@DavidCarlisle Given how the publishers are moving forward, I think that's the unfortunate way (because, we will miss the native font in the figures :/)

@CarLaTeX I'am in tilt :-(With the mouse scroll I didn't go down and I didn't see your answer. For the umpteenth time I was wrong to see. Have you seen the effect of school in Italy?:-)

@DavidCarlisle Have a nice, peaceful day. My best regards.

@Sebastiano Don't worry :)

@DavidCarlisle Sorry I meant the later.

@Raaja I still don't understand your font comment but whatever, it doesn't matter if I follow or not:-)

@DavidCarlisle, @egreg, @UlrikeFischer I'm going to do a release of l3kernel so XeTeX can make active chars (engine permitting): I missed that a couple of days ago

@JosephWright good plan

@JosephWright speaking of which why is ! Bad character code (119808) a bad code? (unicode-math gh)

@DavidCarlisle Don't ask me ;)

@DavidCarlisle Oh, that ....

@DavidCarlisle I guess I can have a look

@JosephWright The CTAN team will be delighted. :-)

@DavidCarlisle I guess that XeTeX didn't change the original interpretation of \char in math mode, so only allowing one byte characters.

@DavidCarlisle Not one byte, but something bad happens at 65536

@egreg yes xetex uses utf16 in lots of places so the internal representation switches to a pair of surrogate points at that point rather then just being a simple mapping between the unicode number and the string representation

isn't this columns of images rather than rows?

if this is rows, how to do columns instead?

@baxx rows it seems (although as posted the OP had two images in one row and then a third below)

how to do rows like matrix rows instead? So one above the other, rather than to the right

@baxx images are placed by the same code that places letters, they come side by side by default unless there is a blank line which puts them in separate paragraphs

this might work actually

@baxx if the images are the same size it's far more markup than needed, just four \includegraphics with a space between each would be enough.

@DavidCarlisle ah, that's correct. thanks

@CarLaTeX Could you please fight a bit with your special friend? (I am just kidding!) Now I am his main target, see e.g. here. He always does the same thing: attacking others. And when he is proven wrong, he would blame it either on the language barrier or come up with another hilarious reason why he can do that. It is never his fault.

@DavidCarlisle And apparently this is not very well accepted by the math engine.

@egreg blame Unicode 3 for stepping past `"FFFF" :-)

@egreg still, we know who last touched the xetex source, so we can blame him.

@DavidCarlisle What I meant was, publishers like Elsevier, Springer (they are just examples) have their own propreitary fonts. So, when we compile the eps to pdf and give them, the font and the font-size at which we compiled will not match with their font and its size per se. But, on the other hand, if we give them the source .tex with text substitutions using \psfrag for example, they mess it completely :D

Thats what I meant with the fonts (it was too early in the morning, to think precisely) :D

@Raaja ah safest would probably to use latex rather than pdflatex for psfrag, but not sure if elsevier would accept that these days

@DavidCarlisle They (not all the journals) are, for the old-time sake, supporting LaTeX :D But some of the well-established journals do provide the "full" support.

For over the past few years, whenever I send them the source files, I am explicitly giving them a "PLEASE_read_me.txt" file.

Which files do define how pdflatex handles eps files?

@skillmon I dont get you.

@DavidCarlisle In a latest event, one of my manuscript, underwent more number of revisions between the copy-editors and me than it went with the review flow.

@Skillmon pdftex.def mostly (but for psfrag it's more complicated)

@Raaja nowadays pdfLaTeX autoconverts eps files to pdf and includes them. This works without shell-escape, so I wanted to take a look at it to evaluate whether a similar route could be taken to handle pngs (because they are soooo slow)

@marmot Just ignore him, I think he's only a know-it-all :)

@DavidCarlisle it wasn't related to @Raaja, just looked as such by accident.

@skillmon Now my confusion got cleared.

