@DavidCarlisle Looking at some test stuff: is tlb-utex-004 actually giving the right result? Shouldn't a TUG-based test be using lmr not cmr? In my stdmain branch I'm getting a failure, but I think it's because the format being used is now right and was wrong

@DavidCarlisle I'm going to push a checkin and would appreciate a once-over

@JosephWright the 2e tests fonttext.cfg makes xetex and luatex default to cmr to reduce diffs (not clear this was the best idea) is your new setup building a normal format?

short question: anyone know how to apply xurl line breaks to \href? We made a \DOI macro using \href but would rather have the line breaks from xurl

\href{URL}{DOI:~\url{text}} seems to work, but I worry that this attempts to generate a hyperlink twice (from \href and from \url)

Ahh of course \nolinkurl .....

@DavidCarlisle Will explain on team list :)

@DavidCarlisle Does mail make sense?

@JosephWright was in a meeting, let me look...

@DavidCarlisle oi

@PauloCereda ?

@DavidCarlisle what did I do this time? :)

@DavidCarlisle weirdness near São Paulo

@PauloCereda and did anyone mention you?

@DavidCarlisle hmmmmmm

@PauloCereda remember, I am not mean

@DavidCarlisle ooh

^^^ preparing for holiday ...

@DavidCarlisle I've read your comment and Heiko's warning and deleted my answer. Thanks!

@AlanMunn sorry:-)

@DavidCarlisle No problem at all. I was skeptical that it would work myself, and was pretty much expecting some dire consequence I hadn't understood.

@AlanMunn what did you do?

@UlrikeFischer Removed the \clearpage commands from \@include.

@AlanMunn Oh. ;-)

@AlanMunn I found Heiko's answer - that's really a memorable warning ;-)

@UlrikeFischer He had typographical advice from a reputable source.

@DavidCarlisle hmmmmmm

@DavidCarlisle yes I directly thought that he had help here ;-)

@DavidCarlisle Had a chance to look at the test file business?

@JosephWright not really (that is, I looked, but can't really run stuff to look properly at present)

@JosephWright probably as well as not using fonttext.cfg -TU test configurations probably ought to be using regression-test.tex or a version of test2e.tex that doesn't hack the luatex font loading quite so much

@DavidCarlisle Sure, we could look at that; just swapping out the test file would be pretty easy

@DavidCarlisle, @UlrikeFischer Have either of you talked with @MartinScharrer about filehook?

@JosephWright not me

@DavidCarlisle OK

@DavidCarlisle Thought @UlrikeFischer was the more likely ...

@JosephWright no you added the last time a comment to the issue. Should I add another one?

@UlrikeFischer No, that's OK, I guess a direct mail is needed

@LaTeX3Team @UlrikeFischer @DavidCarlisle @JosephWright etc. Have you seen the mlist_to_hlist thread on luatex-dev?

@MarcelKrüger yes, sounds like something that should be improved.

@UlrikeFischer Right, I think we should discuss in general which role the kernel should play here. Do we want to keep exposing only the standard callbacks or should we add our own "user-defined" callbacks too?

I currently think that it might be good to create a canonical package where we create more composable "wrappers" for callbacks.

@MarcelKrüger Sure, but that would be beyond-the-scope of ltluatex I think: it's really just 'core allocator' stuff

@JosephWright I agree, so in ltluatex we only need to adjust some types (github.com/latex3/latex2e/issues/188) to get more sensible error messages.

@JosephWright This reminds me of of a related thing I was thinking about: Especially for callbacks like pre_linebreak_filter and post_linebreak_filter, I often use them together. This can lead to problems when multiple people do this and manipulate each others structures. I think it might be better to have post_linebreak_filter run "in reverse", to get proper nesting when multiple packages overload these in pairs. But I'm a bit worried about compatibility...

@Skillmon twitter.com/wtsnjp/status/1179229067377364993

@PauloCereda well, my Japanese is a bit rusted (as a matter of fact, I really learned a bit of Japanese), what does he say? I can see whether I can strip this limitation, but I won't be able to test things. I'll try

@Skillmon Oh I was just pointing out there was a duck in the terminal, courtesy of your package. :) It seems they tried duckuments with some Japanese variant of the engine, but that is as far as I can go. :)

@PauloCereda is there a ctex, I think there was. I'll try things.

@Skillmon I think Norbert might help as well.

@PauloCereda I was hoping that someone would actually find that joke and booom I'm on Twitter :)

@Skillmon I assume they are using ptex (or uptex) do you error with those engines?

@Skillmon ooh

@DavidCarlisle pTeX, yes.

@DavidCarlisle might be, but I think those could work, it was done just to limit support to engines I actually use and can test. Time for another upload...

@Skillmon want me to write a reply telling the author is providing an update? :)

@PauloCereda would be great, but I want to test things first.

@DavidCarlisle do you by chance know how to get the number of pages in a PDF in pTeX and/or upTeX?

@Skillmon sure thing, mr. rabbit!

Does anyone know anything about pTeX and upTeX? Can they include PDFs?

@Skillmon the driver is (x)dvipdfmx imho and so it should be not a problem.

@UlrikeFischer I get a "no BoundingBox" error with example-image-duck.pdf and pTeX and upTeX, that works with pdfTeX, LuaTeX and XeTeX.

@Skillmon did you pass the correct driver option (dvipdfmx)?

@UlrikeFischer 'course not. Thanks :)

@MarcelKrüger imho this is like pdf resources and hooks: it must be managed and the longer one waits to do it right the more difficult it gets ;-)

@Skillmon how would you do it in pdftex? usual \label on last page?

@UlrikeFischer hmm shouldn't that be automatic (in the distribution graphic.cfg) (@JosephWright)

@Skillmon oh you meant in an included pdf, not the one being written:-)

@DavidCarlisle I don't know. I was wondering too as imho Joseph said at some time that this engines/backends can be safely detected.

