@egreg yes currently though aim is to make it simple enough that browser makers don't object to adding support for math, so question is what's a reasonable default, most likely users will be able to adjust via css (depends a bit how it ends up) thanks, generally "do what tex does" is a reasonable guideline so good to get some opinions before suggesting deviating

stackoverflow.blog/2019/11/14/… So long? Have they ever hear about LaTeX?

What takes up more memory, \chardef or \def, assuming \def would hold single character (\def\testName{a}).

@bp2017 I assume def (but I'd have to check the sources) that is never really a concern though as the token space allocated is almost never the limit that you hit in modern tex installations. (The actual memory is the same either way or if you don't do the allocation at all, classic tex doesn't do dynamic memory allocation)

quack

ooh Pokémon

@PauloCereda breakfast!

@PauloCereda ?

@DavidCarlisle oh no

@DavidCarlisle the new Pokémon game is out

@PauloCereda well it's a bit early for dinner

@DavidCarlisle ah it makes sense

oh no

@PauloCereda strangely I missed that.

@DavidCarlisle swordshield.pokemon.com/en-us

I might get this game... for reasons.

@PauloCereda I saw pokemon go for about a week then it went out of fashion and vanished, but otherwise I know nothing, is it like cricket?

@DavidCarlisle You pretty much throw balls at other beings, so I am very inclined to say yes. :)

@PauloCereda throwing is a foul, you have to bowl the ball

@DavidCarlisle ooh rules

For those who does not know the beauty and simplicity of cricket:

You have two sides, one out in the field and one in. Each man that's in the side that's in goes out, and when he's out he comes in and the next man goes in until he's out. When they are all out, the side that's out comes in and the side thats been in goes out and tries to get those coming in, out. Sometimes you get men still in and not out.

When a man goes out to go in, the men who are out try to get him out, and when he is out he goes in and the next man in goes out and goes in. There are two men called umpires who stay all out all the time and they decide when the men who are in are out.…

@DavidCarlisle oh

My past self was clever

ooh

@PauloCereda I have uploaded around 190 packages to ctan this morning

@DavidCarlisle OH MY

@DavidCarlisle Petra will be pleased. :)

@PauloCereda well 95 packages, twice each

@DavidCarlisle ooh

@DavidCarlisle I just wondered ;-). Why twice?

@PauloCereda she didn't like the first lot:-)

@DavidCarlisle :)

@UlrikeFischer to annoy the CTAN blokes reasons unknown. :)

@UlrikeFischer makerobust apparently needed an LPPL statement 5 times longer than the code (and hobsub same) and somehow a rogue luatex.pdf got left in the main oberdiek.zip (the manifest file ensured exactly the right tds zip content but the top level zip used wildcards and I'd moved the git history to old-sources but not actually git removed luatex.dtx so it left it in....

@bp2017 Who cares? It's the token count that's important

@DavidCarlisle Used l3build?

@UlrikeFischer This first round of split packages are special as they are the unused legacy ones so I didn't use the existing dtx, future ones presumably will be based off the existing dtx so less likely to forget LPPL etc. Also I should set up an l3build upload config we can copy into each split off repository as it is easy to miss things on the web form eg the ctan location it guesses the original location under oberdiek, but that's hidden so you don't see it unless you "show all fields"

@JosephWright not for the main oberdiek as I haven't had time to set up the tds mapping, (although actually that would have had teh same issue I think as it uses wildcards by default so would have included luatex.dtx if it were there)

@JosephWright each of the split out repositories is using l3build and as the plan is to essentially dwindle the main one down to nothing, I didn't want to change it too much.

@DavidCarlisle Ah

@DavidCarlisle Makes sense: so for the 'new' packages, we can release without needing a Linux system ...

@JosephWright yes (I release without linux as well, but I know what you mean:-)

@DavidCarlisle ooh Joseph is mean

@PauloCereda we knew that already

@DavidCarlisle surprisingly, Heiko used arara in the early build stages of his bundle. :)

@DavidCarlisle he da boss

@PauloCereda that explains a lot

@DavidCarlisle ooh

@DavidCarlisle uh-oh

@DavidCarlisle :)

@JosephWright bah

@PauloCereda Why? Are there other tools other than arara? :)

@CarLaTeX not as cool, wacky or buggy as arara. :)

@PauloCereda arara is the coolest in the world :)

@CarLaTeX yay

@PauloCereda there's still a lot to be said for using bash (or whatever command line shell you use) when reverse-engineering a build system for something line the oberdiek bundle it's so much easier to do it interactively, copying the lines into a script if they work, and then when you have a working script, just call it done rather than work out how to get l3build or arara to make the same thing.