@Skillmon there was discussion of a new png to pdf convertor making more highly optimised pdf bitmaps the other day on texlive list

@DavidCarlisle I don't see texlive lists, just the latex mailing list. What a coincidence. So I don't have to dig into it myself and can expect them to do it for the greater good?

S-mode woo

Are there any known incompatibilities with pdfpages and other things? Maybe totcount? I tried to insert a page into an exam this morning and it totally failed on the first line that included a counter that was being added by totcount.)

@AlanMunn I think it's mostly safe. :)

I use it almost all the time.

@AlanMunn oh a package conflict...

@PauloCereda Yes I do too. I've never encountered problems before.

@DavidCarlisle I saw that you are in LaTeX3 core team. May I ask you, when can we expect its release :D (atleast a update would be nice)

@PauloCereda Yep, here's a MWE:

\documentclass{article}
\usepackage{totcount}
\usepackage{pdfpages}
\newcounter{points}
\newcommand{\points}[1]{\addtocounter{points}{#1}[#1 points]}
\regtotcounter{points}
\begin{document}
\section{Foo \points{10}}
\includepdf{\jobname.pdf}
\end{document}

@Raaja the standard plan (which we have followed since about 1992) is to have it done in the next few years

@Skillmon I'll dig out a reference

@AlanMunn indeed, I get an error too.

@DavidCarlisle Thanks for the update, this would satisy me for a few-years at best :D

@AlanMunn the incompability is between totcount and calc.

@UlrikeFischer Thanks. Since calc has been around forever, I assume it's really totcount's problem? Is there a workaround?

@Raaja expl3 is updated frequently.

@skillmon Thanks I will note it down.

@Skillmon it was texhax, not texlive: tug.org/pipermail/texhax/2018-November/023367.html

@AlanMunn how can \includepdf{\jobname.pdf} work?

@CarLaTeX Yes. And it is always someone else who is wrong, or unfortunate circumstance, but it is never him to blame. (The moderators seem to like him, though.)

@DavidCarlisle we could load expl3 and then claim that we started ...

@UlrikeFischer ooh

@DavidCarlisle Compile once without the \includepdf then run it again. :)

@DavidCarlisle Seemed simpler than having a random .pdf file to include.

@AlanMunn but even then you seem to be taking a lot of faith in the timing that this run hasn't yet opened the file for writing?

@AlanMunn but \includepdf{example-image} seems a lot safer.

@DavidCarlisle Oh, maybe. Anyway, substitute \jobname with any pdf you have.

@UlrikeFischer True. I forgot about example-image.

@AlanMunn isn't that just the standard fragile command in moving argument error?

\documentclass{article}
\usepackage{totcount}
\usepackage{pdfpages}
\newcounter{points}
\newcommand{\points}[1]{\addtocounter{points}{#1}[#1 points]}
\regtotcounter{points}
\begin{document}
\section{Foo \protect\points{10}}
\includepdf{example-image.pdf}
\end{document}

@AlanMunn although including addtocounter in a heading seems brave anyway, if you use \tableofcontents it'll be double-incremented.

@DavidCarlisle Yeah, that's not an issue really since this is just something I use for calculating points in an exam. As for the \protect issue, you're right it seems to work, but the \protect isn't needed if pdfpages isn't being used.

@DavidCarlisle thanks for digging. Though it doesn't seem like they're intending to create something automated, at least not in that particular message.

@AlanMunn @UlrikeFischer will no doubt be able to tell you that calc package documents that it makes arithmetic fragile as it has to parse the infix expressions and makes multiple internal assignments along the way.

@Skillmon it would be simple to switch pdftex.def to using that but it would have to be available on all platforms, and be accepted to be allowed in the restricted shell list (which mean someone has to do a code review thorough enough to convince Karl et al that it is safe to include)

@AlanMunn and @JosephWright would point you at texdev.net/2018/12/09/floating-point-calculations-in-latex :-)

@DavidCarlisle ;-). I can at least tell that the documentation says that \addtocounter is fragile (and I advance such counters outside titles so actually never had a problem with it).

