« first day (4720 days earlier)      last day (61 days later) » 

8:06 AM
@DavidCarlisle Oh, you got nothing for doing that. :P
@mickep but it's good practice, and @UlrikeFischer is on holiday
@DavidCarlisle I wrote an answer while you were busy editing the question
@mickep summary of your answer: not my fault
@DavidCarlisle The challenge was for you.
@WillRobertson @JosephWright a challenge set by @mickep ^^^
@mickep ctan blames @WillRobertson for that.
The important thing is we can blame someone
@mickep I guess more seriously the question is what (if anything) to do about xetex. That would (probably?) require actually modifying the fonts, i guess.
8:12 AM
@DavidCarlisle You gotta sleep well at night...
@DavidCarlisle I know nothing about how XeTeX works.
@mickep well all you need to know is that it doesn't have a Lua layer to modify the font metrics on the fly...
@DavidCarlisle But I guess one can implement something for one engine that does not work for the other?
@mickep well we could, or we could say it's not worth supporting this engine if only half the stuff works. It's a lot of effort keeping latex working in all three engines (plus ptex variants and hint and ...) and if pdf tagging and opentype math layout don't work in xetex, it's diminishing returns. Not sure we are there yet but....
@mickep xxetex is another possibility
@mickep anyway I approve of "not my fault" answers so I voted +1
@DavidCarlisle Is something working in xetex that is not working in luatex?
@DavidCarlisle haha, thanks
@mickep yes xetex interchar tokens, there is a sort of emulation in xxetex but they are inserted at a place where luatex doesn't really have a suitable callback. They are pretty weird but used extensively in cjk packages, so although the overall aim of cjk linebreaking and fine spacing could be done differently in luatex, there is typically no simple conversion (and the documentation of those packages is difficult, even for a linguist like myself)
@mickep more immediately it must be not that difficult to steal enough of your context math setup to get your example working in lualatex.
8:26 AM
@DavidCarlisle Is there an example somewhere where this is going wrong in luatex?
@DavidCarlisle I'm not sure.
I'm playing with an answer of egregs (left a common on it). In this code
{% #1 (mandatory): a label
\iow_shipout:cn { @auxout }
Wouldn't the IO end up leaving extra space if you have two of there restatethis env after each other?
@mickep there were some questions here, I'll find.. but chat.stackexchange.com/transcript/41?m=49002020#49002020
It is from this answer: tex.stackexchange.com/a/356665/3929
@daleif what do you mean "extra space"? Where should it come from?
@yo' whatsit or similar
I'll do a short mwe, 1min



{% #1 (mandatory): a label
\iow_shipout:cn { @auxout }

\cs_gset:cpn {RESTATETHIS@#1} {#2}%
\begin{theorem} \use:c {RESTATETHIS@#1} \end{theorem}
It is the IO part before the second one.
If one is really nasty one moves the IO into the theorem part of restatethis
8:38 AM
@mickep tex.stackexchange.com/questions/21604/… (can't find the answer i had in mind witha more extensive example)
@mickep, ooh @WillRobertson has arrived to take on your challenge.
8:52 AM
@daleif Oh yeah, I get it now -- sorry for being a bit dumb so early in the morning
{% #1 (mandatory): a label
        \iow_shipout:cn { @auxout } {
@daleif ^^
@DavidCarlisle Maybe one option in lualatex could be to go into a different mathitalicsmode just for integrals?
9:08 AM
@yo' that is an interesting one. Though the end user here, might not be on a system with the hooks. BTW: para/begin ??
9:19 AM
@daleif para/begin will be invoked by the theorem heading, so it's the right spot, and you're sure you're in horizontal mode thus \iow_shipout:cn won't add any whitespace.
But yes, hooks are a TL2022 thing, so "too new" for many. (But not "too new" for Overleaf nor arXiv...)
@yo' ahh, thanks
2 hours later…
10:58 AM
@DavidCarlisle Isn't this the issue that really the fonts need fixing?
@JosephWright yes but technically and politically and in some cases legally, it's rather easier to patch them from Lua. But fixing them would be good (for eg web use not just for xetex)
11:22 AM
@DavidCarlisle I see. It seems that a different type of mechanism/solution is used (not only luatmetatex, but also luatex) to take care of that kind of things. Different script settings.
12:01 PM
No arara complaints so far, so I think it's working again on Windows. :)
@PauloCereda The bird is back in business.
@mickep yaaaay!
@PauloCereda No complaints? Well, there are always Windows users that don't update before opening issues ;P
@TeXnician oh no :)
@TeXnician and we also have some pesky Cygwin users. :)
Oct 16, 2022 at 18:21, by Paulo Cereda
@MarcelKrüger the day I finally come up with a solution for cygwin, David will immediately move to WSL :)
@PauloCereda They never complain, though. Probably because there is only one of them left only trying it on Cygwin to mock you ;)
12:08 PM
@TeXnician ooh
@TeXnician @PauloCereda Hehe, the problem was not that as such, I updated TL2023 together with arara 7.1.2, but, it got corrupted, so I reverted to 7.1.1 ...of course what I didn't do was to update again and try again :( ...anyway everything works OK now :D
Any way to get around this in `l3doc`?
\begin{enumerate}[listparindent=1.5em, labelsep=2em, itemindent=1.5em]
\item \kant[1-2]
\item \kant[3-4]