@DavidCarlisle agreed. :)

@DavidCarlisle I never saw Heiko's script, but I believe it was complicated.

@PauloCereda no neither did I so i just had to start from the original ctan tds zip file and manufacture a script that unpacked the sources into the same tree.

@DavidCarlisle ooh

@PauloCereda that wan't possible to specify in l3build until recently but in bash it is just

# fill tds tree
for i in `cat manifest.txt`
do
    cp `basename $i` $i
done

@DavidCarlisle wait for my l3swissknife module. :)

@PauloCereda where manifest.txt is essentially unzip -l original tds-from-ctan.zip`

@DavidCarlisle pretty straightforward

@PauloCereda but this is the bit you'll like best:

(defun ho-tex ()
  (interactive)
  (goto-char (point-min))
  (re-search-forward "File:\\s-+\\(.*\\)\\.dtx\\s-*\n%\\s-*Version: \\(20../../..\\) v\\([0-9]+\\)\\.\\([0-9]+\\)" 600)
  (let ((today "2019/11/12")
	(pkg  (match-string 1))
	(d  (match-string 2))
	(v1 (match-string 3))
	(v2 (match-string 4))
	)
    (narrow-to-region (point-min)
		      (progn (search-forward "begin{History}") (point)))
    (goto-char (point-min))
    (query-replace d today)
    (goto-char (point-min))
    (query-replace (concat today" v" v1 "." v2)

@DavidCarlisle Brilliant!

Absolutely brilliant!

@DavidCarlisle perfect timing, have you seen this? github.com/mkoeppe/cl-bibtex

@PauloCereda who'd program in lisp these days?

@DavidCarlisle mad people and Clojure fans. :)

Feature request for the kernel: \immediate@protected@write

This definition seems to do what I wanted

Reason: I was about to release a new memoir version when I noticed a problem with memoirs index generation (non ascii chars are expanded), I need to use \protected@write, but I also need it to be immediate because of the manner in which memoir handles indexes (the index is first written to the aux, and then at the end of the doc, written from the aux to the idx.

@daleif I think you can simplify it a lot you don't need the two-step expansion with \edef nor the \nobreak node as \immediate\write doesn't make a node

The above macro seems to work as a drop in replacement, I'll add it as a provide command

@daleif untested, but more like

\long\def\immediate@protected@write#1#2#3{%
\begingroup
#2%
\let\protect\@unexpandable@protect
\immediate\write#1{#3}%
\endgroup
}

@daleif my version expands \thepage your version writes \thepage as \thepage probably my version does what is required more often as it has a chance of reflecting the current page but either could be wrong of course, if you are writing page numbers then probably you shouldn't be using \immedate\write:-)

@daleif memoir has something already it seems?

$ grep -r immediate@protected@write
memoir/memoir.cls:\newcommand{\immediate@protected@write}[3]{%
memoir/memoir.cls: \IfFileExists{\jobname.ent}{\immediate@protected@write\@notefile{}{#1}}{\mempnofilewarn}%
pagenote/pagenote.sty:\newcommand{\immediate@protected@write}[3]{%
pagenote/pagenote.sty:  \let\@pnwrite\immediate@protected@write
volumes/nowtoaux.sty:  \immediate@protected@write\@auxout
volumes/nowtoaux.sty:\long\def \immediate@protected@write#1#2#3{%
volumes/nowtoaux.sty:  \immediate@protected@write\@auxout

@DavidCarlisle ahh, right, didn't even notice....

@daleif as the first two use newcommand, if we added it they would break, usual problem with breaking contrib stuff if you do anything....

@DavidCarlisle I'll change it to \providecommand in my end

@DavidCarlisle as mentioned indexes are a two stage process in memoir, so the page number has already been resolved here where I wanted to use it.

@daleif it should drop the \if@nobreak\ifvmode\nobreak\fi\fi bit as well, really, that is there to avoid the write not adding a page break

@daleif yes sure, you can say that in the use you have in mind in memoir, but in the format we'd have to make \thepage do something and making it not expand so expand to whatever page it is when you read the file is probably the wrong thing.

