@DavidCarlisle yes, at least in our cases. People could embedded files with other methods and we wouldn't catch that. The main question is if the message looks ok.

Package pdf Warning: PDF standard validation failure.
(pdf)                Embedded non-PDF files detected.
(pdf)                Switching standard from PDF/A-4 to PDF/A-4F

@UlrikeFischer ^^ that looks good but I'm confused about something else but it's been a long day, I'll look again in the morning

@UlrikeFischer despite what I wrote in the comments at davidcarlisle.github.io/tagging-project/documentation/…

also if you are still there can you remind me how to get a structure tree dmp using the bfo jar file?

'twas the week before Christmas and the perfect time to tell staff you're halving their teaching hours ....

@EmanueleNardi I guessed your nationality from your name :D Another Italian in chat, yay! The no-pineapple-pizza party is growing!

@mbert You're right, it should rather be \fp_eval:n { firstoftwo ( 1 , 2 ) + firstoftwo ( 3 , 4 ) } ...

@mbert No, I wasn't able to do that either. So, for such cases, some kind of built-in function would still be nice

@DavidCarlisle yes the AF-keys avoid that files are attached to the structure but the mathml is still read in if it exists. We perhaps need a meta key math/onlystructelem. Any suggestion for a name?

We could have a onlySE key but I don't really see why if math/mathml/AF is false that you input or attach the mathml file ? @UlrikeFischer

@DavidCarlisle the AF is a local key, you can enable/disable it for every formula, but the mathml file is read in globally at the begin and the fragment are embedded to the PDF directly. We would have to keep everything in memory if we wait until use.

@DavidCarlisle but perhaps we should try that, it shouldn't put excessive stress an memory and would allow to adapt the fragment on the fly.

A nice video about the bookbinding process: youtu.be/H4UZu7uCscs?si=3RyScl34snYK45ur, really nice channel

@UlrikeFischer ah yes I see, but maybe even if we don't go the hold in memory root a single /math/allmathml/AF=false or something to disable would be good I think. using sources= to prevent the sources feels natural enough in the original version without luamml, because if you supply no mathml you get no mathml, but with luamml generating mathml on the fly just a single key that says don't use mathml AF would be good I think. But I'll add sources= to the examples in the md file for now:-)

@DavidCarlisle perhaps a multichoice key math/activate={structelem,mathml-AF,tex-AF}?

@UlrikeFischer would be good, or mathml-SE rather than structelem for symmetry?

@egreg when to add the pineapple to the sauce?

@Skillmon at the very end, after eating the spaghetti

@DavidCarlisle yes, and perhaps math/setup instead of activate (which sounds as if you can deactivate it again).

@UlrikeFischer sounds good

@Rmano I got this recipe from @egreg the bbc says it gives that true Italian flavour bbc food

@DavidCarlisle I prefer maintaining the Italian flavour using Cattenan's bananas instead of pineapple. A bit more expensive, but you now... cooking is art.

@DavidCarlisle I'm now testing the pdfmanagement in latex and if everything goes well will sent it to ctan. Then we can merge the latestkey-latex branch and then I can add the new keys to my luamml-math branch.

@Skillmon The pineapple should be boiled in water. Then you throw away water and boiled pineapple and start with guanciale/bacon, eggs and spaghetti.

@Skillmon Replace pineapples with pine nuts and apples, make pesto from the pine nuts and eat the apples as dessert to the pasta

@samcarter That's been fixed now. Initially in a plugin, now merged into the core product. github.com/PgBiel/typst-tablex

@TRiG Ah, thanks for the ping (and whatever chat-archaeology you had to do to find the message :) )!

@samcarter Archive spelunking is a hobby of mine anyway, but in this case I was specifically searching for Typst.

@TRiG :) Good thing it has a distinguishable then!

(For certain applications, Typst is definitely better than LaTeX. I'm thinking of building a webshop CMS which needs to generate a PDF invoice from variables passed to a template. You can do that kind of thing in TeX, but why would you? I've previously used TCPDF for it, but I think I could make something look nicer with Typst.)

(And of course TeX and friends aren't good at that kind of thing, because they were never designed to be. This is no bash on them.)

@samcarter Definitely easy to search for, yes.

@TRiG There is a really nice latex package: ctan.org/pkg/zugferd

(all the addons like creating the xml etc. can be switched off if not needed, but the calculations etc might still be useful)

@samcarter Interesting. There's a lot going on there.

@TRiG I've seen a talk about the package at a tex meeting, it was very impressive.

They perhaps should have put a bit more care into their manual, though.

@TRiG It is open source, you can suggest improvements on github.com/TeXhackse/LaTeX-ZUGFeRD/issues

@samcarter I'm going to go for a walk before I run out of daylight, but when I come back I may just do that. It might be good for me to dust off my LaTeX again. (I was never great at it, but I did know some things.)

@TRiG Here, it is already dark, so I'm envious :P Have a nice walk!

(I've only done one complicated document in Typst, and of course I ended up needing to use a feature that it's admittedly not very good at yet: text wrapping. The result worked, but I suspect LaTeX could have done it nicer.)

Bye.

@TRiG in germany that wouldn't work. Such invoices need to be accessible and use zugferd standard soon.

@TRiG I use a LaTeX template for my invoices for years. I use a database that I can access via a local server and PHP and this way I replace placeholders in the TeX file and run LaTeX. Works like a charm

<3 Magdeburg

@yo' oh I just saw the news

Yeah :'‑(