12:13 AM
@cfr Seems like you are also asking about a crystal ball. ;-)

1:09 AM
@marmot One of those would have been useful, indeed. At least if it could read egreg's mind.

@cfr Maybe he lost is, so there is nothing to read. (JUST KIDDING!)

2 hours later…
2:44 AM
@marmot Well, he does think 'the missing brackets should be fixed' in a comment on a question involving multiple issues and multiple packages with multiple issues constitutes a clear bug report. So you might be right. Or I might just be extremely dumb. Or, of course, both.
@marmot At any rate, a crystal ball would have meant the bug got fixed a year ago (and several other bugs nobody has told me about but which follow by implication). So it would have been useful, for sure, and saved some headaches.
@marmot Can you use glossaries with memoir, do you happen to know? It has its own glossaries thing, but I can only get it to take short descriptions.

2 hours later…
5:02 AM
@cfr Sorry, I never used memoir (and did not make any real progress with the file you sent either). The latter is because I use up all my frustration tolerance on a grant proposal (and also on a talk that I have to give very soon...)
@cfr What I meant to say is that I have no clue what is going on there and gave up.... (not that any of these is frustrating).

3 hours later…
7:53 AM
@DavidCarlisle re your comment on the deleted post: “luatex itself may be coming to the end of the line as the luatex developers are thinking of newer less tex-compatible engines” – I had been under the impression that luatex would likely supersede pdftex in the future, but I don't remember where I got that impression from. I'd certainly be sad to see luatex disappear.

@HaraldHanche-Olsen It's not going to disappear
@HaraldHanche-Olsen They mean it's 'finished' in terms of new features
@HaraldHanche-Olsen Hans et al. want to be able to have freedom to focus purely on what ConTeXt needs, and that means for example dropping any features they don't use (so for example all PDF-focussed TeX primitives, all support for DVI mode)

@HaraldHanche-Olsen it's not going to disappear but if its current developers decide to develop a less tex-compatible variant then it may need some new developers to maintain a tex compatible variant, or it may not, depending on how things turn out.

@HaraldHanche-Olsen LuaTeX can't replace pdfTeX since it's not 100% back-compatible: the plan 10+ years ago was different

@JosephWright Ah, so I have my impression from 10+ years ago … that makes sense.

Nov 27 at 23:05, by Ulrike Fischer
@DavidCarlisle @JosephWright I got today the newest texnische Komödie, and it says that on the context meeting they decided to concentrate after texlive 2019 on a stripped down luatex called MetaTeX, without dvi and bitmap fonts, and which only cares about context.

8:04 AM
@DavidCarlisle Thanks. That's something definitive, at least.

@HaraldHanche-Olsen dropping dvi wouldn't be a problem as it's not clear there has ever been a usable dvi driver for luatex output anyway, bitmap fonts also as most by now are in scalable form as well, but the "cares only about context" may or may not be a problem for lualatex, depending what it means

@DavidCarlisle They've mentioned in the past that ConTeXt doesn't use .e.g. \pdfvariable at all: it's all done from Lua

@JosephWright If I had that mindset I'd probably end up with something more like sile + a thin tex macro layer, which is not necessarily a bad thing but may not be a good base for latex

@DavidCarlisle Yeah, that's probably true: for all I know, that's where they want to go over time, but at each step they still need ConTeXt to work!

8:24 AM
@cfr I tried to send mail to the address listed in the doc of adforn, but it bounced. Removing the bug is very simple: the line \Pifont{OrnementsADF} should be {\Pifont{OrnementsADF}} (that is, with a pair of braces).

@DavidCarlisle dropping dvi could perhaps be a problem for tex4ht.

> Microsoft is throwing in the towel with Edge and is building a new web browser for Windows 10, this time powered by Chromium, news blog Windows Central reported Monday.

@JosephWright one can see in back-pdf.mkiv how they remap the primitives. And seeing there \unexpanded\def\pdfbackendsetpageattribute #1#2{\clf_lpdf_addtopageattributes {#1}{#2}}: we need a command like this too ;-).

@UlrikeFischer Sure :)
@UlrikeFischer We do need to think about page attributes: I've been a bit busy with other stuff, but it's one for the next few weeks

@JosephWright don't worry. I'm busy with other stuff too ;-)

8:55 AM
@ChristianHupfer hi

9:11 AM
@UlrikeFischer Question is whether to alter the 2e kernel: I think we have to
@UlrikeFischer, @DavidCarlisle Is there a case for saying that graphicx/color will always be loaded? We could run them into the kernel permanently, then there will always be a graphics driver loaded. The only minor issue is dvipdfmx/dvisvg selection, but that is solvable.

@JosephWright not sure if the whole stuff needs to be loaded (for color there is the xcolor question), but the drivers would not be bad, but how to you want to solve the dvipdfmx problem?

