DavidCarlisle ooh

@Diaa didn't Joseph answer that in comments? (as in answering the "why" it's not supported in v3)

@DavidCarlisle He talked about the retaining of subdivision; here I am wondering why I don't get the \cdot of inner-product printed as egreg found in his output.

@Diaa oh, OK. anyway I'll leave this to Joseph:-)

I am wondering can someone please help with this question: tex.stackexchange.com/questions/615418/…

\documentclass{article}
\usepackage[utf8]{inputenc}

\usepackage[a4paper, total={6in, 8in}]{geometry}
\usepackage{algorithm}
\usepackage{algpseudocode}
\usepackage{amsmath,amssymb}
\usepackage{amsthm}

\begin{document}

{
 \floatname{algorithm}{Code}
\begin{algorithm}[h]
    \caption{Rule evaluator for classical attribute grammar}\label{alg:ag-eval}
    \begin{algorithmic}
    \Procedure{\texttt{AG\_EVAL}}{$r, Val$}
        \If{$r \equiv v_0 \texttt{=} g( v_1, \dots, v_n)$}
            \State $Val(v_0) \gets g( Val(v_1), \dots, Val(v_n))$

@Node.JS ^^ try that

It worked!

Thank you so much

@Node.JS You're welcome!

@DavidCarlisle ^^ Javascript alert :)

Should I delete that question?

@Node.JS You could write a self-answer so future users with the same question can find it

@Diaa I think your method to pass an optional argument to the \unit command doesn't work. Everything is fine if I set up the options before the tabular. Where do you got this code from?

@DavidCarlisle -- Hooray! Calligraphic!

@barbarabeeton one day, now we need font support....

@PauloCereda sensible language: no jvm

@DavidCarlisle -- Yeah, yeah. Speaking of font support, am I the only person around here who is seriously bugged by this: tex.stackexchange.com/q/36907 ? While I think that there's too much of a gap between closing quotes and a following comma or period, I think a full overlap is going too far -- it confuses the situation. I think a kern between closing quote and the punctuation, comparable to that between F or P and punctuation would be preferable.

@DavidCarlisle oi

@barbarabeeton did you mean that link?

@DavidCarlisle -- Er, no. I meant this one: tex.stackexchange.com/q/369077

@barbarabeeton ah I'm used to obscure comma compaints from you, but that one was making me scratch my head...

@samcarter ooh twitter.com/hyxpk/status/1438170758023286786

@PauloCereda ohh

Does anyone know what the exact equivalent of \expanded{ ... } is in expl3?

@Noldorin well depends what you mean \tex_expanded:D but basically the e argument type.

@Noldorin which means \use:e{...} is probably the closest as that does \expanded expansion of the argument then \use:n which just uses it

@DavidCarlisle Ah right, makes sense. How does \tex_expanded differ? I don't see it documented.

@Noldorin \tex_expanded:D is \expanded (\tex_expanded isn't defined)

@Noldorin all tex primitives have expl3 internal names tex_something:D where D means do not use and something is usually the original name but normalised so for example strcmp even though it's \pdfstrcmp in pdftex and \strcmp in xetex

yes, I elided the :D just for brevity, sorry.

@Noldorin see texdoc source3 page 324 (l3names implementation)

triple backtick doesn't work here:-)

@DavidCarlisle Curiously, the above works fine (or with \expanded in place of \tex_expanded:D of course), but not with \use:e

@DavidCarlisle yeah, annoyingly...

\exp_args:NVne \RenewDocumentCommand \citecmd { s D<>{} } {
\exp_not:n { \preparenextcite{##2} }
\exp_not:N \exp_after:wN \oldcitecmd \exp_not:n { \tex_expanded:D { \IfValueTF{##1}{*}{} } }
}

@Noldorin like that ^

Ah, how'd you do that?

@Noldorin Indent with four spaces.

oh right. old-school Markdown. cheers.

@Noldorin paste it in then a "fixed format" button appears out of nowhere (which actually just indents as Marcel says)

anyway, any idea idea why \tex-expanded:D works here but not \use:e?

@Noldorin \use:e needs two steps of expansion, therefore it breaks when you only expand it once..

