@DavidCarlisle No, luaotfload isn't supposed to work in plain dviluatex (see github.com/latex3/luaotfload/issues/240). there we don't have the callbacks needed to make this work properly. You can still use it if you adjust your output routine to call the callback though.

@MarcelKrüger ah, thatnks, i did know that once... But main point of my comment seemed correct, make an issue at gh and let you fix it:-)

@DavidCarlisle Did you see the post about \ProcessOptions* on the TL list?

@JosephWright yes, but it was late, I didn't read in detail, I meant to ask them to post an issue at GH

Friday quack

@PauloCereda breakfast

@DavidCarlisle oh no

armv8 support for IoT TL images!

@PauloCereda cygwin over windows on arm?

@DavidCarlisle oh no

@DavidCarlisle go WSL

@PauloCereda it's more fun to torture you with cygwin support requests

@JosephWright sent holding reply

@DavidCarlisle ooh

I always struggle with \str_if_eq:nnTF { abc } { \seq_item:Nn \l_tmpa_seq { 1 } } ... how to solve this? Do I need e somewhere?

@JasperHabicht Yes, as the n type args here are literally what you've typed

@JasperHabicht Yep, \str_if_eq:neTF or \str_if_eq_e:nnTF

@Skillmon ah, so this is the problem actually! then I never read the spec right =D thanks!

And, o would not be enough in this case, I assume? So, it needs to be e and I should define a variant for this?

\str_if_eq:eeTFis predefined, so I use this, maybe ...

@JasperHabicht o only works if you have exactly one expansion, so typically low-level stuff: no good here

@JosephWright thanks!

@JasperHabicht yes, use that, abc isn't expandable, so if you have control over the first argument and know it won't expand you can use \str_if_eq:eeTF (it'll be faster on pdfTeX and XeTeX, not sure about Lua).

(I think \str_if_eq_e:nnTF doesn't exist, my brain tricked me because \str_case_e:nn exists)

@JosephWright regarding the \DeclareOptions-message on the list: I'd argue that this was a misbehaviour, and shared keys shouldn't be \RequirePackaged, but \input.

@Skillmon meanwhile chatting to <todays random user name> in comments never seems to help, it just encourages more pointless questions and more user names.

@DavidCarlisle I noticed that after a month or so the boolean now follows the expl3 naming scheme ;-).

@DavidCarlisle sorry :(

@UlrikeFischer ... so in approx. 5 years the questions about their list might finally stop?

@samcarter only if we are fast enough with comments before the question disappears again ...

@samcarter hopefully before then he'll stop spamming here and move to other places like topanswers.xyz ...

@DavidCarlisle I will make sure to keep you informed (in great detail!) about each of their questions there :)

@DavidCarlisle :)

@DavidCarlisle we could name him Randolph

I wonder whether Randolph read about our heuristics, as of lately there are no question marks in the questions anymore.

@Skillmon They now also start new accounts with a question so sparse, that it is hard to identify them.

I'm off the palindrome track, so this is the best I can do. :)

@AlanMunn -- But our questions are very close: 255,550 questions . So watch. We're going out to dinner, so I can't. (Happy day after Thanksgiving.)

@MarcelKrüger at the context list Hans complained about the way tikz "registers function where it assumes that it's the only package in the tex universe". I think he spoke about the following definition. What is wrong with it?

  \directlua{
  local lft = lua.get_functions_table()
  lft[\string#lft+1] = function()
      local lhs = token.scan_string()
      local rhs = token.scan_string()
      if lhs < rhs then
          tex.sprint(-2, "-1")
      elseif lhs == rhs then
          tex.sprint(-2, "0")
      else
          tex.sprint(-2, "1")
      end
  end
  token.set_lua("pgfsys@strcmp", \string#lft, "global")
  }

@UlrikeFischer It's a good line :)

@JosephWright ;-) yes, but I wonder if they can/should do it better under the assumption that they want to stay compatible to all formats.

@UlrikeFischer It kind of assumes that it can use some I'd just because it's free, leading to issues if someone tries to assign functions based on a e.g. counter (as LaTeX's newluafunction before that got adjusted to work with such usage). Also if he was talking about lmtx he might be referring to the use of lua.get_functions_table which in LuaMetaTeX at some point deleted all defined functions. (Not sure if that's still an issue, I always just patched it to work more sensibly)

@MarcelKrüger but I guess tikz can't really do it better unless it tries to detect the format?

@UlrikeFischer I guess that Hans actually wants exactly that: tikz should try to detect the format and then use a way native to the format.

@MarcelKrüger well if he wanted that he should open a tikz issue and not simply define \pdfstrcmp to work about the tikz code ;-).

Answer accepted after 12 years and 21 days!

@egreg they waited to get a palindrome 12.21 ;-)

@DavidCarlisle How should I read \sp? SubscriPt?

@egreg Wow! What question?

@mickep ah well one or the other

@DavidCarlisle Oh, so \SuPerscript... Well well.

@mickep as I wrote in my comment, like ^

@DavidCarlisle Yes, now it is clear.

@mickep :-)