12:08 AM
Hey!
I think that sometime ago I found a question asking how to insert a dashed line using array environment, but without any package: it was {clr:c}. But in LaTeX is not possible to do that; one option is we have to add \usepackage{arydshln}, but I never heard any of that package, so I found that, for example, in math.SE and also in other sites we can use : inside array without any problems, very easy!
So I asked myself: does MathJax load that package? My answer was: I do not know. Do you know if it yes? In any case, is there another way to use dashed lines in array? P.S. I tested it on Overleaf

@manooooh mathjax doesn't load any package (it doesn't use tex)
@manooooh arydshln is the usual way to do that, what objection do you have to using the package?

12:42 AM
@DavidCarlisle ok, I google but I could not find anything related to "does MathJax load that package?", so now I know why. Thank you
@DavidCarlisle I have a remote thought that at some point I made use of that command but I do not remember having loaded a package with that name (arydshln). That's my objection. I tried with mathtools and array but it did not work...

@manooooh no arydshln is the main package for dashes
@manooooh mathjax is javascript not tex

@DavidCarlisle but I only need for a couple of matrices... Maybe | and that's all. What do you think?
@DavidCarlisle oh, nice!

@manooooh that's why it's not enabled on this site

I got confused with the supported LaTeX commands
@DavidCarlisle for being nice? ;)

@manooooh yes that is the list of commands it tries to emulate

12:48 AM
@DavidCarlisle yes... but that is not the list of packages... I got a very confusion lol
@DavidCarlisle there will be a day when you ask a question on the TeX.SE page?

@manooooh I'm waiting for egreg to show me how, but he is a terrible teacher
5

@DavidCarlisle gotcha

4 hours later…
5:06 AM
@marmot prepare for a killer question... Probably I will start a bounty for colecting other answers, or simply to make good the hard work that is going to take who wants to answer

5:40 AM
1

I would like to plot S1: -(x-1)^2+y^2+z^2=1, x=1 and z=0 and their intersections using tikzpicture environment: Using this post about the equation of the hyperboloid of a leaf I end up with two type of equations. Let x^2/a^2 + y^2/b^2 - z^2/c^2 = 1. Parametric equation: x=a*cosh(u)*cos(v)...

@marmot ^^^^ :P

6:04 AM

6:28 AM
@DavidCarlisle Not currently, though not a bad idea
@DavidCarlisle Probably we need a bit more functionality for choosing TDS destinations: at the moment its largely hard coded

@manooooh Need to go hibernating now... ;-)

7:31 AM
Quack!

7:41 AM
@JosephWright I think in theory you could tell docstrip to use the directory layout and then copy that to tds but I didn't trust it and when taking over the bundle it just seemed safe to make the manifest first by listing the zip file form ctan and then forcing that the build always re-creates the same list.
Apr 15 at 9:38, by David Carlisle
@PauloCereda dinner!

@DavidCarlisle oh no
@JosephWright so... the membership value will be recalculated...?

@PauloCereda I'll forward something

@JosephWright Thanks

2 hours later…
9:34 AM
@UlrikeFischer @JosephWright I'll need to adjust the obediek test results again as the checked in tlg have truncated paths to the packages (../accsupp.sty but if I build locally before running the tests I get (accsupp.sty with local copies.

9:46 AM
@DavidCarlisle Such differences between current directory and texmf is why I normally prefer to build local texmf trees. With luaotfload-fonts it is even more problematic as the current directory is not search first.

@UlrikeFischer but l3build really wants the files copied to the local directory. My current plan is to change the mkctan script so it builds the unpacked sources in build as l3build would and then adjust the test results to match. (It used to take an hour to run that script before, if I add the l3build check at the end it's even longer:-)
@UlrikeFischer I just zapped the pre luatex 0.65 versions and the oberdiek. version of luacolor, tried to run the test suite and every test failed which was a bit of a shock:-)

@DavidCarlisle ah yes, just got the reference

@UlrikeFischer because of the ../ so unrelated to luacolor other than I'd unpacked everything locally, why the "wow" anyone can do that:-)

Wow I managed to reference a later comment ;-)

@UlrikeFischer on bibtex and \zap@space I'd thought of that too but I'd really not introduce more uses of that, I suppose bibtex ought to be fixed to parse the argument more sensibly but failing that I think we need to remove space around comma
someone at work just pointed at this
5

10:05 AM
@DavidCarlisle I know, but now that one can set a local texmf.cnf I would like an option to do "l3build install" and then "l3build check" without copying (apart from the support files) and unpacking to the test directory. Imho this would give more realistic tests as the location in the tds must be correct too then. @JosephWright
@DavidCarlisle and if there were some way to install only the changed dtx it would save you a lot time before the tests ...
@DavidCarlisle Sure changing bibtex would be the way to go. But is there any chance that this could happen?