Any native English speakers? We're having an argument, which is more corerct: "Monday to Thursday" or 'Monday – Thursday', I do not like the latter in writing.

@UlrikeFischer phew I'm glad you could confirm (I din't of course check the doc first:-)

@daleif the first is better

@DavidCarlisle No @JosephWright would be more helpful than that. :)

@daleif It would probably depend on the context. If it's part of a sentence, then I would agree with you an use 'to' but in other contexts (e.g. a table or the like) I would probably use the hyphen.

@DavidCarlisle oh

@marmot The moderators bore even worse users :)

@AlanMunn Here the sentence is a bullet point: "* Lunch Monday to Thursday"

@daleif Then I might go with the hyphen there. I don't know if @DavidCarlisle would agree.

@CarLaTeX I am actually not blaming them. They are not police officers. It is just sad that certain people use this site to exhibit unfortunate behavior patterns.

@DavidCarlisle Good idea

@DavidCarlisle I know :( But I don't think I'm a qualified reviewer.

@CarLaTeX The moderators try to be fair all round: this is not always easy, and it's not always obvious to others when action has been taken

@DavidCarlisle we could decide to use \Uchar in \symbol and you could make ctan happy by uploading everything anew.

@UlrikeFischer :)

@JosephWright ;-). Did you see that the development luatex in w32tex has now a luatex.dll and a luatex52.dll? and the experimental luatex has only a luatex.dll and lua53.dll? It is drifting over to lua 53 ...

@daleif -- i claim that "monday to thursday" (which i prefer to the " - " form) doesn't necessarily include thursday. to be sure it does, "monday through thursday" is more precise.

@barbarabeeton did you like our TUG advertisement?

@JosephWright Yes, the lack of trasparency sometime creates puzzlement, but I understand there are privacy reason for that

@barbarabeeton Really? Is there any real usage of it with that meaning?

@CarLaTeX -- the one in the extravaganza? yes, i did. and i liked the whole presentation, including the music choices. (although maybe the promenade from "pictures at an exhibition" would have been appropriate for the arts segment.) you guys did a super job!

@AlanMunn -- "office hours are monday through thursday, from 9 a.m. to 5 p.m."

@barbarabeeton No, I meant the one with 'to' not including Thursday.

@barbarabeeton Or do, you mean the time version in your example? That's an interesting difference.

@AlanMunn -- i think i'm just being pedantic again. but as for the contrast between "through" and "to", yes, that's why i used that example.

@barbarabeeton That seems more to be a difference between what a time denotes and what a day denotes. A time is a point (roughly) while a day is the name of an interval.

So I would say there's no ambiguity in the M to Th. example.

@barbarabeeton we thought about pictures at an exhibition too, but it actually started only as a mona lisa and then the music idea stayed.

@AlanMunn -- you're most likely right about that. i've just been exposed too much to situations in which texts were going to be translated and the multiple versions had to have the exact same meaning. so i tend to get overly pedantic. (and perversely "misunderstand" directions whenever possible -- if i can predict what someone else might get wrong, there's a better chance to fix a problem sooner than later.)

@UlrikeFischer You could also have done Pearls on a Chain, Olivia Newton John; Andy Warhol, David Bowie; Scream and Holler, Van Morrison; (I've forgotten the other painting).

@barbarabeeton Yes, I can definitely see that. I agree that the 'through' version is unambiguous.

@barbarabeeton Great music suggestion for the sequel next year :) [there are plenty of nice pictures left]

@AlanMunn ooh greased lightning

@PauloCereda, @AlanMunn, @barbarabeeton, @UlrikeFischer, @Skillmon @egreg @marmot, @PhelypeOleinik good afternoon and good work.

@Sebastiano Good afternoon. Can you do me a favour? If you say hello, don't ping everybody. Just say hello, and those who are currently viewing will notice, the others won't get pinged. I don't want to sound mean, I just get pinged every time, thinking there is something important or stuff, stop whatever I'm doing to look.

