@FaheemMitha and that's the problem: it's not a generic solution

@FaheemMitha strong in sense of being able to handle millions of requests a day

Notice that the only really publically available universal DNS servers are the two ones offered by Google

@tohecz Can't OpenID do that currently?

Right, I've registered a StackExchange OpenID!

@tohecz Really?

@FaheemMitha OpenID is basically a protocol, not an institution

@JosephWright Me too.

@tohecz True.

@FaheemMitha Well, the only ones I have ever heard of are 8.8.8.8 and 8.8.4.4

otherwise you're usually provided with whatever your ISP gives you

@tohecz there are other free dns servers out there.

@FaheemMitha ok, I was not aware of that

granted, if i had to think of one off the top of my head, i probably could only come up with 8.8.8.8

@tohecz Well, actually I don't think dns servers require auth, so you can use anyone you like.

though i could be wrong about that

i've tried using random dns servers before, and they work

@FaheemMitha usually restricted to the IP space of whoever runs them

At least as I use my own site for OpenID I can shift the forwarding

I mean, at home, I have a public IP, the DNS server has a public IP, but you can't access it from the outside

@tohecz Oh? I see. Maybe I was wrong.

"free public dns server" gets some hits

@FaheemMitha ok, good to know :)

Yay, got internet in the auditorium!

@PauloCereda lol. Conference participation \to 0 :)

@tohecz \qed <3

@PauloCereda \qedhere ;)

(since chat is a sort-of list environment...)

@tohecz \randomqed :)

Does anyone know how long processing a new TUG membership application typically takes? Obviously not a critical issue, but it's been 2 days and I thought I would have heard something by now.

@PaulGessler no worry, if you said 2 months, I would be worried :)

@tohecz thanks, like I said, not an issue but I just wanted to know the relative speed of the organization :)

Guys, am I crazy:

\begin{filecontents*}{\jobname.dat}
2.76658    -0.07990     0.10551    -2.94131     0.00840     0.23832    -2.56759     0.04593
-0.21476     2.37350    -2.30670     0.11634    -2.36124     2.62034    -0.32261     2.36860
 0.01118    -2.42926     2.42470     0.08698     2.45065    -2.39544    -0.16931    -2.32933
-3.04568    -0.15376     0.03051     2.74830    -0.07237     0.02359     2.96758     0.05812
-0.12791    -2.50370     2.63524     0.15000     2.26310    -2.39198    -0.05032    -2.41050

No way I'm getting Jake's output. :(

@tohecz I posted a question on Genealogy.sx. Though I don't know whether it is on-topic or not.

Does someone have the canonical »You need to run bibtex« question at hand?

i looked at the allowed topics, but only after I posted the question.

@Johannes_B it's in the "Often referenced questions" at meta

Thanks

@tohecz I am pretty sure i bookmarked that once

@tohecz Thank you very much

Once again saved to Favourites

@Johannes_B you're welcome :)

Those important meta threads come up immediately if you go to meta and sort questions by votes :)

gotta go, see you probably tomorrow!

@tohecz Gotta go to to, bye. Have a nice evening

@PauloCereda Yep I also don't get anything close either.

@egreg Proof-of-principle for doing expandable UTF-8 case changing in Lua without case changing macros:

\documentclass{article}
\usepackage{expl3}
\long\def\luachangecase#1#2{%
  \directlua{
    local unicode  = require("unicode")
    local utf8     = unicode.utf8
    unicode.string = string
    tex.sprint(unicode["utf8"].#1("\luaescapestring{\unexpanded{#2}}"))
  }%
}
\def\foo{a}
\ExplSyntaxOn
\cs_new:Npn \tl_to_upper_case:n #1
  {
    \etex_unexpanded:D \exp_after:wN
      {
        \tex_romannumeral:D
          \__tl_to_case:nn { upper } {#1}
      }
  }
\cs_new:Npn \tl_to_lower_case:n #1
  {
    \etex_unexpanded:D \exp_after:wN

Could be tided up, I'm sure, but the mapping has to be done in TeX I guess

Issue of catcodes remains: you get those in force at point of use, not point of original tokenization of input

@JosephWright I don't think you need to use require("unicode"), this library is preloaded in luatex

@michal.h21 At present, at least :-)

@michal.h21 I'm not saying it's perfect: I've got little Lua experience, so there will be mistakes

@JosephWright yeah, I hope they just add new unicode library and don't remove slunicode, I use it quite often

For example, I know there is some business about [[ ... ]] being possibly better than \luaescapestring, but I forget the detail. Plus of course for a format you'd want to use bytecode, ...

@michal.h21 Seems clear it will be removed from the docs, but I'm not sure whether that means unicode.... will not work

@michal.h21 Part of the pain is that Lua itself is 8-bit, so it makes life more complex than perhaps would be ideal

@JosephWright selene lives in its own namespace, so I hope it will stay.

