« first day (5189 days earlier)      last day (40 days later) » 

cfr
cfr
00:51
@Jasper FWIW I would only use LPPL for latex-related stuff. I'd prefer GPL. but it is easier if latex-related stuff uses LPPL. then you don't (mostly) have to worry about licence compatibility.
@DavidCarlisle I understand that. but, in the case the OP has presumably found it already since set is from the OP's comment. and I seriously doubt anybody else is going to be interested in my answer.
01:14
@Jasper I do the same as @DavidCarlisle LaTeX stuff LPPL but any other code, MIT.
 
2 hours later…
03:07
@cfr I unfortunately still cannot get the hvmath fonts to work. My output of `latex hvmi10test.tex` is
```
This is pdfTeX, Version 3.141592653-2.6-1.40.26 (TeX Live 2024) (preloaded format=latex)
restricted \write18 enabled.
entering extended mode
(./hvmi10test.tex
LaTeX2e <2024-11-01> patch level 1
L3 programming layer <2024-12-25>
(c:/texlive/2024/texmf-dist/tex/latex/base/article.cls
Document Class: article 2024/06/29 v1.4n Standard LaTeX document class
(c:/texlive/2024/texmf-dist/tex/latex/base/size10.clo)) (./hvmath.sty)
Output of `set TEXINPUTS=.;dpi300;;` is `dpi300: The term 'dpi300' is not recognized as a name of a cmdlet, function, script file, or executable program.
Check the spelling of the name, or if a path was included, verify that the path is correct and try again.` I guess `;` isn't appropriate here. I don't an error with `:` though.
The output of `set TEXINPUTS=.:dpi300::
>> dvips hvmi10test.dvi` is
```
This is dvips(k) 2024.1 (TeX Live 2024) Copyright 2024 Radical Eye Software (www.radicaleye.com)
' TeX output 2025.01.14:0834' -> hvmi10test.ps

kpathsea: Running mktexpk --mfmode ljfour --bdpi 600 --mag 1+0/600 --dpi 600 hvrm10ex

The command name is C:\texlive\2024\bin\windows\mktexpk

kpathsea: Running mktexmf hvrm10ex.mf

The command name is C:\texlive\2024\bin\windows\mktexmf
name = hvrm10ex, rootname = hvrm10ex, pointsize =
dvips wants the metafont sources for some reason, which have not been provided
btw, how do you type code blocks here? enclosing it in ``` does not seem to work
 
