« first day (4915 days earlier)      last day (40 days later) » 

7:10 AM
user image
6
@PauloCereda ^^^^ found in an issue ...
 
 
2 hours later…
8:45 AM
@UlrikeFischer more improvements from the ua/2 examples:-)
 
@DavidCarlisle ? I see nothing new?
 
@UlrikeFischer same examples but they validate faster, see secret email
 
@DavidCarlisle wow ;-)
 
@DavidCarlisle So, the validator got updated instead of the examples? :)
 
@mickep yes bible validation goes from 40min to 30s so noticably faster
 
8:59 AM
@DavidCarlisle Oh, with what validator? vera?
 
@mickep yes (next release)
 
@DavidCarlisle Oh, future bibles are so quick.
@DavidCarlisle Now you only need to push down the compilation time as well.
 
@UlrikeFischer ooh
 
@mickep use context instead of latex, obviously
 
@DavidCarlisle Oh, but the contexters are using the much faster King James bible. Not comparable.
I even asked @PauloCereda about that.
 
9:11 AM
@mickep I used that first, but I would be locked in the tower for using that, as I'm british and the Crown has extra rights
 
@DavidCarlisle Haha! Risks one has to take.
 
@mickep wasn't prepared to take that one, hence the ASV
 
@DavidCarlisle I assume that the version does not have a big impact (unless they skip the old part...)
 
@mickep well the original source forms I got were a bit different so I had to adjust things a bit, and I had hand tweaked a few bad page break \looseness=.. here and there to keep last lines of books falling to a new page, so it was a bit of work to get the sources aligned, but the basic tex markup was more or less the same
@mickep the real choce is not whch version you use (which just affects some words in some places) but whether you include all the footnotes or frontmatter or other commentary that are in some versions I didn't do that here (and the plain text sources I picked up had already stripped them) It might be fun to try to fully tag one with multiple foot and side notes
@mickep frontmatter of the 1611 KJV more interesting than the main text structurally:-)
 
9:30 AM
@DavidCarlisle Hm, I do not remember the XML versions we found having footnotes/other notes. But if they had I guess it would just have worked.
Regarding the format, there are some interesting words that will not be hyphenated, but that seems to be the "main problem" if one has a decent page setup.
@DavidCarlisle Oh, you found the context output? :P
 
10:04 AM
@mickep yes I had to add a few hyphenation exceptions, but not so many, and I only looked when tex gave up and made an overfull box, I didn't read the whole thing looking for linebreaks that could be improved
@mickep the KJV has lots of sidenotes:
@mickep and some interesting exercises in metapost for you:
 
10:20 AM
@DavidCarlisle Oh, they seem to be missing from the XML :( I must say that I'm impressed by their typographical work on these bibles, though.
 
@mickep yes that was my thought too, they could do that in 1611, we haven't really improved...
 
@DavidCarlisle I wonder if the annotations are a bit funny, like those in Concrete Math.
 
 
2 hours later…
12:00 PM
Thanks @DavidCarlisle, 3x faster here (not 80x)
 
12:28 PM
git bisect FTW
 
12:50 PM
@mickep on your document, you mean? or you don't see that improvement on ours?
 
@DavidCarlisle On the context one.
 
@mickep yes there seems to have been some specific construct that triggered the behaviour previously the time for that one was completely unlike the times for all the others. Still, even 3x isn't a bad improvement
 
cis
@JohnKormylo @John Kormylo Could you maybe make a command from your code?
-----> https://tex.stackexchange.com/q/715634/46023
 
1:06 PM
@DavidCarlisle Indeed, thanks!
 
1:19 PM
@cis if the @ ping does not auto-complete then it will not work, you can only ping people who have been in chat recently.
 
1:49 PM
@DavidCarlisle Hi, one question (just out of curiosity), what is the purpose of the file/texlive/2024/texmf-dist/tex/latex/latexconfig/lualatexiniconfig.tex?
@DavidCarlisle I have run grep -rn and it is not loaded by anything usable (only lualatex-doc)
 
@PabloGonzálezL No longer used
 
@JosephWright Ok, I understand, so my impression was correct :D
@JosephWright I was struck by another thing... In that directory /latexconfig/ is located epstopdf-sys.cfg ....but I got lost with the name (in my mind it should be epstopdf-base)?
 
