« first day (4975 days earlier)      last day (29 days later) » 

5:18 AM
Do I need to be able to release test files if a package is LPPL? I generally would, but I think LPPL prohibits distributing one of them ....
5:55 AM
@cfr Huh?
@JosephWright Probably I don't know how to ask the question.
6:13 AM
@JosephWright Never mind. I'm an idiot. I figured out lying to LaTeX made it easy to test, but I wasn't quite nefarious enough ....
1 hour later…
7:31 AM
@DavidCarlisle, @egreg Spaces ... github.com/latex3/latex2e/issues/1378
Hi guys! am here to ask some questions
@JosephWright lookquitenormalforagerman
@JosephWright hmm not my fault:-) I thought that came up before while testing, presumably we missed one, grrr
@DavidCarlisle Indeed - I'm not even sure why I had the test in siunitx either :)
7:40 AM
@JosephWright amazing we didn't have a test for that in tools
@DavidCarlisle Yes, I thought it was odd
@JosephWright most of the older tests are testing good input gets handled correctly rather than checking error exits, but still....
@QingHongBoon hello
@DavidCarlisle Yes, certainly for new code I always test all of the messages I can
@QingHongBoon Go ahead, but anything 'significant' is best on the site
@JosephWright we can still release tools without base?
@DavidCarlisle Of course
@DavidCarlisle You going to take the bug?
@DavidCarlisle I want to do github.com/josephwright/siunitx/issues/733 first
7:46 AM
@JosephWright unless Frank wants it:-)
@DavidCarlisle :)
@DavidCarlisle probably the tests are in base and so do not catch array changes?
@JosephWright PR to main for a push straight to latex, or to dev ?
@DavidCarlisle PR to main is easiest as then I can merge to both main and dev
looks like doesn't even need ~ just move \ExplSyntaxOff up a line
8:18 AM
@JosephWright damn I branched off main but took default PR to develop, I'll adjust
@DavidCarlisle :)
that looks better @JosephWright
8:30 AM
@DavidCarlisle quack
@PauloCereda breakfast
@DavidCarlisle oh no
8:51 AM
@DavidCarlisle Merged: I'll sort a release
@JosephWright thanks
Meanwhile, can we find someone to translate learnlatex.org into Greek I wonder ...
@JosephWright θα μπορούσα να το κάνω
@JosephWright we have several unfinished translation forks, tr, zn, jp, ...
oh and gr
@DavidCarlisle Yes, true
@DavidCarlisle I've posted on Twitter - let's see if I get any interest
@JosephWright was theere a request?
9:11 AM
@DavidCarlisle No, just reminded me ...
@JosephWright he said he would start a fork but I don't see his name in the list of forks so I'm not sure if that ever started
@DavidCarlisle Tools with CTAN
@DavidCarlisle We should probably clean up a bit ...
I might look over lunchtime
9:28 AM
@JosephWright see skype
Good to see Khaled's Noto Sans Math 3.0 is released (last week but he just mentioned it today) Finally a Noto Sans Math font that is sans serif
9:49 AM
@DavidCarlisle which language is oh?
Must be the language of Brazilian ducks.
@DavidCarlisle Where was that?
10:10 AM
A customer just came across "LaTeX Error: Too many math alphabets used in version normal. " but the same tex file with an exact 1:1 copy of texlive (i hope) runs without issues on my local machine. Is there any tex config option that could explain this?
Stupid question: I'm assuming it is save to pass a seq variable as an argument to a macro. But what is the proper procedure to work with it at the receiving end?
10:23 AM
A: Noto Sans Maths truetype font support for Unicode-maths

خالد حسنيVersion 3.000 of Noto Sans Math has reworked and improved MATH table (which was first introduced in version 2.539), and should generally work better with unicode-math package: https://github.com/notofonts/math/releases/tag/NotoSansMath-v3.000

