@JosephWright Thank you, but as you might have seen, I'm doing very little for the site. I visit it mainly when I receive a mail notification like in this case. There are other enthusiastic users who deserve TUG membership much more than me. So, the answer is no.

Hi folks. What's the most painless way to get text like "Batič" to render correctly in regular pdflatex. Or do I need to switch to something that support UFT8 (or whatever)?

@karlkoeller Understood

@FaheemMitha If your file is utf8 encoded and you load package inputenc with the corresponding option, inputting the word just like that should work.

@FaheemMitha \usepackage[utf8]{inputenc} should do it

@JosephWright Thanks, I'll try that.

@JosephWright That worked, thanks.

@Johannes_B Thanks as well. I think you were saying the same thing, right?

@FaheemMitha Yes :-)

Does anybody see a good answer for the following question or should it be closed?

@Zarko Did you see my question re. the TUG thread?

@Johannes_B I voted to close as dupe.

@TorbjørnT. egreg just posted an answer. Thanks @egreg

@Johannes_B The correct answer would be “why on earth are you using a style that puts a space before semicolons?”

@egreg :-)

@Johannes_B Oh, sorry, that was in reply to the LyX question you commented on, not the one you mentioned here in chat.

@TorbjørnT. That explains why i didn't see the vote. I just saw the question in the review queue and voted as well.

@JosephWright, no I didn't see. A recent days due to cached flu I only occasionally beside my PC :-(. Mainly looking if some simple question are on TeX. SE (for more I'm still not fit). I check now, but there now I don't see any your question.

@Zarko I nominated you here: meta.tex.stackexchange.com/q/7212/101651. The question is: do you accept the nomination?

@Zarko Do you have a regular TUG membership already? If not, you can win a one-year-membership ;-)

@JosephWright, I'm honored with this, I accept, but I still not found the question. What I should do now?

@CarLaTeX I think you can undelete?

@Zarko Nothing, I've undeleted my post!

@JosephWright, no still not have regular TUG membership. I was elected last year here.

@ChristianHupfer Done!

@CarLaTeX, thank you very much for kindness. I will work further to improve my answers :-).

@CarLaTeX Voted ;-)

@Zarko You're welcome!

@ChristianHupfer We have all the 8 positions covered, now!

Now I go back to bed. Hopefully I will back on the evening. Best regards to all.

@Zarko All the best

@CarLaTeX Hm, yes. I find it a little bit strange that compared to 2015 there are so few self-nominations. It's difficult to get the list filled, actually

@Zarko Get well soon!

@Zarko Buona dormita!

@ChristianHupfer In the end it's not meant to be a pressure situation: the mod team would sort out some names if we didn't have 8 takers

@JosephWright That's another pressure then... ;-)

@ChristianHupfer It's very strange... perhaps it is not clear that it's a gift!

@CarLaTeX It's not a loan, it's a gift @JosephWright ;-)

@ChristianHupfer Yes, I know: more than that it's a way of widening people's experience

@JosephWright @CarLaTeX: I think it is just a matter of public outreach. We have lot of users, but not many know the Meta site and the 'advert' of Meta on the right panel on the main site. So it is limited in outreach and those who are active on Meta have a membership of their own already. In addition, you won't catch the careless user who comes in for one (stupid) question how to format his thesis and at best, write the thesis for him or her ;-)

@ChristianHupfer We should invent an ad for the next year!

@CarLaTeX The timing doesn't work out

@JosephWright Is there a fixed date until the names must be known and transmitted to the SE Staff?

@ChristianHupfer It's incredible that there are users only interested in their thesis! :):):) (@PauloCereda)

@ChristianHupfer No, but to make sense this needs to be done before the first TUGboat of the year

@CarLaTeX @PauloCereda is a counter - example :D

@ChristianHupfer That's why I pinged him!

@JosephWright — I sent off a new version of fontspec today; hopefully the last I need to do for a little while :) But I never got Travis CI working! Just to confirm — if you get an error in a test in Travis, you have no idea what that error was until you re-run the test suite locally? I'm wondering about an addition to l3build to output the diff if it fails in -H mode.

@WillRobertson Yes, that's right: Travis-CI just tells you 'something went wrong'

@JosephWright Quack! I am a duck and I am going to join your colony.

@WillRobertson For a lot of test systems I guess the name of the failing test would be enough ...

@JosephWright We have one year's time to refine timing and advertising!

