@JosephWright This should be resolved now. GitHub had trouble finding the cache since some paths were represented differently which got added to a hidden part of the cache key.

@MarcelKrüger thanks

@MarcelKrüger Great thanks!

Quick question: with tlmgr update --all I get the message Local TeX Live (2023) is older than remote repository (2024), which is obviously true, as I haven't installed TL2024 yet. But I would like to update TL2023 to the last (frozen) version. Ideas?

@campa the simplest way is to do that last week, after the freeze and before the switch, otherwise you need to switch the repository URL (tlmgr option repositiory ....) to the URL of an archived 2023 tlnet (which exist at a URL I forget)

@DavidCarlisle Ah, so I just need to get my time machine out of my cupboard... can do :-)

Anyway thanks for the hint

@campa thank me this time next year:-)

@yo' Beginning of July I have a flight from Prague. Hopefully the sculpture is still there by then. Want to see it with my own eyes.

@AlexG Beginning of July is too early for TUG'24 :P

@samcarter ^^ this!

Hm hm, I wonder why they (debian) make auctex depend on a lot of texlive packages.

(@DavidCarlisle Not that I really want to use emacs, of course...)

@mickep for the preview feature?

@DavidCarlisle It was long time since I tried it, so I just wanted to see what it was now. But maybe I can equally well live without it.

@mickep I refuse to learn about any features added after 1990 anyway

@DavidCarlisle Oh, yes, that does not surprise me.

@mickep any interseting new texlive 2024 features to try at texlive.net/run?context ?

@mickep did you see that there is a VS Code addon providing native dviluatex rendering (dvi straight to the VSCode canvas, not via conversion to pdf) ?

@DavidCarlisle You can try this: shorturl.at/oKY35 (I had to shorten the url, hope it works). Uncomment the limitstretch line.

\setupalign[depth] is essentiall \flushbottom. With limitstretch, the (positive) stretch can be limited, not to get too results in some situations.

@DavidCarlisle Oh, so now we go back to dvi. Or just to be quick in the editor?

@mickep I think so, question came up here and I thought user was confused and I asked why dvilualatex and they said it was required by a plugin, I tried it and it seems to work, the home page shows images working but i couldn't get that to work and no information about what image format could work. weirdly interesting choice, I guess you'd have to write a new dvi back end for lmtx though...? see marketplace.visualstudio.com/…

@DavidCarlisle I have a feeling that going dvi is not an option. :)

@mickep hint then?

@DavidCarlisle Hint?

[1/4, ??:??/??:??] update: knuth-hint [20242k] (67373 -> 70667) ... done

@mickep hitex output format

@DavidCarlisle Aha, that. Maybe html/mathml is a better option (even if not equivalent).

@DavidCarlisle What timeout do you have on texlive.net? I get an error "Timeout/Error status: 137 [context][document]" on a small thingy that takes 0.65s to run locally.

Seems there is some font cache generated or so (but I do not change font).

@mickep ah I run a few lualatex documents on the commandline for a new texlive to seed the font cache, I guess context is left out... is there a commandline call I can do to cache stuff (or just try a few documents) I think it was 1s and memory is limited as well

@DavidCarlisle Maybe I have to sign up for the paid version :)

@DavidCarlisle You can try mtxrun --script fonts --reload --force or something like that.

@mickep done

fonts           | names | registered fonts : 4662
fonts           | names | read files       : 11370
fonts           | names | skipped files    : 26
fonts           | names | duplicate files  : 26
fonts           | names | total scan time  : 16.009 seconds

@DavidCarlisle Works, indeed. :)

@mickep actually I remember now I did push up the limit, you get 15s on the web interface but you probably just exceeded that on a first run

@DavidCarlisle Maybe, but the error showed earlier than 15s.

@mickep It's a bit arbitrary and a real time timeout and the machine may be busy doing something else (like me running some online tagged pdf generation trials or whatever) basically you get what you pay for, but I would guess here it was mostly about the cache building.

I just learned that \GetFileInfo{foo.sty} doesn't require \usepackage{foo} in the preamble.

Except now it does. Sheesh.

Every time I think I've figured something out I embarrass myself.

Hi, can anyone try the line luaotfload-tool --find="times"? in win11 TL24... return C:\texlive\2024\bin\windows\runscript.tlu:941: ...4/texmf-dist/tex/luatex/luaotfload/luaotfload-arabic.lua:19: attempt to index a nil value (global 'node') :(

$ luaotfload-tool --find="times"
...4/texmf-dist/tex/luatex/luaotfload/luaotfload-arabic.lua:19: attempt to index a nil value (global 'node')

$ luaotfload-tool --find="times"
...4/texmf-dist/tex/luatex/luaotfload/luaotfload-arabic.lua:19: attempt to index a nil value (global 'node')

@PabloGonzálezL ^^ cygwin and ubuntu/wsl

@DavidCarlisle The usual culprit? : -)

@PabloGonzálezL yes I blame @PabloGonzálezL

But you could ask @MarcelKrüger as he is here.