@UlrikeFischer Well xcolor just overloads anyway, so that's not an issue
@UlrikeFischer For non-standard drivers, one could have a stub graphicx.sty that simply alters the driver. Or have a macro used before \documentclass (or \begin{document}?) that sets the driver. I suspect very few people are using dvipdfmx and not also using (u)pTeX
@UlrikeFischer My point is that to solve the wider issues, one probably needs to be sure a driver is loaded, at which point we could fold in a good number of standard tools. We've done that with inputenc for example
@DavidCarlisle ^^^

@JosephWright sounds ok. I think I would prefer a macro before \documentclass to ensure that everyone gets the message. Perhaps even a more general setup command that e.g. can also set the language or draft mode - than one would no longer need the documentclass options and this could help with the keyval problem.

@UlrikeFischer Does integrate with the same question, yes

9:37 AM
@UlrikeFischer, @DavidCarlisle I guess we should talk on the team list ...

@JosephWright ok, but probably not today. I have to manage a chess tournament with lots of children in the afternoon and to sort out a document (the *** publisher offers a style, which breaks all sort of things by setting 1pt=1bp, redefines all the footnote stuff and then say they don't like the result ...)

@UlrikeFischer Oooh
@UlrikeFischer Sounds fun

@UlrikeFischer I thought the length units in tex were immutable? You can't set 72.27=72, either.

@HaraldHanche-Olsen They set \p@=1bp. It is the same as here tex.stackexchange.com/q/352580/2388 (different publisher but same package author ...).

10:03 AM
@JosephWright ;-). Did you know that enumitem has the interesting feature, that the value of the series key shouldn't be the identical to a key name? Which means that the following breaks with the newest enumitem:
\documentclass{article}
\usepackage{enumitem}
\begin{document}
\begin{enumerate}[series=left]
\item bbb
\end{enumerate}
\end{document}

@UlrikeFischer ooops

@JosephWright it is hilarious. series=right is okay, series=left breaks ;-(.

@UlrikeFischer Really confusing, I'd say!

@JosephWright @UlrikeFischer Good morning and good work. Best regards.

10:25 AM
@JosephWright probably it should check graphics.cfg (possibly renamed) so it could default to different dvi drivers without needing document changes, but Ok, perhaps (would need to decide whether to turn on the extra colour safe box groups by default...
@UlrikeFischer arrrrrr I wish that would go away!

@DavidCarlisle I did put a bit pressure on it and they said they asked the authors about it.

@UlrikeFischer I sent a bug report years ago. How can they ask the authors about it, how are the authors supposed to know it breaks latex?

@DavidCarlisle not the author of the document, the one who made this style.

@UlrikeFischer ah, well they shouldn't ask him either, they should tell him it breaks latex:-)

10:45 AM
@DavidCarlisle ;-)

11:01 AM
@DavidCarlisle One for the team list?

@JosephWright yes

3 hours later…
1:42 PM
@UlrikeFischer Eeeek!!

2:04 PM
@HaraldHanche-Olsen ;-). I see you sympatize ;-)

2 hours later…
4:11 PM
Is pdfLaTeX able to include eps files now?

The rabbits in Super Mario 64 DS are called MIPS.
@Skillmon ^^ since I read EPS, thought of PS, then thought of MIPS, then thought of rabbits in a game. :)

@PauloCereda nice associative chain.

@Skillmon We ducks like stuff. :)

4:28 PM
@Skillmon define "able"
@Skillmon for a decade or so you have been able to use \includegraphics{foobar} and foobar.eps will get included (but pdftex itself can't include eps)

@DavidCarlisle took a look, seems like it's auto converting them to <file>-eps-converted-to.pdf and includes that one (which fails with the file at hand because that one has some corruption it seems)

@Skillmon yes (is this for the q on main site, or for one of your eps?)

@DavidCarlisle neither.
Which q on main?

@Skillmon oh that's a coincidence then, as identical question just got asked as:
0

I'm trying to include .eps images in my document. However, for one particular image, LaTeX gives me the following error: !pdfTeX error: pdflatex (file /compile/Images/T_xz-eps-converted-to.pdf): reading image file failed ==> Fatal error occurred, no output PDF file produced! If I try changin...

5 hours later…
9:58 PM
@AlanMunn -- there's a message on the xetex mailing list that you may be able to help with. it's about using some fake italics in the context of a linguistic example text. (if you don't follow that list, and don't know where to find the archives, let me know and i'll look it up.)

1 hour later…
11:08 PM
@egreg Hmmm .... I'll check the address. I know it is very simple. My point was just that I didn't know about it until yesterday. (That is, I didn't know there was a bug.)

@cfr OK
@cfr Did you get my mail?