 
1 hour later…
2:58 PM
@mickep ooh
 
@PauloCereda Oh indeed
 
3:12 PM
Brining back a feature from 2009 ...
 
@JosephWright Oh, that is old stuff
 
@mickep sounds pretty modern to me
 
@mickep Turns out it was a good iddea
@DavidCarlisle \DelayEvaluation ;)
 
@DavidCarlisle Hehe, well, that depends.
 
3:31 PM
@JosephWright sounds vaguely familiar but i don't recall what it does, it sounds like \unexpanded
 
@DavidCarlisle ooh that means you are old
 
3:54 PM
@PauloCereda You are confusing me with that Italian bloke
@PauloCereda albatross question: tex.stackexchange.com/questions/715453/… I thought of responding "how can you trust software written by a duck...."
 
@PauloCereda `$ uname -a
Linux rayzentex 6.8.5-301.fc40.x86_64 #1 SMP PREEMPT_DYNAMIC Thu Apr 11 20:00:10 UTC 2024 x86_64 GNU/Linux`
 
@DavidCarlisle Turned out we had already sorted it :)
 
4:23 PM
@DavidCarlisle oh no
@PabloGonzálezL go away
@PabloGonzálezL <3
@PabloGonzálezL how do you like it? We are less than a week for F40 stable!
 
@PauloCereda I can't complain, it's stable and has no crashing problems...as always fedora doesn't disappoint me :D
 systemd-analyze
Startup finished in 6.003s (firmware) + 4.245s (loader) + 1.431s (kernel) + 6.502s (initrd) + 5.427s (userspace) = 23.611s
graphical.target reached after 5.404s in userspace.
 
@PabloGonzálezL cool! I cannot complain either
@PabloGonzálezL are you using flatpaks? Just out of curiosity
 
@PauloCereda You are on flapjacks...
 
@mickep oh no
@mickep you are mean
 
@PauloCereda oh no
 
5:09 PM
Hm, re-reading about protected commands ... I think, I should implement one macro in jsonparse differently: \JSONParseGetValue has a starred variant. The starred command typesets things, so should probably be protected. The non-starred command may be used as input to parsing again (especially if it outputs an object). But if I understand things correcty, I should better provide a separate command (not protected) for the latter and change the former command to be protected?
 
@JasperHabicht Yes
 
@JosephWright Thanks! This is a concise and nice answer =D But the longer I think about it, the more I also come to this conclusion
And for terminology, would \JSONParseGetExpandableValue a reasonable name?
 
@JasperHabicht The team tend to use get in things that carry out an assignment, so I'd favour just \JSONParseValue
 
@JosephWright Actually, the function names are a bit of a mess currently ... JSONParse is also the name of the package, so the parsing command should be named \JSONParseParse if we want to be precise ...
 
@JasperHabicht :)
 
5:30 PM
@JosephWright What about "use"? There is a command named \JSONParseUseArrayValues which can be used to plug each value of an array inside a function. I guess, "use" is also used differently? =)
 
@JasperHabicht Well at the expl3 level that's normally item in that case, use means 'insert an entire variable', e.g. \dim_use:N
 
@mickep you are not mean
 
@JosephWright Right. So, I should think of another name for this as well. The command iterates over all items actually, so item would not be the right word.
 
@JasperHabicht That's map
@JasperHabicht Can be use with a multi-entry thing if it's like say \clist_use:Nn
 
@JosephWright Yes! It is map ... I will use map. Thank you!
 
5:55 PM
Last week I received a question whether one of my LaTeX packages could be used in Word...
4
 
@Skillmon Which package, go on, tell us!
 
@JosephWright xistercian :D
 
@Skillmon That I would not have guessed
 
@JosephWright me neither, but apparently someone wants them to be part of their dissertation.
 
6:14 PM
can tabularray span mutiple pages pls help
 
@FerriusUndermine Yes, with longtblr for example.
 
6:59 PM
\str_casefold:n seems only to exist as of TL24 but "Interfaces" says that one should not \str_lowercase:n for caseless comparisons ... What would be the correct way to support TL23 then?
Ah, there is also \str_foldcase:n ... seems to still (?) work, but is no more documented
Found it. So, would it be save to do \cs_if_exist:NF \str_casefold:n { \cs_new:Npn \str_casefold:n { \str_foldcase:n } } ?
 