@Noldorin a guiding aim of the expl3 design was that functions would just have argument specs of :nneo to control expansion and the days of interleaving \expandafter and \noexpand would be gone. Something very suspect in needing \exp_not:n

it's a second-order macro. I've tried every possible way to do this, and this is working at least, so I'm not going to tinker with it more. it's fairly readable too.

Oh, it just occurred to me that \exp_not:N \exp_after:wn \oldcitecmd \exp_not:n { { \IfValueTF{##1}{*}{} } } should work... although it doesn't

loads of errors still.

@Noldorin Instead of \exp_after:wN \something \tex_expanded:D{...}, you would just use \exp_last_unbraced:Ne \something {...} in expl3.

@MarcelKrüger Ah yes, fair point, thanks. I wrote that line just a week or so ago before I learnt about exp_args!

Oh, let me try your updated comment...

(And it seems the chat only recognises code blocks indented by spaces, not tab.)

@PauloCereda this page doesn't exist

@UlrikeFischer hold on, the artist erased it. Let me find another link...

@UlrikeFischer twitter.com/hyxpk/status/1438171237767737359

@UlrikeFischer From Joseph's previous answer tex.stackexchange.com/a/600854/2288

@Diaa Yes I found it. I already added a comment, I think there is something wrong. The options are ignored with this code.

@UlrikeFischer I would be grateful if you could give a new working answer or edit the accepted one.

@Diaa I don't know an answer. You will have to ask @JosephWright.

@FaheemMitha Yes many times, but not recently. I used to think I would retire to Chapel Hill but it's just too expensive to do so now.

@LaTeXereXeTaL Ah, a fan of the Triangle. Me too.

@Diaa I didn't include passing on the options: really I didn't think this was useful for a unit column

@FaheemMitha I used to make special trips just to visit Bull's Head Bookstore and the used book shops on Franklin Street.

@LaTeXereXeTaL Yes. I tell people sometimes that in my head I think of Chapel Hill as home.

@FaheemMitha I need to get back there for a visit soon.

@LaTeXereXeTaL I'm a fan too. My first job was at UNC. Crook's Corners was my favourite restaurant.

@AlanMunn -- I've gotten a report from a TeX friend that sending the clause "fruit flies like a banana" through Google translate to numerous non-English languages either treat "flies" as a verb (Italian, Danish, Norwegian, Swedish) or fail to provide a verb at all (French, German, Spanish). i haven't tried it myself, but intend to, with Russian and Greek. Isn't this a rather odd failure?

@barbarabeeton Why is it odd? If the verb "flies" is more common than "fruit flies" it automatically defaults to the more common one. It can't do context if it doesn't have precedence. Obviously there wouldn't be a verb if it translates "fruit fly" correctly but quite possibly skips the verb if there are enough quite similar contexts that it can find. It functions on matching, not actual translation.

@AlanMunn Oh wow I had no idea. I liked Carolina Coffee Shop.

@Plergux -- Not saying that the original expression isn't odd, but for a translation service that is advertised as being competent, this seems an unfortunate lapse. (I'm probably just expecting too much, but I'm still "recovering" from editing several hour-long interviews for which I was provided auto-transcriptions. The time involved ultimately amounted to just under 30 hours each. I really expect better.)

@AlanMunn How long did you live in Chapel Hill?

@barbarabeeton FWIW, I agree with @Plergux. I think that's a tricky sentence for a bunch of heuristics to translate.

That's why us humans have the upper hand over silicon. For now.

@FaheemMitha -- Yeah. Humans aren't perfect either. At a conference in Paris which had simultaneous translation, "deux colonnes" was translated by the human as "two doves" ("deux colombes"). Knowledge of subject matter is obviously important. But I've never seen an adequate warning on an online translation service.

@barbarabeeton Humans are obviously extremely imperfect. But they tend to be better than a machine at things like translation. For now, at least. Though if a program got too good at translation, that would be worrying.

@barbarabeeton Do you expect Google to have a health warning, like "we are not responsible for the consequences of an erroneous translation."? :-)

@FaheemMitha -- I have a rudimentary understanding of a lot of languages. (I refer to is as "restaurant competence" -- there are few places where I will find it impossible to obtain a satisfying meal.) But if one is writing documentation, and wants it to be understood even by readers who may have to resort to automatic translation, the writer has to be careful. When crafting web pages, I would present them to Google translate to make sure the result was at least plausible.