3 hours later…
cfr
cfr
05:52
@ApoorvPotnis sorry, but what is actually in hvmi10test.tex or where is it from?
@ApoorvPotnis you have to paste nothing but code into a comment. then one of the buttons on the right turns into use fixed font or something like that. you click it and it sets the comment as code. the code must be at least 2 lines. I have no idea why it is this way ;).
@ApoorvPotnis try chat.stackexchange.com/transcript/message/66995089#66995089. that is certainly a better suggestion than mine. (my real solution would say: install linux. but you might think that is overkill just to view some fonts.)
06:10
@JosephWright forgot unit tests for \int_step_tokens:nnnn. Bad?
06:32
@DavidCarlisle Because that's, how I read it. I thought, I can't be the only one.
@MaestroGlanz One * too many. TeX MetaFont.
07:14
@Skillmon I'm on it
07:37
@JosephWright thanks :)
08:06
@DavidCarlisle ooh CDL
08:35
@Skillmon Dev release to CTAN today, so in the main release by the end of the month
yo'
yo'
08:46
@mickep Yeah. It's "fun", but suddenly one day you wake up and all non-tagged PDFs are obsolete and for formal purposes irrelevant/unacceptable. In this sense *TeX is on the risk of getting completely forgotten, if not for the work of the accessibility team!
@yo' I see no reason whatsoever that will happen.
@JosephWright uhh, then I need to update the answer that it's already in -dev :) Btw. I thought L3 doesn't have -dev, only the 2e level.
@Skillmon l3kernel-dev goes out to be built into pdflatex-dev, etc., and if there are no issues after a couple of weeks I send the same as l3kernel to go into pdflatex
@JosephWright ah, ok. Thanks for the explanation :)
@Skillmon After the issues when we extended prop internals, FMi in particular is keen to get some testing outside of the suite
09:04
@JosephWright evil prop internals!
@ApoorvPotnis that's a windows error you need quotes around the path, also you want TEXINPUTS to end in a ; otehrwise you stop latex finding article.cls
@Skillmon Indeed
@Skillmon It's a bit frustrating .. my feeling is that was a one-off
@mickep Do you and Hans want to lose another day looking at tfm font metrics? tex.stackexchange.com/questions/734953/…
@yo' I assume you have no (La)TeX alternatives for Overleaf, for example, do you?
@DavidCarlisle Oh, very difficult to motivate to look at the stix fonts, even more so for stix type1... :)
@mickep I thought you would be thrilled
09:15
@DavidCarlisle As far as I remember, even the OpenType fonts are a bit strange. Maybe Xits was a bit better than Stix.
@mickep stix had no Math table so couldn't be used as an opentype math font, Khaled added one and and added some other fixes to make xits
@DavidCarlisle Ah, so then they can try xits instead, if they dare to run luatex and opentype math...
@mickep yes which is what I just put in a comment:-)
@DavidCarlisle So you are postponing your closer tfm study this time. To after the tea break.
@mickep my plan was to postpone it until whenever you find convenient.
09:22
@DavidCarlisle I'm just making tea. But I also need to write a review report...
09:36
@JosephWright until we change the internal representation of the next datatype :P Though afaict the change to single tokens for flags was not as problematic (most likely because they weren't used)
@mickep it's an almost documented feature:-) Looks like on loading you copy the numeric fontdimen to the named opentype-ish parameters \Umathsubshiftdown and so if changing them later you need to change the named ones
yo'
yo'
@mickep Yeah, no reason whatsoever, because things like this don't happen at all: accessibility.works/blog/…
(it's just an example, the thing is big in the US now)
@DavidCarlisle Oh, so a bit the opposite of what xetex does for opentype fonts, iirc.
yo'
yo'
@mickep That's a fair question that I choose not to answer, sorry...
@mickep yes because I realise you try to make it easy for latex to support cross-engine code:-)
yo'
yo'
09:42
IIRC this talk by Jeff Kuan covers the topic quite deeply: youtu.be/e71-lIhJnZk
@yo' For web pages, I understand it better. There are also other things that are big in the US that we do not need to import. (Yes, I know that the EU is already on it...)
@DavidCarlisle I did not know that luatex did map to the opentype parameters. But, if you identify as a opentype engine, I think that makes more sense than doing the opposite.
@mickep you could move to the UK:-)
@DavidCarlisle Haha, well, I would be surprised if there is nothing like this going on there as well? :)
First siunitx and l3kernel-dev of 2025 sent to CTAN
@Skillmon :)
@Skillmon Unless we do linked lists for seq, seems unlikely
09:57