@UlrikeFischer of course the bundle is ridiculously large, I hope to split it up at some point but the hobsub thing means there are hidden dependencies in the build
@UlrikeFischer get @JosephWright to make a pull request against the texlive sources?

10:26 AM
@DavidCarlisle yes, I saw the issue from Karl about ifpdf etc. But the latex sources are quite large too. How do you manage to test something in the git version without waiting all the time for the unpacking?

@UlrikeFischer drink coffee (travis is also much faster than my machine, I really should try to work out why one day, we are not using any obvious parallelism in the build and check scripts so it's surprising that the vm they provide is that much better than this machine which was a basic work laptop, but is an ssd i7 so shouldn't be that slow)
@UlrikeFischer also the main time is building the doc, and normally in base we don't build that, just unpack and check, but the mkctan script here does both (and I didn't have any option to just unpack to a working tds arrangement without building the doc)

10:47 AM
News from the wild duck cluster:

11:08 AM
@DavidCarlisle well it takes arount 10-15 minutes on my machine to unpack both latex3 and latex2e, time for a good tea ;-). But I seem to miss the format files if I do l3build install --texmfhome=E:\latex-install, they are created and there are some in the build folder but I can't find them in the texmf or some other texmf.
@DavidCarlisle an regarding slow. Do you agree with The code isn't so good. It uses a lot of \ifthenelse switches and similar which are quite slow. ?

@UlrikeFischer I tried to give hints in the ifthen doc that you shouldn't use the package but no one believed me

@DavidCarlisle did you saw my benchmark results?

@UlrikeFischer you can run the 2e checks in that time? i really must get a faster machine
@UlrikeFischer not in detail:-)

@DavidCarlisle no. Only unpacking. (I have an old, slow machine).
@DavidCarlisle tex.stackexchange.com/a/458937/2388. I should rewrite the package ...

@UlrikeFischer oh it's your code, sorry:-)

11:19 AM
@DavidCarlisle I was young ...

@UlrikeFischer distinguishes you from @egreg who's always been old.

@DavidCarlisle I even had forgotten that it used ifthen - I'm telling people not to do it (just a few days ago on a german list) and then looked in my code and thought ups.

@UlrikeFischer it could have been a lot slower, if we'd have just left the 2.09 implementation

12:20 PM
hi guys. I have a bad problem.
I am from Iran. The website in which the pictures of TeX.SE is uploaded is blocked in this country.
So I can not see the pictures of the questions and answers at all.
Is there any way we can change the root of pictures so that they be loaded?
Or is it possible to change the website of the pictures?

@EnthusiasticEngineer That's not really a tex question. See meta.stackexchange.com/questions/209161/… for some options.

1:11 PM
@EnthusiasticEngineer I don't think it's possible. Imgur is used by the whole StackExchange network.

1:52 PM
For anyone interested: I found a German LaTeX introduction referencing tikzducks while searching for a completely unrelated topic :)
4

2:11 PM
@TeXnician Well, that should teach you a lesson. How dare you searching for other things than ducks? ;-)

2:30 PM
Didn’t luatex print how many nodes were used at the end of the log/terminal, or is my mind making things up?

@KhaledHosny Do you mean this?

@TeXnician ooh

 381 words of node memory still in use:
2 hlist, 1 vlist, 1 rule, 2 glue, 3 kern, 1 glyph, 4 attribute, 45 glue_spec,
4 attribute_list, 1 write nodes
avail lists: 2:15,3:2,4:1,5:21,6:1,7:17,8:1,9:6

I am back!

2:46 PM
@PauloCereda From where?

1 hour later…
4:04 PM
Apr 15 at 9:38, by David Carlisle
@PauloCereda dinner!

Is there a way to smuggle an \hspace{0pt} at the begin of an list item?

@UlrikeFischer \item\hspace{0pt} :-)

@DavidCarlisle and now again with nomencl ;-)

@UlrikeFischer never used it

@DavidCarlisle lucky you. I have a
 \item [{TKÜ}]\begingroup Telekommunikationsüberwachung\nomeqref {0}
\nompageref{28}
and wonder how to press this in a twocolumn layout. (Don't ask me about the begingroup ..)

4:11 PM
@UlrikeFischer put some inter-word spaces in that sentence?
@UlrikeFischer \leavevmode\hspace{0pt}\begingroup... ?

@DavidCarlisle sure, but how to get it there? The file is made by makeindex. I will try to redefine \item locally. Should work.