@barbarabeeton That sounds like a bit of a burden to place on potentially deathless prose.

Currently upgrading my Debian system. Fingers crossed. The last upgrade was quite painful. I had to debug breakages using my phone.

@FaheemMitha -- I wouldn't want to claim my prose is deathless. I do try hard, though, to make sure it's accurate and reasonably understandable.

@barbarabeeton Those are most admirable aims.

I think people are two dependent on Google Translate, though. Aren't there better alternatives?

@FaheemMitha -- Actually, if one digs into the bolierplate of "terms and conditions", that probably is there. "As is" is pretty much ubiquitous these days. "The contents of this container may be hot" on a coffee cup.

@barbarabeeton I make it a matter of policy not to depend on Google. So I've never really felt the necessity to read their fine print.

@FaheemMitha -- "two" ==> "too". (Apologies for pedantry.) I'll try a couple of other options.

@barbarabeeton Yes, I noticed, but it was too late to edit. Just outside the 2 min mark.

I don't normally make homophone errors, but it's late here. And no need to apologise. Feel free to point out any errors I make.

@FaheemMitha DeepL seems to be a bit better, but unfortunately there are users who don't understand that the risk of mistranslations grows exponential if they feed some obscure figure of speeches from their native language into the machine instead of keeping things as simple as possible

@samcarter Perhaps find an actual human to translate. Though last I checked, paid services were unreasonably expensive.

@FaheemMitha ask DavidCarlisle to do the translation - I bet you can pay in crispy ducks :)

@samcarter I suspect David outsources.

I hear a lot about ducks here. Sometimes it makes me hungry.

I wonder why my system has so many packages on it. I don't know what most of them are for.

I'm researching fridges that don't suck. Does anyone have any thoughts? Regular residential/home fridges here are kinda crap.

@FaheemMitha move to the south pole and you won't need one

@samcarter That's not particularly helpful.

Also I hear the South Pole may be melting soon.

@FaheemMitha sorry :)

@FaheemMitha yes, but worst case you will end up sitting on a pile of stones. If the same happens when you are on the north pole, things could become a bit more problematic

@samcarter Well, I'm unlikely to go to either place, so that's probably not something I need to worry about. A fridge, on the other hand...

@FaheemMitha twitter.com/nick_craver/status/1335356076221489154?lang=en

Apparently Samsung has something called the Twin Cooling System, which sounds vaguely promising.

@samcarter I don't get all the Samsung hate. I doubt they're any worse than the competition.

@FaheemMitha Not sure how the situation is for other markets, but in Germany most fridges are anyway made at the same factory and the different companies then just stick their logo on it.

@samcarter Yes, I think it's like that in most places.

@FaheemMitha Just a year, unfortunately. It was a one year visiting position.

@AlanMunn Oh. So you liked it there?

@FaheemMitha Yes indeed. I liked the university, and being relatively close to both mountains and ocean.

@AlanMunn I liked the town too. Very different from most other places. Even in NC.

@barbarabeeton The sentence is syntactically ambiguous, so without context, what should any automatic translation system do? As @Plergux says, if the relative frequencies of the various words vary and there's no supporting context, it's going to make a best guess. It's interesting if the targets end up different in different languages though.

@barbarabeeton Take a more prosaic example like "Sara watched the man with the telescope". Without context, either meaning is reasonable. Now consider translating to a language that has an instrumental case for the VP attachment and not for the NP attachment. Without any other biases and no context, it will necessarily be wrong half the time.

@DavidCarlisle I wonder where the fbox problems belong. It is a xcolor bug as it redefines fbox, or a latex2e problem as is should adapt fbox?

@UlrikeFischer er Ive been away, fbox?

@AlanMunn -- Okay. With a little bit of context, "melanogaster fruit flies like a banana" comes out in French "les mouches des fruits melanogaster comme une banane". I guess some expressions are just intractable. And that what I consider relevant context isn't recognized as such by that software. That's bad enough with a written original; now consider how much worse it could be if the original is spoken.

@samcarter Ich habe absolutes Vertrauen in Google

@UlrikeFischer ah found the issue. Ideally neither should be redefining fbox

@DavidCarlisle -- My German is a bit shaky, but I think "Vertrauen" isn't the best word in this situation. @UlrikeFischer -- Is there something better?