@JosephWright True... perhaps the 'competition' should take place, say, two weeks earlier and with a longer time for announcement? (Given that the settlement between SE and TUG is basically fixed and does not have be negotiated each year again)

@JosephWright Which I guess is fine for all programmatic style failures…

@JasperLoy Ducks are always welcome here!

@CarLaTeX In some languages, duck means male prostitute, LOL.

@JasperLoy Hahahahahaha

@JosephWright I'll see if I can get something working quickly in l3build...

@ChristianHupfer Yes but I have to know that they are continuing to pay ...

@WillRobertson Easy enough to do: perhaps -d?

@PauloCereda see @JasperLoy message above! Hahaha

It is very strange that there is no review of More math into LaTeX on tug.org when that book has been out for years. The same goes for Guide to LaTeX.

@CarLaTeX He's busy so let's not disturb him. =)

@JasperLoy Book reviews are a relatively modern thing

@JosephWright There's a new job for you then. =)

@JasperLoy I forgot he's writing his thesis!

@JosephWright As in "display"? That works. I was thinking maybe -v (for verbose). Or otherwise just make -H do it by default?

@JasperLoy Boris (TUG President-elect) is in charge

@WillRobertson You have changed a lot since your last photo. I was thinking this is your father's picture. =)

@JosephWright That's what I meant -- the settlement (or agreement?)

@WillRobertson -v is (almost) always --version and so it already taken, so it you want that it would be -V

@WillRobertson -d == --diff was my thinking

@WillRobertson -V would I expect not do the 'hiding' of output of any of the runs

@JosephWright Oh of course. makes sense :) I think you're right -- -V would be useful for other things (with -VV -VVV, etc, for different levels of verbositry)

@CarLaTeX I refuse. :)

@PauloCereda Resistance is futile ... you must comply ;-)

@JasperLoy So I need to cut my hair and shave. I still have a mug.

(This is me right this minute of course)

@WillRobertson This looks like Hugh Jackman. Congrats!

@WillRobertson You look suspiciously like Tony Stark. :)

Ooh, so these two Hollywood actors actually look alike, since they both look like Will.

@JosephWright Perhaps I misunderstand, but what about e.g. travis-ci.org/wspr/fontspec/builds/200626212? "Check failed with difference files - ./build/test/01-pkg-load-um.xetex.diff" Isn't that more specific than "something"?

I am guessing that looking alike is an equivalence relation. =)

@PauloCereda I think I will buy Enderton's Computability Theory after all. I currently have his two other books.

@JasperLoy I got an ebook, it looks very good.

@JasperLoy Haven't you ever noticed that you've never seen me and Hugh Jackman OR Tony Stark in the same room?

@WillRobertson Oh my

@TorbjørnT. Yes, it just tells you the name of the test that failed. But there's no way to get the contents of that diff file.

@WillRobertson Ah, I see.

@WillRobertson And, by the way, have you ever seen @PauloCereda and Spiderman in the same room?

@egreg Nor have I seen @egreg and President Coriolanus Snow in the same room.

@WillRobertson At least during set-up a local VM is still useful: working out which TeX Live packages one needs to get everything working is non-trivial, particularly for fonts

@JosephWright @WillRobertson @UlrikeFischer tex.stackexchange.com/a/353431/1090 any ideas?

@DavidCarlisle What about checking if 204E exists, and if not falling back to ... erm ... something?

Damn.

It obviously doesn't work to fall back to another font.

Maybe we need to alter the definition of \@fnsymbol (or whatever typesets the mark) to use *

(without superscripting it)

@WillRobertson trouble is if the whole footnote mark is already defined to be superscripted, and that's just printing the counter inside that

@DavidCarlisle \egroup*\bgroup

@DavidCarlisle (that was a joke)

could define it to be ifcharwd"2217=0*\else \char2217 ? so at least it does something?

Check for 204E and if nothing, take the *, lower it, and scale it up!

@DavidCarlisle Isn't \iffontchar easier?

I know you're anti-eTeX and all, but

@WillRobertson my first thought was \setbox0.}\ifdim\wd0 takes a bit of catching up with these new fangled tests:-0

I agree that's the least we can do

@WillRobertson I thought I was doing well to remember etex \ifcharwd thing!

@DavidCarlisle, @WillRobertson Any thoughts on the timing of the next team meeting?