@topskip As I said, docs seem clear on that but you never know

@JosephWright I hope they will not remove it. in my other project, I found that it was broken for half a year because of new lua version in luatex and I didn't notice it

@michal.h21 Yes, this sort of thing worries me with my 'LaTeX3 hat' on!

(I am now glued to 0.71.0, so I'll probably have selene for a long time)

I'm more concerned at the moment with the business in the docs about \lowercase and \upeprcase

OK, we have some Lua experts so perhaps you can tell me an answer to this one

How do you deal with the fact that Lua doesn't respect TeX catcodes in material it returns?

@JosephWright (too high for me, I mostly use tex.sprint(-2, ....) to make sure that tex never sees any "bad" catcode)

@topskip Doesn't really help, though. My earlier example was case-changing abc $ y = mx + c$. If I've already tokenized the input, but then alter the catcode of $, doing a Lua-baed case change screws up the math shifts

I send Lua tokens, but it doesn't send them back :-(

@PeterGrill Thanks Peter, I was afraid it was only me.

@Jake: Hi Jake, sorry to bother you. Is there any secret to run your code from tex.stackexchange.com/questions/155291/… ? Here's my code: chat.stackexchange.com/transcript/message/13631010#13631010

@JosephWright you use tex.sprint(), right?

@PeterGrill Possibly. It can also be some internal change in some revision.

@topskip tex.print: I don't quite get the difference

Hi @topskip! Miss ya, buddy!

@PauloCereda me too!

@JosephWright tex.sprint makes TeX run. :)

@topskip Perhaps my example had sprint, but the problem remains

\documentclass{article}
\usepackage{luacode}

\begin{document}

\begin{luacode*}
tex.sprint("$4+3=2$")
\end{luacode*}
\end{document}

works fine

I guess I'm annoyed that tokens passed to Lua from TeX aren't stored as an array of <char>-<catcode> (as in TeX), as that's how I think

So I don't know where the problem is with catcodes.

@topskip Yes, but what about:

\def\tokens{$y = mc + c$}
\catcode`\$=12 %
\directlua{%
  tex.sprint{"\tokens"}%
}
\bye

@JosephWright right, they are lost. Lua only knows about characters

Catcode of tokens has been messed up

@topskip Yes, but in linking to TeX I'd have stored all TeX-input with them, just as TeX does

@JosephWright That's what I'd expect

@topskip Why? We've tokenized the input, so Lus should respect that

LuaTeX sees the text "$y = mc + c$" at the end of the last curly brace

@topskip Nope, it sees a load of tokens then proceeds to throw away the catcodes

@JosephWright Because Lua doesn#t know anything about tokens

you are probably right, and I should keep my mouth shut

@topskip But that's my point: in using Lua in TeX there should be a way to retain this info

... before I think

My problem is that it makes it hard to use Lua to do things with TeX input when it's not just 'text'

See my need about to do a TeX loop to find control sequences

Anyway, I'm sure this is all workable :-)

I guess it comes down to using the tools in the right way

@JosephWright Here's the response of \tl_show_analysis:N \test

The token list \test contains the tokens:
>  A (the letter A)
>  B (the letter B)
>  C (the letter C)
>  D (the letter D)
>  E (the letter E)
>  \^ (control sequence=macro:->\OT1-cmd \^\OT1\^ )
>  F (the letter F)
>  Ą (the letter Ą)
>  É (the letter É)
>  Д (the letter Д)
>  ῼ (the letter ῼ)
>  Λ (the letter Λ)
>  \foo (control sequence=macro:->a).

@egreg Sounds about right :-)

@JosephWright Yes; now for something more difficult.

@JosephWright I added {}~#^? and got

>  { (begin-group character {)
>  } (end-group character })
>    (blank space  )
>  # (macro parameter character #)
>  ^ (superscript character ^)
>  ? (the character ?).

@egreg All fine

@JosephWright Seems so.

@egreg I think it all works with the proviso that catcodes are 'current' ones in all cases

If you are dealing with 'text' after expansion then that's not really an issue, I guess

@JosephWright Yes, _ gives letter, of course. Now we would like to avoid uppercasing math. ;-)

@egreg Could be fixed using the optional arg for tex.sprint here, or of course by using at a document level

@egreg The math mode problem is different, of course, and applies to any approach to case-changing

@egreg I guess you have to do this in TeX even with LuaTeX (so you can check catcode of $ or similar)

@JosephWright Probably; some trick like in textcase.