@JosephWright ^^
@Lupino local math alphabets
@daleif just use any command starting\seq ?
@DavidCarlisle But those alphabets have to be defined in some style files/packages, no? Those should be equal in both systems, so the only explanation would be that they aren't in fact equal, right?
@Lupino yes, something must be different.
@UlrikeFischer thanks, just wanted to make sure. anyoneknow how to diff entire texlive installations -.- ?
@Lupino I would start with a diff of the log-files.
@DavidCarlisle Very nice! I am still hesitant to having the mathematical italic small a and others in sans, when there are sans slots...
10:40 AM
@UlrikeFischer of course. I also wanted to make sure I didn't miss some texcnf parameter to increase the number of possible math alphabets, like you can do for max stack sizes.
10:50 AM
@DavidCarlisle in German there is an abbreviation "ff" for following pages. Is the English "ff" the same?
(I think it's based on some latin, but can't remember)
@Lupino probably not, but even in such a case, I would start with the obvious and check versions in the log.
@Skillmon biblatex uses this: sequentes = {{sqq\adddot}{sqq\adddot}}, in english.lbx
@Skillmon most English publications we produce also use "f" and "ff". It comes from latin "folio", see en.wiktionary.org/wiki/ff.
11:21 AM
@Lupino the handling of math alphabets changed to avoid that message (sometimes) so check the exact patch level of the format
@mickep I think that's definitely the right thing to do though (that was a major problem with the first version, that it put serifed letters there)
@Lupino no it's 16 in pdftex and 128 in luatex you can't change it as the number affects the syntax of \mathchar which has the slots bit-packed into an integer
@UlrikeFischer, @Lupino, @DavidCarlisle thank you very much!
11:43 AM
@DavidCarlisle Let us agree to disagree on that point (I think tha could easily be handled by the typesetting system, since one should in a sans serif font be able to use some serif characters, just as one is able to do the opposite for serif math fonts.) A worse thing: Some variants have themself variants that point to earlier sizes, so there are loops...
@mickep this was me agreeing with someone complaining about the serifs, and I got pinged when Khaled updated the font not to have serifs github.com/notofonts/math/issues/14#issuecomment-1479854432
@DavidCarlisle So it is essentially your fault. :P
@mickep No just that Khaled and I are right and you are wrong, seems a better way to describe it
@DavidCarlisle I was just chatting with Hans about it. We might make a companion font that replaces the serif slots. Then we can set it up to be a math sans font, but where one could get the serif slots with \mathrm and \mathit...
@mickep If I wanted that I'd just define \mathrm to use Noto Serif
11:51 AM
@DavidCarlisle Yes, but not sure the glyphs are the same (side bearings, ...)
@mickep same as what? \mathrm should pick a text font with smaller sidebearings and ff ligatures and stuff.
@DavidCarlisle You are a LaTeX user.
@mickep isn't everyone?
@DavidCarlisle I cannot speak for all...
@mickep but even in plain tex if you use \mathop{\rm diff} x you want diff typeset from a text font with ff ligatures not a serifed math font.
12:08 PM
@DavidCarlisle I think that shall be the choice of the user.
12:25 PM
@DavidCarlisle :/ turned out, they used a different xmltex call and got the wrong sub-layout which loads way more fonts than the intended "classic" layout… to the skript maintainer it goes… (again thanks to @UlrikeFischer for the hint to check the log files of both runs. TeX installations, versions, and paths were the same each, but lucky for us, we pass the cli call down to the log, so finding the culprit was pretty easy from there)
(well, not "we" pass it down; tex does that on its own, i believe)
1:12 PM
@DavidCarlisle Ta.
1:31 PM
@DavidCarlisle \seq_set_eq:NN \l__foobar_seq {#1} seemed to work ok
@daleif should not have {} for an N argument (although you can usually get away with it)
@DavidCarlisle then I need a variant, there is no Nn version
@daleif no just remove the braces
@DavidCarlisle why does that work?
N is for a single token, and who knows what #1 is here
@daleif well if #1 is not a single token, then you should not use it there.
1:35 PM
@daleif why would it not work? presumably you have arranged that #1 is a single token so it's like doing \let\foo#1 where \let\foo{#1} would be the wrong thing.
@daleif If #1 is not a single token then \seq_set_eq:NN \l__foobar_seq {#1} will die horribly anyway.
I just saw that as a data structure, so something more complicated that one should be more careful with
@daleif but all commands with names like ..set_eq: are really (or conceptually) \let so they require two tokens and one is defined to the other, you can not givethem arbitrary content
@DavidCarlisle all I'm looking for here is how for a macro to handle \foo{\l__foobar_seq}, where \foo should be able to work in the contents of the seq but not be able to destroy the original (as in arg copy vs arg ref in other langs)
@daleif then \seq_set_eq:NN \l__foobar_seq #1
1:58 PM
Noted. BTW is there a N variant of \seq_pop_right:NN, that is I just want to pop the right most element, but not do anything with it, or just throw it into a tmp var
@daleif No, you always need a 'dump' variable
ok, noted
We'll see if these rewrites work. The MWE did, so ..., we'll also see if there are any performance issues, but I do not think so. As the lists are small
@DavidCarlisle did you see this new package spelatex?
@UlrikeFischer no
@DavidCarlisle interesting but nothing works ...
2:13 PM
@UlrikeFischer speaking....
@DavidCarlisle ?
@UlrikeFischer spelatex
@DavidCarlisle sorry is there some pun I don't get? ;-).
@UlrikeFischer no I was just confirming I looked up the package after your ping and that it's about generating speech
@DavidCarlisle ah. I tried their example.pdf but while it starts the media player it then claims that I'm not allow to access a file. And their example document doesn't compile, it complains on ! Missing $ inserted. for every section. And I had to create a folder for their output files first.
2:33 PM
@UlrikeFischer I just looked quickly at the doc on ctan, that did play a reading of figure 1 at one point
@DavidCarlisle which pdf reader did you use?
@UlrikeFischer firefox
@DavidCarlisle oh, so inside the browser?
@UlrikeFischer yes I was just browsing ctan page clicked on doc link and was paging through the pdf quickly and at some point it started talking
@DavidCarlisle the figure works for me too, but not the small triangles (which are by the way the source of the math errors ...).
2:40 PM
@UlrikeFischer as I say I wasn't really looking, I just noticed it started talking:-)
@DavidCarlisle apart from the triangles it now speaks to me too. But basically everything is an active area in this document, as soon as you click you hit something ;-)
2:59 PM
@JosephWright Could you please have a look at github.com/josephwright/beamer/pull/898 ? Can you think of a reason to keep the inconsistent spacing for right-aligned frame titles?
@samcarter Will do a bit later
@UlrikeFischer meanwhile you broke arydashln
@JosephWright No hurry at all!
The code has been there since 10 Nov 2004, a couple of days, weeks, months or years won't matter too much :)
@DavidCarlisle ?
Q: LaTeX2e 2024-06-01 conflict with arydshln package?