@barbarabeeton you can't argue with an infallible oracle

@AlanMunn -- Okay. I've played around a bit more. I can't explain this result: "Bees like ice cream" ==> "les abeilles aiment la crème glacée" but "flies like ice cream" ==> "mouches comme la crème glacée", "aiment" is produced with all sorts of other insects; only "flies" has a problem. "these flies" solves it. Hmmmph.

The plot thickens: "flies like ice cream" ==> "mouches comme la crème glacée" but add a period "flies like ice cream." ==> "vole comme de la crème glacée." I give up!

@barbarabeeton no, it is a quite accurate translation.

@UlrikeFischer -- Thanks. As I said, my German is really rusty.

@DavidCarlisle yes, but I assume that xcolor tried to correct something, so the question is if we can/should move that to latex2e?

@UlrikeFischer yes of course I read the doc:

% The main difference to \LaTeX's implementation is that box construction and frame drawing are split into separate operations, such that the frame is drawn \emph{after} the box contents has been constructed.
% This ensures that the frame is always on top of the box.
% \People{Donald}{Arseneau} improved speed as well as memory requirements of this approach.
% Furthermore, a new macro is introduced:

but the resulting definition is \protect robust but hyperef kills it in the bookmark handling, etoolbox \robustify makes it etex \protected which seems safer with hyperref. Shouldn't \protect be enough for the bookmark code?

@DavidCarlisle Same as xbox implementation then

@DavidCarlisle well it isn't safe in an edef:

\documentclass{report}
\usepackage{xcolor}
\begin{document}
\edef\blub{\fbox{World}}

@UlrikeFischer you're not supposed to use edef with latex:-)

@DavidCarlisle tell that hyperref ...

Where is \cmditem defined?

@UlrikeFischer which edef there are hundreds:-)

@DavidCarlisle I guess this one in \pdfstringdef explodes here: \xdef#1{#1\HyPsd@empty}%

@UlrikeFischer \pdfstringdef is such an unreasonable definition

@UlrikeFischer simplest might be to make xcolor use \protected, since it is redefining things anyway

@DavidCarlisle well xcolor doesn't redefine \fbox as far as I can see. So if we want to make if protected it should probably be done in latex? It would also be nice to be able to close the gnats bug, but I don't think that one can simply copy over the xcolor code. It seems to contain color stuff too.

@UlrikeFischer yes well xcolor redefines \@frameb@x which is the main bit (shared with \framebox)

@DavidCarlisle making \@frameb@x protected avoids the error too, but at the cost of much more Package hyperref Warning: Token not allowed in a PDF string warnings. When \fbox is robustify one gets only one:

Package hyperref Warning: Token not allowed in a PDF string (Unicode):
(hyperref)                removing `\fbox' on input line 20.

@UlrikeFischer so can we pdfstringdef fbox by default

meanwhile my husband sent me a citation: "He was a gentleman on whom I built an absolute trust (Act I, scene 4)"

@DavidCarlisle ? what do you mean?

@UlrikeFischer Du solltest Dante lesen, nicht Shakespeare

@UlrikeFischer \protected or \protect just stopis it blowing up but can't we stop that warning by allowing \fbox in pdfstrings (doing nothing) like the existing \let\mbox\@empty in pdfstringdef

@DavidCarlisle well yes one can always do it, but do we want add all sort of commands like \fbox \framebox or \duck to pdfstringdef only because once in a year someone uses them in a section?

@DavidCarlisle "Dante and Shakespeare divide the world between them“

@UlrikeFischer no as I say it's a mess already I suppose just silently dropping all unknown commands is bad as well ... so xcolor in addition to changing the internal @frame@box make the top level \protected\def\fbox{\protect\fbox } ?

@DavidCarlisle or latex. Or is there any reason why it should be not protected generally?

@UlrikeFischer probably not but why just fbox if we start theres \parbox and \makebox and ... do we make them all etex \protected (and make @JosephWright happy?)

@UlrikeFischer I'm tempted to say that it's hyperref that should support classic latex \protect better, but I don't want to look tonight

@UlrikeFischer \fbox is a robust command it shouldn't be randomly expanding and dying pre-expanding \@makeother\<%` catcode normalisation