@egreg Yes

@JosephWright what about token_filter callback?

small sample:

it doesn't work correctly, because luatexbase.remove_from_callback("token_filter", "hello") doesn't remove the callback and all tokens until end of the document are processed, but I hope it can be fixed

@JosephWright Just to get sure: Did you receive notices for chat.stackexchange.com/transcript/41?m=13614908 (if not: see also two following postings of egreg and me) and chat.stackexchange.com/transcript/41?m=13616906 ?

@egreg Deep inside I believe to remember, that you once answered a very similar topic, but I do not know how to search for: MikTex doesn't accept “ù” character. Seems, this is from an Italian. Otherwise, since there was no reaction anymore, we can probably closed this as unclear.

@Speravir Yes: I guess I need to take a look!

@JosephWright Even if you decided to do nothing I just wanted to know …

@PauloCereda I used the second data format, which the OP had provided in a Dropbox link. That has been deleted in the meantime. I've edited the question to include a sample of the file format.

Anyone know of any cool uses of labels? Beyond just the standard cross-referencing stuff? I only recently started using labels even for that purpose, but I can't help but feel that there must be some cool uses of labels that escape me. Don't really have anything in particular in mind, hence the chat question rather than a question on the main site.

@NicolaTalbot After a bit thinking: In MiKTeX the use of binary executables has clearly security reasons. With the binaries a check of config file scripts.ini is involved, and only when there is an entry (with the same name) the program is progressed. This config file is signed, so alteration would not work. Starting a binary where no entry does exist fails, too. The second part, that obviously the path is kept in config file, seems to be the culprit for several issues in past.

@Speravir Well, saying that the document is UTF-8 encoded and that \usepackage[latin1]{inputenc} is called is a clear indication of confused ideas. ;-)

@Speravir voted

@egreg Yes! :-) But did you see the edit?

@egreg Ah, oh.

@Dennis Well, you can label arbitrary chunks of text and paste them into other places.

@Speravir Yes: the document is probably encoded in some Windoze code page rubbish.

@FaheemMitha Interesting use of labels there. Do you do that often? (trying to see if it's a useful use or merely a possible one...)

@egreg I think so. Do you know by any chance, whether win-1252 (= cp-1252) used in Italy, too? Then ansinewwould have been the right choice.

@Speravir I know nothing of what Windoze users do.

@egreg :-) :-D

@Speravir I tell students: whatever system you have, use UTF-8, as there's no other encoding.

@egreg Cf. discussion of engines

@egreg Yepp, and this works in recent Windows versions as well. BTW first result in guggle search: scratchpad.wikia.com/wiki/… Yes, it's how I thought.

@Jake Oh thank you very much! Sorry for the trouble.

@PauloCereda How was the conference today?

@egreg Quite interesting, but extremely tiring. :) I'm in the steering committe, so lots of things to do in between keynotes. :)

How are things in here? What did I miss?

@PauloCereda Uppercasing frenzy by @JosephWright

@egreg Oh no, I think I need to be here in order to act as a referee. :)

@FaheemMitha Stack Exchange OpenID. Oh, I realize you already know …

@Dennis Not really. Did you find the relevant question?

@Speravir Thanks.

@Speravir Given the security issues with executables, I wonder why MiKTeX uses an .exe wrapper instead of a .bat wrapper. Karl said the best portability option is to use Lua, so I probably ought to get around to learning it!

@NicolaTalbot Well, as already said I’m just guessing: It must be avoided, that something is shipped out from a third party in the MiKTeX bin folder. I think he wanted to have it hardcoded. A batch could be altered. (But the scripts, too …)

@Dennis In case you care, this is the question

Does someone know, what happens, when I want to close a migrated question as off-topic? See MikTex Download Failure - toptesi.tar.lzma. Will the migration be rejected? ping @JosephWright

I thought $\alpha = 3.81$ m/s was the most correct....

@Canageek Knuth himself thinks that numerical values count as part of the maths.

@Canageek $\alpha=\SI{3.81}{\meter\per\second}$

@Canageek Of course this requires \usepackage{siunitx}. None of your attempts is correct. If you don't want to use siunitx (but I can't see why), $\alpha=3.81\,\mathrm{m/s}$

@egreg I've been trying to talk him into using siunitx ;)

@Canageek $\alpha=3.81\,$m/s or $\alpha=3.81$\,m/s, that's the quick way to go, given that you are sure the text will be typeset in \normalfont, and not in \emph, \textbf or such.

He is using @egreg's suggestion about mathrm, but can't tell the difference between it and textrm

Now he decided to try siunitx

@Canageek Try \textrm in an italic context.

@egreg I am unsure and cannot remember, what Joseph said. What do you think, unclear or off-topic,? Cf. comments: Texworks compiling acting non-deterministic. Difficulty adding images to document.

@Speravir I'd say "unclear": judging from the comments, the example compiles.

@egreg OK, then I do it in this way. One could also say a not reproducible problem is off-topic.