$ cat -n $(kpsewhich l3doc.cls) | grep '\\listparindent'
510 { \__codedoc_oldlist:nn {#1} { #2 \dim_zero:N \listparindent } }
I am trying to document a list environment using l3doc (It is basically written in expl3), but I can't show as an example the use of listparindent...I think it is a bit aggressive the way l3doc is set up.
12:55 PM
@egreg the example it not too hard to adjust for the user to specify the theorem env they want to use other can theorem, though granted it still uses the same common counter (I'm just giving the env name along side the label)
1:11 PM
@PauloCereda It was nice to see how many users use arara :)
@samcarter Well, on the one hand yes, on the other hand it is saddening that there seem to be so many Windows users of arara yet none stepped up from our call to action to help us (and even if it only would be routinely testing arara builds before the actual release) :/
This is like the fourth arara release in one or two years which we had to patch up for Windows only one or two days after release.
@TeXnician selection bias? If a users fancies debugging software, maybe they aren't a windows user?
@TeXnician New plan: make everybody use a unix based system :)
1:33 PM
@mickep I risked an answer to the italic correction post:-)
@TeXnician but albatross on cygwin is a useful way to keep the house warm (until windows rudely interrupts and shuts the machine down to avoid melting the cpu:-)
@DavidCarlisle Nice, at least it looks like it is giving OK results in these cases.
(@DavidCarlisle I was hesitating if I would give you a +1 on this or on the \text{\cite{aa}} answer, but this felt more valuable.)
@mickep you could do both of course
@mickep \mathitalicsmode=work-as-if-there-was-a-character-at-the-end-of-each-math-list would help here.
@samcarter It's not about debugging, it's just about helping people (and testing by running is not that hard). So unless you say that Windows users cannot be helpful I doubt the selection bias ;)
@samcarter I'd definitely like that :D
@DavidCarlisle Dealing with complaints about cygwin is @PauloCereda's responsibility ;)
@samcarter AIX ?
@DavidCarlisle I do not want to promote a single user too much. Don't want to upset the Italian.
@DavidCarlisle You can write Han^H^H^HSanta Claus your wish list.
1:43 PM
@mickep why not? (to your last point) Ask him if he enjoyed the rugby at the weekend.
1:59 PM
@DavidCarlisle Oh, I leave the rugby teasing for you.
2:26 PM
@DavidCarlisle will it run arara?
@TeXnician ... one needs to download the new version, adjust paths etc. For some users that's probably more than they usually do with software.
$ arara --version
  __ _ _ __ __ _ _ __ __ _
 / _` | '__/ _` | '__/ _` |
| (_| | | | (_| | | | (_| |
 \__,_|_|  \__,_|_|  \__,_|

arara 7.1.0
Copyright (c) 2023, Island of TeX
arara is released under the New BSD license.

New features in version 7.1.0:
* Add (Lua) project support to arara.
* Use a defined domain-specific file system API instead of `java.io.File`.
* Use header mode by default (`-w` restores the old behavior).