7:27 PM
```
```\begin{longtblr}{
colspec = {Q[c, 1.4cm] Q[c, 1.4cm] Q[c, 1.4cm] Q[c, 1.4cm] Q[c, 1.4cm] Q[c, 1.4cm] Q[c, 1.4cm] Q[c, 1.4cm] Q[c, 1.4cm] Q[c, 1.4cm]},
hlines,
}
& Appearance & Smell & Water Soluble & Ph & H2SO4 (Sulfuric Acid) & HCl (Hydrochloric Acid) & AgNO3 (Silver Nitrate) & NaOH (Sodium Hydroxide) & BaCl2 (Barium Chloride) \\
NaCl & & & & & & & & & \\
CaCl2 (Calcium Chloride) & & & & & & & & & \\
NaHCO3 (Sodium bicarbonate) & & & & & & & & & \\
CaCO3 (Calcium Carbonate) & & & & & & & & & \\
i have this code, and it produces an hbox too wide error 4 times
why exactly could this be, does the error matter at all
 
News from the minimum image size problem:
@Shadow Not sure where the tone is coming from, but I am not claiming images larger than 160x160 are somehow inherently safer (if I had to guess I'd say the opposite is probably true - unsafe images are probably more common above 160x160). It's that the detection method we are using may need images to be larger than 160x160 in order to classify them appropriately. But we need to take time to investigate those requirements (and other options) more thoroughly. As far as what Imgur is doing - well, I can't tell you that. I don't work there. ;) — Slate ♦ 15 mins ago
 
@FerriusUndermine This is not an error but a warning (just an information even) that things are too wide for a box and will therefore be typeset in a way that is probably not intended (text will overlap to the right, typically). The problem is that TeX does not know how to break strings such as C12H22O11 which are too long for the column you try to typeset into. Try a smaller font size maybe.
 
@FerriusUndermine You need to make your columns larger. If I were you I would split the table into two, the first having first 4 columns to the right of the first, and the second having the last 4 columns. You could make the font smaller as @JasperHabicht suggests, but that's generally not really done (unless all tables are deliberately in a smaller font).
 
7:49 PM
@JasperHabicht We had a rename so the older one is deprecated - we remove those from the docs
 
@JosephWright Yes, but if I put only the new one in the package code and load the package with a TL23 installation (e.g. on Overleaf), it complains. I would like to use the new one of course. It is just for backward compatibility.
 
@JasperHabicht All quite true: depends on whether you expect people to use your code other than installing as part of their TeX system, and to what extend you then choose to support that
 
8:05 PM
@JosephWright Sure, I will need to set some kind of limit. I just thought that supporting TL23 would be a good idea. And I wanted to just make sure that the above line of code would not be dangerous in some way ...
 
"Don't expect an immediate update here, but know that we're aware of the issue and are evaluating approaches to addressing it. Thanks for your patience." this is blocking a large number of images posted to tex.stackexchange unpostable, it is very common to need to show the output of one or two lines of output text and almost always if you select a region and paste in you hit this limit and have to repost with pointless white space. Please could the restriction be removed while you evaluate options rather than making the site so difficult to use while you do the evaluation? — David Carlisle 15 secs ago
 
@DavidCarlisle The powers that be should sent every user a high-definition-ultra-high-k-retina-insert-more-buzz-words-display and even a simple screenshot of just a couple of symbols will have 1000 times 1000 pixel -- problem solved :)
 
Why should images even be handled differently than other user content? As far as I know there is a spam button, and if someone posts offensive pictures (such as of pineapple pizza or roasted ducks or both mixed) then such contents can be deleted easily. Well ... whatever ...
@samcarter Yes. Next step is probably "Sorry, your answer is too short [to be parsed by our AI-powered offensive content filtering mechanism] please add a few more lines [probably using another AI]" =D
 
@PauloCereda The truth is that I use little or nothing :(
 
@JasperHabicht Yes! That's how evolution works :)
 