@DavidCarlisle Jejje...I hope to @MarcelKrüger see message at some point :D

@PabloGonzálezL After careful consideration I came to the conclusion that this shouldn't happen.

@PabloGonzálezL In luaotfload-tool in line 1466 there is some config containing live = false. You can change that to live = true to resolve the issue.

Ok, I found the actual issue now. The detection that luaotfload was running in the tool was broken in some places. It's somewhat interesting that this didn't cause issues earlier. Anyway it will be fixed in the next release.

@MarcelKrüger thanks

@mickep You can ask emacs to get AucTEX for you, rather than asking Debian. I can't exactly remember the syntax, but I'm pretty sure you can get it from within emacs using M-x package-install. I know, because I had the same issue (and didn't want to install the Debian-packaged tex-live).

@UlrikeFischer Package tagpdf Warning: Parent-Child 'P/pdf2' --> 'verse/user'. can't be too bad can it?

@DavidCarlisle hm, sounds wrong. Wasn't verse mapped to Span?

@UlrikeFischer I only get it 28650 times so hardly any warnings at all...

@UlrikeFischer I'm using essentially the same macros but with an ASV bible not KJV so I don't get locked in the tower. I think I missed something.

@DavidCarlisle did you push already?

No I think I might so I don't lose it, give me a minute...

actually in KJV verse is mapped to P: add-new-tag=verse/P. So the question is why you have another P around that.

@UlrikeFischer pushed ASV in sibling directory \verse and \bookchapter now have an (unused) argument with the verse number (as the numbers were explicit in this version and I thought keeping them might be a good check)

you might want to add \endinput after Genesis to start with:-)

@DavidCarlisle this looks odd: ` #1 verse number not used`. Doesn't that print something before the structure?

@UlrikeFischer sigh teach me how to do comments in tex? I knew I should never try to document code...

@DavidCarlisle do you have the bible.cls in your path? it is not in the folder.

@UlrikeFischer it was in mine:-) try now

@DavidCarlisle hm, still warnings.

@UlrikeFischer yes I thought I saw it without but maybe I was back in kjv I see them now

@UlrikeFischer I'll cut it down to a couple of books tomorrow and have a look, If you feel tempted feel free to push but no worries of not doesn't have to be fixed tonight:-)

It must be something simple the structure is pretty identical to the kjv

@DavidCarlisle one obvious difference is that \bookchapter in KJV start the paragraph, here there are in a paragraph of their own. So one need to check if the logic is still right (probably not ...).

@UlrikeFischer well I can change the source to match, I just started with a plain text file and added the tex markup, let me check...

@UlrikeFischer oh no blank line after bookchapter...

@UlrikeFischer fixed

@DavidCarlisle behaves better now ...

@UlrikeFischer before I was using something installed from one of your branches, I'm using a standard tl2024 lualatex-dev now, should that be enough?

@DavidCarlisle I think so, there is not much open currently. my pr contains mostly some math fixes.

@DavidCarlisle one overfull box that just past by said "David smote ..."

@UlrikeFischer shocking

oh parent/child warnings pop up again around page 500 and something

\bookchapter{2}
(return to top)

@UlrikeFischer ^

never trust plain text files :-)

@DavidCarlisle ;-).

@UlrikeFischer why on earth did wikisource "download plain text" option just put (return to the top) on one chapter out of all the chapters in the whole work

pushed

@DavidCarlisle what have the romans ever done for us?

@DavidCarlisle I still see more than one (return to top)

several

@UlrikeFisch pushed again

@DavidCarlisle first round passed (and I should add a message at the end "busy creating structure objects, be patient" ;-)

@UlrikeFischer only three overfull boxes, two of which suspiciously have the same text

@UlrikeFischer I didn't copy in the setcounter as I thought they were odd anyway but \chapter isn't resetting the \bookcounter count so we are getting very big numbers

@UlrikeFischer can I just add \counterwithin{bookchapter}{chapter}? (I'll try...)

@DavidCarlisle hm. Where was the counter reset in KJV? Ah I see in the files after \chapter.

@DavidCarlisle should imho work fine.

@UlrikeFischer just waiting for you to finish making your tree and write the pdf so I can look at the counters...

@UlrikeFischer you can stop but TeX can't hyphenate Meshelemiah or hypocrisies which are the only two overfull boxes left

@DavidCarlisle hm, hy-po-cri-sies? No idea about Meshelemiah, Me-she-le-mi-ah perhaps?