@DavidCarlisle I'll change it into your suggested version, and double check what it is actually used for.

and now I no longer get \T1\ae errors when using creating an index with non-ascii chars

@DavidCarlisle as Heiko already mentioned years ago ;-) tex.stackexchange.com/a/82085/2388

@UlrikeFischer He knows a bit of tex

@UlrikeFischer hobsub makerobust and oberdiek all installed at ctan:-)

@DavidCarlisle ooh

@daleif desert bus for hope yay!

@PauloCereda that was bussing awesome, I went to bed right before they had the Legend of Zelda sword donation drive which raised $94k in an hour, woke up 3 hours later only to learn they has raised 150k in those few hours

@DavidCarlisle ;-). Do you know if VTeX is still used?

@daleif wow

@DavidCarlisle ooh

@UlrikeFischer ooh

@DavidCarlisle is there a github for csquotes? (not mentioned in the manual), may have found an error

On my just updated TL19 setup, this MWE gives an undefined error

@daleif github.com/josephwright/csquotes

@UlrikeFischer thanks, I'll post it there. Bit strange that I did not notice issues with it yesterday, but was probably too occupied with index error.

\langname didn't I see @JosephWright define that locally in an ltnews doc the other day

#blamejoseph?

@PauloCereda interestingly that blame X is also a thing in desert bus

@daleif :)

@daleif I think there is already a fix: github.com/josephwright/csquotes/commit/…

@UlrikeFischer probably, I just redefined \hyphenquopte in my test book project and now it compiles without errors on current TL19 and next memoir version, so hopefully no #blamedaleif

memoir 3.7i sent off to CTAN

@daleif ooh

ooh a bomb

Hi mr. bomb

(")>/

hello mr duck

ooh :)

i tried to come up with an ascii bomb but it didn't really work

hey, can any of you tikz-heads tell me why the word "Enumeration" is getting chopped in half here, when it has so much room? i was going to make it a proper question, but i don't really want it to be out there on the web for all to see just yet: overleaf.com/read/phfmqgjfntcy

until after i hand in my assignment

quackwack 🦆💣

@voices where exactly, mr. bomb? I could not find it...

umm

the second yellow box on the left

line 131

Ah found it, apologies.

with the overfull warning

i've coded myself into a corner i think haha

Looks to me that the box is quite too short...

yeah, but where is the box? from memory i thought the size was established on line 37

got it. thanks

@DavidCarlisle I see a question was posted that's ideal for you!!

@JosephWright answered

@DavidCarlisle I'd seen :)

@DavidCarlisle I wonder if Petr will appear and do it all in PDF specials ...

@JosephWright I nearly just used \circ, \bullet and \Box since he said they were for item labels

@JosephWright ooh wipet

AGM TOMORROW GUYS

@PauloCereda no need to spoil today with such bad news

@DavidCarlisle oh no

@JosephWright I just tried to run the first time "l3build doc" in the pdfresources, it compiles in the build folder and there is a pdf, but at the end it says " ! Compilation failed". Any idea why?

hey, how do you curve the line between nodes again [bend] something or other i forget

@UlrikeFischer Did it fail say on index creation?

@JosephWright a yes. there is an error.

i thought you could just [bend left] or [bend right]

maybe that's just edges not nodes

@JosephWright Now I have only Input style error in the ilg, but the compilation still fails. Do I need some special .ist?

@everyone: Is there a simple way, to change x- and y-axis in pgfplots's \addplot3 ?

@cis Plotting f(y,x) is not an option?

@mickep I used that "dirty trick" now "({2-3*t+t^3},{t},{0})"

@cis ohh, dirty tricks.

There is a Man, that uses TikZ like a daredevil / go-getter!!!

matheplanet.de/matheplanet/nuke/html/…

Wow!

@norbert hello:-)

@DavidCarlisle Hello David. I apologize for last night, but the Internet line broke down. After a busy day I'm connecting and I hope to go to sleep as soon as possible. Thank you and excuse me again. Now I'm editing my question and answering various comments

@Sebastiano I deleted the comments as you edited the question, but as I note in the answer, looks like a bug that you get an undefined command error

@DavidCarlisle I will inform you later of the reply given by the author of the package. Thanks again. The Internet line is very slow. My sincere greetings.