8:20 PM
Why can't I pass a \msg_new:nnn message declared later in a code to \IfPackageLoadedTF?
\documentclass{article}
\usepackage{multicol}
\ExplSyntaxOn
% \msg_new:nnn { mymodule } { pkg-load } { The ~ '#1' ~ is ~ already ~ loaded. }
% \msg_new:nnn { mymodule } { pkg-not-load } { The ~ '#1' ~ is ~ NOT ~ loaded. }
\IfPackageLoadedTF { multicol }
  {
    \msg_info:nnn { mymodule } { pkg-load } { multicol }
  }
  {
    \msg_info:nnn { mymodule } { pkg-not-load } { multicol }
    \RequirePackage { multicol }
  }
\msg_new:nnn { mymodule } { pkg-load } { The ~ '#1' ~ is ~ already ~ loaded. }
 
@samcarter Or they could just interpolate the missing pixels using some AI to make images that are too small to parse larger. Problem solved.
@PabloGonzálezL Why should that work and why would you want to do that? You define the message after you call it ... The code is processed from top to bottom.
 
@JasperHabicht Let's replace all too small images with cute ducks :)
3
 
@JasperHabicht Most of the messages I use I document at the end (just for that reason) and you don't necessarily have to have them stated in that order.
 
@PabloGonzálezL Well, I think you do need to have the messages defined before you call them, but in some cases, for example if you place \msg_info:nnn inside a function, you can define the message later, because the function is called only later in the document, so still after the message has been defined. But \IfPackageLoadedTF is called immediately. At least this is what I assume here
 
@JasperHabicht I'm used to declare them at the end (instead of at the beginning), but you're right...maybe that's the reason why it doesn't work for me in this case...snifff
 
8:32 PM
@JasperHabicht I think that images and text are very different things to automate content control over. You can (relatively) easily scan text for content you don't like but images are much harder to automate.
 
@AlanMunn This is, of course, a good point.
 
@PabloGonzálezL Use DocStrip to produce 'live' code in a different order to your sources
 
@PabloGonzálezL why would you expect \msg_info:nnn to act differently to every other tex command? It has to be defined before it is used.
@PabloGonzálezL at the point that you are making the message, \msg_new:nnn { mymodule } { pkg-load } { The ~ '#1' ~ is ~ already ~ loaded. } has not even been read from the filesystem and parsed by tex, so naturally the message is not defined.
 
8:52 PM
@JosephWright It is an option...mmm..but it would make my life more complicated :D
@DavidCarlisle The truth is that I used to have these things written with wlog and I decided to modernize the code :( ...and I use this way the messages for the functions written in expl3 ...I thought that the same thing would happen.
 
@PabloGonzálezL the same thing does happen but \wlog is defined before it is used
@PabloGonzálezL normally you would define the message in your package but do the test \AtBeginDocument to allow multicol to be loaded later and so the problem doesn't arise.
 
@DavidCarlisle That trick works...any problem using the \AtBeginDocument hook?
 
@PabloGonzálezL no, that's what it's there for. I wouldn't call it a "trick" it's the intended form
@PabloGonzálezL but why have the test at all, why not simply have \RequirePackage{multicol} that is already doing the same test.
 
@DavidCarlisle I never stop learning...thank you very much :D
@DavidCarlisle I think I once read here in the forum that it was not a bad idea to load the package only if it was not already loaded in memory.
 
@PabloGonzálezL \RequirePackage and \usepackage never load a file twice, they test if loaded and skip input if the package is loaded already.
@PabloGonzálezL yes but that means using \RequirePackage instead of \input. That's what it does.
 
9:07 PM
@DavidCarlisle Mmm...so your recommendation is to just use \RequirePackage?
@DavidCarlisle And related...what is the correct form \input <filename> or \input{<filename>} at .sty file level?
 
@PabloGonzálezL Braced
 
@JosephWright Thanks... 2.7.1 Loading other files in clsguide need a mention :D
 
9:55 PM
@PabloGonzálezL yes why not.
 
@PauloCereda dinner
 
@DavidCarlisle oh no
 
 
1 hour later…
11:26 PM
What is the best way nowadays to have an environment where things are not carried out in a local scope, no environment-hooks interfere, you don't need to have the body of the environment collected as a whole?
 
@UlrichDiez why use an environment? \def\foo{\endgroup}\def\endfoo{\begingroup} I guess
 
@DavidCarlisle The decision on this is not on me. I guess it is because the concept of environments exists and people are sort of used to it and to abusing it. ;-)
 

« first day (4915 days earlier)      last day (40 days later) »