@UlrikeFischer ooh experimental :)

@DavidCarlisle exactly :)

@DavidCarlisle well, whether it should or shouldn't, it wasn't

Is there any manual or reference for pTeX or upTeX that is comprehensible?

@PauloCereda you could reply to him that I can fix this incompatibility, but only with a reduced feature set (no random duck images in example-image-duck for pTeX and upTeX since I don't know how to get a PDF's page count in those engines)

@PauloCereda but the update will take one or two days to make the code look good (current test version just does things which are never really needed for pTeX and upTeX and disable the feature at one point only)

@PauloCereda Also he might file a request in the issue tracker on github: github.com/Skillmon/ltx_duckuments/issues

@yo' congratulations!

@Skillmon like this:

\documentclass{article}
\usepackage[dvipdfmx]{pdfpages}
\begin{document}
 \makeatletter\def\AM@currentdocname{example-image-duck.pdf}\AM@getpagecount
 \show\AM@pagecount
\end{document}

@UlrikeFischer wow, how did you find this?

@Skillmon well I assumed that pdfpages knows how to get it ...

@UlrikeFischer now that I see you're including it... I'll take a look at pdfpages implementation. Thanks for pointing me!

@Skillmon it is comprehensible if you read Japanese

@Skillmon github.com/latex3/graphics-def/issues/22

@DavidCarlisle well, I don't. I learned Japanese for less than a year, nothing got stuck in my head, unfortunately

@Skillmon that's almost a year more than me:-)

@DavidCarlisle :)

@DavidCarlisle それでも、どういうわけかあなたは完全な流fluさを達成しました。

@AlanMunn i can read that (super slow, and I have no idea about the Kanji), but I have no idea what it means.

@AlanMunn that's because I'm a real linguist

@Skillmon yes by "variants" I meant uptex euptex eptex as well as ptex

@DavidCarlisle somehow overread that "variants"

@Skillmon changing graphics.def would be trivial but I'd like Hironobu Yamashita to confirm any change doesn't break any patches made in Japanese TeX distributions.

@DavidCarlisle Want to guest lecture in my grad class this afternoon? :)

@AlanMunn subject?

@DavidCarlisle Parametric variation in syntax.

@AlanMunn ooh \def\zz#1{\def\zz##1{##1}}?

@DavidCarlisle Not quite, I'm afraid.

@DavidCarlisle ooh

@Skillmon Thanks!

The “In futurum” section of Erwin Schulhoff’s 1919 Fünf Pittoresken (starting here at 8:08) consists entirely of rests.

The direction tutto il canzone con espressione e sentimento ad libitum, sempre, sin al fine means “the whole piece with free expression and feeling, always, until the end.”

@yo', @AlanMunn, @barbarabeeton ^^

Floppy-eared rabbits, also known as lop rabbits, are more susceptible to serious ear and tooth disorders than upright-eared domesticated rabbits, according to new research. The findings are raising important concerns about the appropriateness of breeding lop rabbits, adorable as they are.

@Skillmon ^^ oh no

@DavidCarlisle Looks like a backend change for expl3/(u)pTeX ... probably do that plus handle options with no backend=, though likely as part of \DeclareDocumentMetadata

@UlrikeFischer ^^^

Hello friends. Nowadays, what is recommended to insert sub figures: subfig, subfigure, subfloat or some other? I checked the documentation and they appear to be old packages. What do you recommend?

@Sigur I would use the one from the caption package.

@DavidCarlisle, @UlrikeFischer See new branch ;)

@JosephWright yes probably means can't really devolve all this to the format and have a different format for each driver (as we discussed at one point) needs an option (probably)

@JosephWright did @MarcelKrüger's ltluatex change make it to the tagged release?

@DavidCarlisle Yes, looks that way, so some common meta-data stuff is a good plan

@DavidCarlisle Not at present, but as we are sitting on the release awaiting @MartinScharrer's update to filehook, I could cherry-pick

@JosephWright did you hear back from Martin?

@DavidCarlisle No

@DavidCarlisle Was wondering about your take on beamer, Etc.

@JosephWright lppl 1.4 that adds a clause "latex team may update this package on ctan as they deem necessary...)

@DavidCarlisle I had seriously considered something like that: not in the LPPL but as for-example a statement on the CTAN upload page

@JosephWright I started to trace the beamer/sans math fonts. bm stuff but was at work and ran out of time and lost the will to carry on, it shouldn't be needed but beamer's font setup is ... different... and if it's working now I'm inclined not to break it...

@DavidCarlisle Yes, I'm not really sure what Till was up to, but I'm stuck with it

@MarcelKrüger not really a view on the other callbacks yet will think about it, but I wanted to sign off on the "clear bug" cases so that they could sneak into this release at short notice.

@UlrikeFischer, thanks for attention. So, you mean to search for such package in the caption doc, right?

Or to use only caption pkg for this porpouse?

@Sigur it is subcaption ctan.org/pkg/subcaption.

@UlrikeFischer, I'll take a look

@UlrikeFischer, nice, it is updated, from 2019. Thanks.

@DavidCarlisle The ltluatex changes cherry-pick across to master just fine

@JosephWright so how does that then work with the auto-release do you delete the tag and re-tag to trigger a build, or live with a second tag or?

@JosephWright chemisty google doodle I see

@DavidCarlisle ?

@JosephWright the ltx is unpacked but not installed.

@DavidCarlisle If you force-push the tag, Travis will trigger a new build. Travis will skip the deployment if the filename already exists in the GitHub release, so to ensure that the new files are uploaded, the GitHub release, or at least the files associated with the GitHub release, have to be deleted first.