@Skillmon Ok. sorry

A quick question: Is expl3-generic.tex supposed to work when input to ConTeXt?

I am so tempted to write to TL list saying that "Cannot wait to have TeX Live on my Xbox One", but they won't get the joke or the reference. :)

@TeXnician Yes, hopefully

@JosephWright Well, I get a "fatal error: kpse library is accessed (key: find_file)".

\input expl3-generic.tex

\starttext
Test
\stoptext

@TeXnician Ah

@TeXnician Hmm, I'll need to look at that I guess

@TeXnician I think Hans must have got more aggressive about file lookup

@JosephWright May be, it is the first time I tried to explicitly include it. But as this even happens with the LuaTeX and ConTeXt in TL2018 it can't be that recent…

@TeXnician Been a while since we tested it: it works n TL'16

@JosephWright not only with this. I checked yesterday evening the newest context fontloader and the lualibs files and I think he removed the unicode-library (or hides it) and nothings works.

@UlrikeFischer Hmm

@UlrikeFischer What's the ConTeXt version of kpse.find_file()?

@JosephWright not sure resolvers.findfile probably.

@UlrikeFischer Perhaps I should ask on the ConTeXt list ...

@JosephWright what do you need it for?

@UlrikeFischer See above: it's used by the pdfTeX-primitive-emulation code in expl3, and that fails now on ConTeXt

@JosephWright Do you think it is worth to try to get expl3 working again in context?

@UlrikeFischer Personally, I would really love to see it working again, because some TikZ libraries (e.g. tikzmark when adapted) rely on expl3…

@UlrikeFischer Possibly not ...

@TeXnician I'll try :)

@JosephWright Thanks :)

@TeXnician well perhaps you should then complain to context and ask them to adjust to expl3 instead of expecting constantly us to keep up with the context changes.

@UlrikeFischer What is slightly odd is that I don't think there's any Lua in the problematic block ... this is odd

@UlrikeFischer That might be a plan, but didn't they decide to put forward a "stable" version of LuaTeX (I guess including the Lua interfaces) for TL19?

@TeXnician ? sure but why should this be relevant?

@UlrikeFischer Maybe this means the need for constant adjustments would vanish :)

@JosephWright well the error is when the expl3.lua is loaded. And this is already enough to trigger it:

\directlua{local kpse    = kpse
           local kpse_find  = kpse.find_file}
\starttext
Test
\stoptext

@Sebastiano don't understand me wrong, it is fine to ping somebody sometimes, just don't ping everybody always. For instance, @PauloCereda and I do ping each other sometimes to say "hi" in a ducky/rabbity (does this exist) manner.

@Skillmon ooh a ping... I should pong! :)

@UlrikeFischer Yup, I got that

@UlrikeFischer I've nearly sorted that issue, but there is a second one ... should be solvable too (TeX layer)

@JosephWright what did you do? "if context then use resolvers.findfile"?

@PauloCereda got that pong. Now I feel challanged: pong-2.com

@Skillmon ooh

@TeXnician the problem is not so much a stable luatex but a stable context.

@UlrikeFischer More-or-less what I've done

@UlrikeFischer, @TeXnician I've sorted a second issue, more-or-less: will be working again soon. I did a release today, so one for CTAN tomorrow ...

@PauloCereda controls don't work in online multiplayer :(

@PauloCereda and I can't choose with whom to play.

@Skillmon oh no

@JosephWright what was the second issue?

@JosephWright Thanks.

@UlrikeFischer \newread in ConTeXt now does a check before altering definition ...

@UlrikeFischer, @TeXnician I've just checked, I have things working again :)

@UlrikeFischer I need \cs_undefine:N #1 before \newread #, but that's a bit more tricky as we also have to cover plain where \newread is \outer ...

@DavidCarlisle \outer ^^^

@UlrikeFischer I'll release later today or probably tomorrow now (I'll likely move \tl_range:nnn to stable first)

@JosephWright do them a favour and redefine it without \outer:-)

@DavidCarlisle :)