@JosephWright unlike for prop there doesn't seem to likely be a good reason to switch to linked lists for seq (except for very large sequences which could be split every $n$ elements or so)
@JosephWright Lua tables for prop ?
@DavidCarlisle /panic
@Skillmon Quite
@DavidCarlisle Possibly, although I'm not sure how easy it would be to manage catcodes
@DavidCarlisle catcode and grouping nightmare. Though, iirc, someone had a working PoC for them
@Skillmon Quite
10:00
@JosephWright catcodes are the lesser of the two problems though.
@Skillmon Fair enough - neither aspect though is natural
10:24
@mickep web accessibility is not only about html pages. You can not hide content e.g. in a doc or a png or a pdf, distribute that over the net and claim web accessibility doesn't matter here. If you put something on a server and add link to it on a web page it is a part of the web.
@UlrikeFischer Yes, of course.
But if one does as arXiv (ar5iv), providing html versions of the pdf files as well, everything is fine, as long as it works OK.
yo'
yo'
10:54
@mickep Yeah, as long as they are actually accessible.
@yo' Maybe someone inhere has tried to use a screenreader on their html output. How does it work compared to doing the same on a tagged pdf?
yo'
yo'
@mickep I'm certainly not the one to answer that question.
11:24
@mickep yes as long as it works ok. So try to read such articles on a flight or in other places where you have no good internet access. You can download articles as pdf and take them with you. What should blind people do?
@UlrikeFischer Just save the file. It is essentially text, and that can be saved... If the reader cannot read offline, then it is not much of a help.
@UlrikeFischer Look, I do not complain that you do tagged pdf for LaTeX, you do a great job! I just say that I do not think tagging pdf is the right way to go, at least for ConTeXt. It makes more sense (for us) to export to html/xml.
@mickep well did you ever try to save and store and read later offline in a sensible way html pages? Pages that e.g. call complex javascript (e.g. mathjax), or rely on your session cookie or load webfonts or ...?
@UlrikeFischer Mathjax or webfonts are not so important if the html/xml/mathml source is there. The reader should not care for that, should it? We shall not make it difficult for the users.
(I guess a blind person could not care less about what font a document is using)
@mickep So you mean you won't have a problem to replace for a week all your local pdf uses by (local) html uses instead?
@UlrikeFischer If I was blind and I wanted to read something, I assume I would not mind to use html if that worked better for me. It is difficult for me to imagine that situation, though.
11:52
@mickep you don't need to imagine. You could simply try a bit. Avoid pdf if you can. You don't need to use a screen reader (they are scary), simply store html when you would store a pdf, try to read it locally when you would read a pdf. If there is no html equivalent, ask for it, or generate it with context and use that and so on.
12:20
I've no idea of what's going on, but the answer is clearly vim.
@DavidCarlisle ^^
@PauloCereda poor fella, is missing his meat on his diet.
@Skillmon ooh
12:37
Yay, after quite some time I managed to be in the "watch your health" area of reputation earnings (aka >200 daily rep) again. For @egreg that means everyday, 8am.
5
13:01
@Skillmon Yesterday it was 210 just because of –50 due to user “what does TeX should mean” leaving. ;-)
@Skillmon I can restrain from upvoting you, if you so prefer.
@egreg ooh no! :)
@egreg yesterday it was 10 (though for the same reason of user leaving)
@UlrikeFischer No, I'm happy continuing to look at the pdfs as long as I still can.
@DavidCarlisle Oh, you have another secret mail.
14:19
@mickep secret reply
14:31
@DavidCarlisle now I'm all @PauloCereda and want in on that secret secret.
@Skillmon but then it wouldn't be a secret
@DavidCarlisle what is the point of a secret if you can't tell everyone in secret what the secret is?
4
15:05
@UlrikeFischer The point is of course to tease @PauloCereda. He is now furious.
15:23
@Skillmon I simply had a question/comment about spacing in a document (not mine). Not more fancy than that.
@mickep luckily you pinged me and not someone else, so someone else (like @PauloCereda for instance) still has no idea what the secret secret is.
15:39
@mickep you could tease a small snippet of secret that it's all @barbarabeeton's fault. But maybe you should keep that secret.
@DavidCarlisle I haven't seen her after her question on reading habits. I don't want to scare her away.
@DavidCarlisle And she might be reading something on the building of paragraphs, so I don't want to annoy her either. :)
16:05