@UlrikeFischer ah I thought you meant that the package had already redefined \item so just adding \hspace wouldn't work

@UlrikeFischer form a nearby city. :)
@DavidCarlisle oh no

@DavidCarlisle thanks rubber ducking helped ;-)

@UlrikeFischer I'm still undecided whether I should edit the oberdiek tlg for packages in the current directory, or modify the build to put them somewhere that l3build doesn't see and set a local texmf.cnf for the tests

4:27 PM
@DavidCarlisle I'm undecided either. If I were sure that the second option works well for all targets (save, check, install, etc) I would prefer it as it reflects better the setup in a real user system. But I fear that there are hidden traps

Yes, I don’t seem to get it here. What I get in the log is:
PDF statistics: 81 PDF objects out of 1000 (max. 8388607)
49 compressed objects within 1 object stream
0 named destinations out of 1000 (max. 131072)
1 words of extra memory for PDF output out of 10000 (max. 100000000)

@KhaledHosny plain luatex?

@UlrikeFischer Yes

@UlrikeFischer I did a multi column thing to nomencl once. See if it does what you need: tex.stackexchange.com/a/432863/134574

@UlrikeFischer Looks like I need to set \tracingstats=1, but I’m sure I have seen it before without doing anything special.

4:36 PM
@KhaledHosny latex does set tracingstats to 1

@UlrikeFischer Aha, that is it then. Thanks!

@PhelypeOleinik thanks but the problem is not to get multicolumns but the hyphenation of long german words.

@UlrikeFischer Ooh, sorry. I misunderstood it :P

1 hour later…
5:43 PM
> Strategy guide maker Prima Games is shutting down
oh no

5:59 PM
ooh there's a cauldron
@cfr cwac! :)

@PauloCereda @UlrikeFischer @CarLaTeX @marmot @barbarabeeton ctan.org/pkg/tikzlings
3

@samcarter "The requested URL /tex-archive/graphics/pgf/contrib/tikzlings/tikzlings-doc.pdf was not found on this server." ;-((((( (the fifth mirror then found it ;-)))

@UlrikeFischer I only got the confirmation that it was installed a couple of minutes ago, so nice surprise that one of the mirrors already has it :)

6:24 PM
@samcarter yay!

@samcarter Yay, let's party!

/party
@samcarter OH MY WE CAN MAKE THE MARMOT DANCE
@JosephWright ^^ why isn't UK-TUG funding this amazing package? ^^

2
@samcarter @UlrikeFischer @CarLaTeX ^^^ duck math @Skillmon

HEY
IMPORTANT QUESTION
WHO ATE THE CHOCOLATE I WON ON BEHALF OF @samcarter?

6:53 PM
@marmot :)

7:16 PM
@PauloCereda they need to save their money to buy crispy duck pizza

7:34 PM
@DavidCarlisle I managed to get a more meaningful error from my TeX document, although I don't know if it is the right error. Using list and make (with my real document), I get
\g__prg_map_int
as undefined.
But I'm getting it for a picture that compiled fine with the default mode.

7:45 PM
@DavidCarlisle The hobby package uses it, though this can't explain the error I was getting as the archive I sent you didn't include any significant TikZ stuff apart from the externalisation.
@DavidCarlisle Never mind. This is just the outdated hobby error.

@CarLaTeX @PauloCereda yay, party!
3
@marmot Great math! Nobody can say this is wrong :)

@samcarter Of course not. And the derivation is crystal clear. ;-)

@PauloCereda I'd blame the tikzmice - I saw them with cheese, maybe they also have the chocolate
@marmot One does not need a crystal ball to for it :)

@samcarter I also think it must have been the mice.... ;-) (Luckily the chat does not transmit chewing noises ;-)

@samcarter the Bär suggests an improvement for the docu:

8:00 PM
@UlrikeFischer :) But he'll have to wear his team shirt!

8:21 PM
3
@marmot ^^

@cfr that's just a scratch count register used in lots of l3 code
/build/doc/l3prop.dtx:      { __prg_map_ \int_use:N \g__prg_map_int :wn } \@@_pair:wn
./build/doc/l3prop.dtx:    \int_gincr:N \g__prg_map_int
./build/doc/l3prop.dtx:        \int_gdecr:N \g__prg_map_int
./build/doc/l3prop.dtx:          { __prg_map_ \int_use:N \g__prg_map_int :wn }
./build/doc/l3seq.dtx:    \int_gincr:N \g__prg_map_int
./build/doc/l3seq.dtx:    \cs_gset_eq:cN { __prg_map_ \int_use:N \g__prg_map_int :w }

8:34 PM

@marmot you will not add the intersections of surfaces?? Hahha, no problem. I will find the way to do it, thank you!!

@manooooh I'd like to if I knew what you're asking. Which intersections are you referring to?

@marmot the curve of y^2+z^2=1 and the hyperbola. You can see it un the first image (orange and green curves)

@manooooh Ahh, you mean projections. Will do it asap.
@manooooh ^^^ like this?

9:07 PM
@marmot a cross section of your burrow?

@DavidCarlisle No, that is obviously the mass shell of the mouse that ate @samcarter's chocolate before @PauloCereda could get it. ;-)

1 hour later…
10:33 PM
@marmot yes, now it's a full answer! Sorry for my bad spelling