scottkostyThe following example now fails to compile for me: \documentclass{article} \usepackage{color} \usepackage{colortbl} \usepackage{arydshln} \begin{document} The following table... \begin{tabular}{|c|c|c|} \hline \textcolor{green}{a} & \textcolor{red}{b} & \textcolor{red}{c}\tabularnewline \hline ...

3:08 PM
@DavidCarlisle well that is easy, as soon you move you break something there ;-(
2 hours later…
5:21 PM
Working a bit on xbeamer
3 hours later…
8:25 PM
@JosephWright working on trying to fix arydashln :-)
8:54 PM
@DavidCarlisle any progress? It looked as if it would die while processing the preamble.
@DavidCarlisle ah, just saw your answer.
9:06 PM
@UlrikeFischer you and Frank have mail:-)
@DavidCarlisle yes saw it. I don't have many example, only somewhere one where it broke lines in a longtable .... But I did read a bit in the dtx and it looks as if the main reason for some of the hacks is to count the columns and we are already doing that now, so perhaps they can reuse that.
@UlrikeFischer yes I doubt he'll want to rewrite it to use the new features though
@UlrikeFischer what surprised me is I didn't think we were expanding the preamble more than previously, so I didn't see exactly why it broke, may look again later
10:00 PM
user image

« first day (4975 days earlier)      last day (29 days later) »