About to start the .cls for the upcomming Duden volumes and I decided to finally transition to Expl3. Is there a expl3 version of the clsguide?
@Lupino no, and there are no class-specific functions in L3. The go-to document is interface3.pdf
@Lupino oh, and: evil Duden! :)
^^ why evil?
@Lupino I don't agree with some (very few) things that end up in the Duden, that's all :)
@Lupino had no idea that they're using LaTeX, btw. Very interesting to hear.
@Skillmon oh, well, with some 300k entries, there will always be some questionable things for everyone.
@Skillmon they don't, until now. They used to use something proprietary, but now switch to LaTeX.
@Lupino ah, ok. May I ask how you got in on that?
16:19
@Skillmon the company I work for won the tender
@Lupino <insert obvious follow up question here>
@Skillmon unfortunately, I can't tell too much because of NDA-foo
@Lupino the obvious follow up question was which company (I'd never try to squeeze information out of you that's behind an NDA)
@Lupino ahh, heard about them once or twice (but obviously had no idea -- and if I ever had I forgot about it -- that you're working there)
16:30
@Lupino I don't know how much you are allowed to say about the texnical aspects, but I'd love to hear about the process to create the duden! Have you considered a talk at Dante?
@Skillmon yeah, there were some things in the past...
@samcarter i actually did consider that, but made no final decision, yet. also depends on how much the publisher allows me to tell.
@Lupino I'd love to hear that talk as well (if peer pressure helps convincing the publisher...)
@Lupino Oh good, so I can keep hoping :)
Potato internet of doom: Pushing to Github with 220 Kibit/s
Uh, it's picking up its pace, now at 235 Kibit/s
@Skillmon Can you please translate to bauds?
16:37
@Skillmon i also noted recently that github is terribly slow, despite a 1Gbit connection…
@Lupino hm, culprit could be 3 things for that particular upload, potato internet at home, potato connection to corporate VPN, or potato servers at Github.
@mickep is that serious?
@Skillmon my guess is the latter…
@Skillmon Will you eat potato salade tonight?
@Skillmon Not really
@Skillmon more probable guess is that all input is routed through some "AI" training facility first…
16:47
@mickep phew, I was afraid there for a second
@Skillmon Dodos are not so quick...
@mickep well, my dayjob has quite a lot to do with bauds (since I'm programming some CAN-speaking SW)
@Skillmon Ha! Lucky coincidence.
@Lupino I hope that this would be illegal (the repository is a paid private corporate one, but hosted on their server)
@Skillmon you seem lucky: "Does GitHub use Copilot Business or Enterprise data to train GitHub’s model? No. GitHub does not use either Copilot Business or Enterprise data to train its models. " (github.com/features/copilot#faq)
16:57
@Lupino Cool
@Skillmon @DavidCarlisle ooh secrety secrets
@Lupino xmltex?
@DavidCarlisle no, unfortunately not. It would be better, thou, given the data come from XML. But upper management want to show-off their xslt capabilities.
@Lupino shame it would have been good if all the Germans were relying on me to know correct German.
6
17:18
@cfr I removed my comment regarding the \` vs \par`, thanks for pointing it out. Another thing: My user name is spelled with an "o" not an "a" (but keep the comments, it doesn't annoy nor insult me or anything, I'm totally fine with it, everybody knows who was meant)
@Skillmon Skillmon is a skill man
@mickep it's a leftover from my not so bright teenage years :)
17:35
@mickep ooh mr. rabbit
17:50
hm, I wonder what it'd take to extend LaTeX's output routine to allow for three columns…
cfr
cfr
@Skillmon now I feel awful: why on earth didn't you tell be before?
 
1 hour later…
19:14
@cfr you really don't have to, and I didn't because I noticed just today
As I said, it doesn't bother me at all.
19:28
@mickep -- what she's doing right now is trying to gather all the documents needed to change the registration of her deceased husband's car, because the registration runs out at the end of this month, and if I can't reregister it in my name, I'll be reduced to begging rides or walking. The original of the one I need the most was mailed to me in mid-December, but hasn't yet been delivered. The lawyer has requested new sealed copies, but the state is not quick.
@barbarabeeton Oh, I'm sorry to hear. I hope it works out in time!
@Lupino -- I suppose it wouldn't be proper for me to include two sentences on this in my next TUGboat column?
@mickep -- Thanks! So do I!
 
1 hour later…
20:40
@DavidCarlisle ... so you are now the hunter of spurious spaces?

« first day (5189 days earlier)      last day (40 days later) »