In expl3, if I use V expansion on a boolean, the result is 0 or 1 which is fine, but I can then not access this value using \bool_if:nTF it seems

So, I should probably not use V expansion, but something like bool_if_q:n?

The answer seems to be yes. At least this way it works

I wonder whether V expansion of booleans shouldn't expand to \c_true_bool or \c_false_bool?

This discussion on character sets is something which may interest some people here (at least, it interested me enough to want to share...) retrocomputing.stackexchange.com/questions/31315/…

@JasperHabicht Yes

Ah, manual says: v type expansion is not supported for bool. I guess there was a long discussion about this

@JasperHabicht There is no \bool_use:N so there is no V-type support

@JosephWright Yes. Makes a lot of sense

@JosephWright @DavidCarlisle blames you for things he doesn't expect when you are not here ;).

@cfr :)

@cfr I'll take a look in a bit - just rebuilding .tlg files ATM

@JosephWright or you could blame him for what he expects?

it is not what I expected either, but that is par for the course so I am probably not worth the trouble of blaming.

@cfr :)

@DavidCarlisle FMi

@cfr, @DavidCarlisle I'll see if Frank remembers why he set it up this way

@JosephWright it looks like it is designed to handle dos text file tests on linux, but I'm not sure why

@DavidCarlisle We could change in l3build without impacting on the 2e repo: the latter uses it's own test2e.tex definition

@DavidCarlisle I wonder if it was to do with @UlrikeFischer's need to use Windows files - but as they are all generated files, I don't see where the issue comes from

@JosephWright you mean about \SHOWFILE? I don't remember any discussion about that. When was the ^^M implemented?

@JosephWright yes my thought was to blame @UlrikeFischer's tests but like you I couldn't see a workflow where this would show up

@cfr as you see from the above, it may be that we can blame Frank or Joseph or Ulrike, but the important thing to take away from your question is: don't blame me.

Why are you \SCREAMING so much these days? ;)

@DavidCarlisle Quite

@UlrikeFischer Frank Mittelbach 6 years ago (September 28th, 2018 10:45 AM)

@UlrikeFischer SHA 3e11288bb8c070ce93893229f92992dbcf91596a

@UlrikeFischer github.com/latex3/l3build/commit/…

@JosephWright can't be me, I joined in october 2018 ;-)

@UlrikeFischer :)

@cfr Could you log an issue about \SHOWFILE with an example of the issue - best to discuss there rather than here or on the team list (not visible to others)

@cfr ?

@cfr (@DavidCarlisle, @UlrikeFischer) I will though also prod Frank directly

@cfr but why are the ^^M a problem for you? I saw them too in some texts (e.g. when I show the created mathml in some tests) but I simply ignore them. From a point of the test the only thing that matters are changes ,

@mickep because unit tests don't like to be ignored.

@Skillmon OHH!

@mickep all the commands defined for unit tests are in all uppercase.

@UlrikeFischer scary and monstrous!

(from the tooltip point of view)

@UlrikeFischer Ȧ̇̇̇̇̇

@UlrikeFischer A̸̡̭̜͈̠̽̓̐̚A̵̢̟̲̾̈́̏̂͑̚A̷̻͆̈́̎̀̓̈́͗̐͛̌̍͂̕̕͝Ả̵̡͇̫̜̗̭̜͇̖̱̗̝͓̘͋̈́̓̏Ǎ̵̡̡̢̼̖̝͕͎̗̟͎̙̽͋͊́̀͗͘͠A̷̡̳͔̯͈͖͙̦͕̪̎͋́̎̂̌̉̈́͐̅͝ͅͅȂ̷̻̾̓͂̿A̵̧̦̘̹̩͇̼̱̞̻͆̿͒͠Á̸̧̺̘̯̙͈̠͙͜ͅȀ̷̛̰̭͙̯͔̗͕͍̟̖͍̑́A̸̤̾͆̑Ạ̴͍͈̱͚̜͉̞̒͛͛̒͋̑͝͝͝

scream cipher with unicode support?

@UlrikeFischer AAH, very well!

@Skillmon Oh, does it compile?

@mickep With LuaLaTeX and Noto Sans ... yes

@JasperHabicht Beautiful!

It is actually Chinese using the above cipher ...

Getting back to xbeamer: hopefully I can sort out 'decoding' the overlay spaces this week

Really want something I can use for my own lecture slides by the summer

@JosephWright with the risk of me not being productive (there or anywhere): keep me posted and ask for help if you need me.

@Skillmon :)

@Skillmon I want something that vaguely works before I check it in: I get the feeling there will be several rounds of refinement (\beamer@masterdecode does various things, and I'm getting the sense they might need to be separated out, at least outside of the 'innards' of reading overlays)

@JosephWright the problem (and advantage) most likely is that I don't know the innards of beamer and only the macroscopic behaviour, so I'd implement something most likely completely different on the architecture side that only behaves similar from the users' perspective

@Skillmon Don't worry - the plan as it stands it more about the user behaviour than the inners, and even then some of the user behaviour (some aspects I think don’t make sense to carry forward for a new class)

For overlays, I will try to get something similar to the current parser, but I am writing from scatch

@JosephWright good, and agreed :)

(remember that I'm part of the xbeamer "team", so I know that the plan is to implement user behaviour, not innards, I'm just one of the guys interested in the project who knows almost nothing of the beamer innards, unlike you and @samcarter)

So for me it comes quite naturally that I don't (re)implement some internal structures because I simply don't know beamer's

@Skillmon :)

@Skillmon best preconditions for a procrastination project

@UlrikeFischer er ... mostly because I didn't expect them. and I didn't realise l3build rewrites a tlg for lua, so not only I had differences I didn't expect at all, but I also have engine differences I didn't expect, including weird characters showing up for xetex and pdftex, but not luatex. and these are showing up in the logged content of .memos, which are identified by checksum, so could not figure out where things diverged. (actually, I still don't know why I now need different tlgs.)

@JosephWright but is it an issue or just user ignorance?

Someone broke CTAN?