See the full changelog of this release at
@samcarter ^^ it does that at least
2:56 PM
@samcarter Well, you don't need to adjust paths for testing. It's actually java -jar downloaded-jar.jar -argument file. But yes, I see your point. Still, we do seem to have a sizable chunk of Windows users and I just can't believe that there is none that is able to download a file and run a command-line call on a regular basis. But maybe I'm really just positively biased by our Linux and macOS users who manage the same.
3:14 PM
@TeXnician sure I can, simply poke me when you need a test (and when I'm not on holiday).
@UlrikeFischer Thanks, will note that for the future before we do our next release that could potentially upset Windows users :)
@TeXnician I could do it but to make a reasonable test I'd have to admit I was on windows and install a windows tex
@DavidCarlisle ooh
@DavidCarlisle Well, yes, I think I take the chance to forego the lands of cygwin and take Ulrike's offer. We just use blind induction: when it works on one Windows machine, it will work on the others too ;)
@TeXnician and it means if it fails again on windows you can blame @UlrikeFischer
3:29 PM
@DavidCarlisle Well, I think as long as I can blame the duck, she is pretty safe ;)
@TeXnician @DavidCarlisle once more escape the work/fun by blaming he uses cygwin. Last time it was some engine compilation...
@TeXnician oh no
user image
@PauloCereda good one
@daleif :)
@UlrikeFischer that was quick, seen it before or are you just better at type1 encoding files than I am:-)
3:36 PM
@DavidCarlisle I checked the afm and made the guess that the right name has something to do with the unicode code point.
@UlrikeFischer so, the latter of my two suggestions :(
@DavidCarlisle since I encoded all the chess fonts I'm good in finding glyph names ;-)
4:16 PM
@Skillmon That's why there are also ones which use lenses :)
@PauloCereda A comedy show recently made a joke about the increased prices of food along the lines of: "Garlic is 50% more expensive, danger of Vampires coming back" :)
4:34 PM
@samcarter ooh
@samcarter but the ones being able to look really far use mirrors, so we might know there are no space-vampires on Mars, but near Alpha Centauri?
@PauloCereda -- Yes, I listened to an interview about this yesterday. I don't think they covered anything except the qwerty keyboard, but I'd like to learn about azerty. So I'll have to read it.
@barbarabeeton ooh :)
@Skillmon Maybe we should scatter some Garlic around the Oort cloud to protect from the Alpha Centauri Space Vampires?
@samcarter there is always a relevant XKCD!
5:05 PM
@Skillmon yes!
@samcarter ooh can we get a space vampire TikZduck
user image
@PauloCereda ^^^
5:24 PM
@samcarter OOOOOH
@UlrikeFischer ^^ we need this for teh Extravaganza!
@PauloCereda That's actually modified from teh Extravaganza (pig replaced by vampier duck)
@samcarter ooh pigs
@PauloCereda Or as @DavidCarlisle would call them: dinner :)
@samcarter oh no
@PauloCereda :)
5:53 PM
@samcarter did someone say dinner?
@DavidCarlisle You're hungry? :)
2 hours later…
7:33 PM
@barbarabeeton I believe French needed all letters with accents on top in the top row, and all letters with accents on the bottom in the bottom row, for technical reasons. But I could be wrong...
@yo' -- That does make sense. When I started working at AMS I was assigned a typewriter with lots of accents on dead keys, so I could handle almost any European Latin-letter language or transliteration. (It was an IBM but not Executive model; it had Prestige Elite type, of which I'm still very fond -- that's the type style on the manual portable I'd used through college.) I still kick myself that I didn't purchase it when it was swapped out for a computer terminal.
7:55 PM
@barbarabeeton Did it have ň and ů? :) (btw, you have 2 mails, sorry that the first one took me weeks to respond to!)
@yo' -- It did have a tilde (Spanish and Portuguese far enough west), but no ring, ogonek, or the under-accent used in Romanian. It did handle anything that was needed for the cyrillic transliteration from Russian or Ukrainian -- in other words, anything needed for the languages AMS typically translated from or reviewed. (Re email, I'm reading it now.)
@barbarabeeton oh, no pushing with the mails!
8:41 PM
@samcarter If it's a vampire duck, then turn it into morcilla.
9:21 PM
@TeXnician It's not an issue of will (I think), it's more an issue of experience with Windows, I've been using Win/Linux in parallel for years and I don't have problems with setting the variables environment or permissions, but it's not the common user (I see it everyday)...If you make an issue for test on Windows I'd be happy to help just ping :D
9:41 PM
@AlanMunn :D
10:37 PM
could someone help me understand this: I have index.tex which does \input{listing_of_CAS_systems_tested} and then I modified the input .tex file and added new code. But latexmk when asked to build index.tex says nothing to do for index.tex ! This never happened before. Below shows the file to be input is more recent than index.tex
>ls -lrt listing_of_CAS_systems_tested.tex
-rwxrwxrwx 1 me autologin 990 Oct  3 17:28 listing_of_CAS_systems_tested.tex
>ls -lrt index.tex
-rwxrwxrwx 1 me autologin 2970040 Oct  3 14:30 index.tex
>grep -i listing_of_CAS index.tex
>latexmk -pdflua  -shell-escape -jobname=LUA --file-line-error index.tex
Rc files read:
Latexmk: This is Latexmk, John Collins, 4 Apr. 2023. Version 4.80.
Latexmk: Nothing to do for 'index.tex'.
Latexmk: All targets () are up-to-date
the file listing_of_CAS_systems_tested.tex is not only more recent, I have actually modified it now. (since latexmk goes by hash code and not timestamps like make). So I expected it to run lualatex on index.tex. Very stange, and this is first time I saw this happen.
btw, I just did full update for TEXLIVE 2023 just few hrs ago. do not know if this has anything to do with this.

« first day (4720 days earlier)      last day (61 days later) »