@DavidCarlisle, @WillRobertson Somewhat related: tex.stackexchange.com/questions/353393/…

@JosephWright Sorry for my ever-slow replies on that — I should be okay most of the time at the moment...

@WillRobertson Don't worry: you are not alone :)

@JosephWright lunch time or 9am work for me monday or tuesday this week or any day this month after this week, I think

@DavidCarlisle OK: I'll see what Frank says

\DeclareTextSymbol{\textasteriskcentered}{TU}{%
\iffontchar\font"2217 "2217 \else"002A\fi}

@WillRobertson @JosephWright ^^ I suppose? Looks a bit odd but....

I never realised there were such big differences in the shapes of the letters of mathptmx and Times New Roman

Anyway

I think it's not optimal but we could get away with it for now. I take it you're against any shifting/scaling of the glyph?

@WillRobertson I propose \raisebox{-.5ex}{*} if U+2217 doesn't exist.

@egreg I think 0.3ex is better :)

@WillRobertson Whatever, except the xunicode default

@egreg Yep :)

Or equivalent...

\documentclass{article}

\usepackage{fontspec}
\setmainfont{Times New Roman}
\newfontfamily{\termes}{TeX Gyre Termes}

\UndeclareTextCommand{\textasteriskcentered}{TU}
\DeclareRobustCommand{\textasteriskcentered}{%
  \iffontchar\font"2217 \char"2217 \else\loweredasterisk\fi
}
\newcommand\loweredasterisk{\raisebox{-.5ex}{*}}

\begin{document}

x\textsuperscript{\textasteriskcentered}{\termes\textsuperscript{\textasteriskcentered}}

\end{document}

This is with -.3ex

@WillRobertson ^^^^^^

@egreg Definitely font-dependent! I'm happy with 0.5ex as a nice round number.

ASTERISK [^^^^002a]

LOW ASTERISK[^^^^204e]

ASTERISK OPERATOR[^^^^2217]

FOUR TEARDROP-SPOKED ASTERISK[^^^^2722]

FOUR BALLOON-SPOKED ASTERISK[^^^^2723]

HEAVY FOUR BALLOON-SPOKED ASTERISK[^^^^2724]

FOUR CLUB-SPOKED ASTERISK[^^^^2725]

HEAVY ASTERISK[^^^^2731]


EIGHT SPOKED ASTERISK[^^^^2733]

SIXTEEN POINTED ASTERISK[^^^^273a]

TEARDROP-SPOKED ASTERISK[^^^^273b]


HEAVY TEARDROP-SPOKED ASTERISK[^^^^273d]

HEAVY TEARDROP-SPOKED PINWHEEL ASTERISK[^^^^2743]

BALLOON-SPOKED ASTERISK[^^^^2749]

EIGHT TEARDROP-SPOKED PROPELLER ASTERISK[^^^^274a]

@egreg @WillRobertson I just tried ^^ with times new roman, only the first works:(

@DavidCarlisle All the almost-the-same symbols

@DavidCarlisle TeX Gyre Termes only has U+2A and U+2217.

@JosephWright :)

@JosephWright — sorry I missed the (luc: thing — there have been so many commits recently it got lost in the mix, and when I grepped for "luc" nothing came up. You were too clever for me!

@WillRobertson No problem: I though about just covering that but was trying to find a broader pattern

@JosephWright Yes, I'm very impressed!

@WillRobertson Bruno's been busy :)

@WillRobertson @DavidCarlisle Of course none of this is correct: the asterisk in this case should not be superscripted. But that's a different ballgame altogether.

@egreg Right, we have to make do with what we've got

@egreg Did you get my Donald Sutherland joke? I thought it was quite flattering

@egreg yes but that means making the footnotemark command encoding-specific (and then using superscripted characters) rather than just the counter

@WillRobertson I had to google for it: not a fan of that kind of movies. But I was talking of real persons, not fictional. ;-)

@egreg I just picked a recent movie I knew he was in :)

@egreg Are you saying Spiderman is real??

@WillRobertson So you don't deem M.A.S.H. a recent movie?

@WillRobertson Of course!

@egreg Well...

\DeclareTextCommand{\textasteriskcentered}{TU}{%
\iffontchar\font"2217 \char"2217 \else\raisebox{-.3ex}{*}\fi}

@egreg I'm not saying there's anything wrong with older movies, but the likeness was probably less so then :)

@WillRobertson @egreg this in tuenc.def ^^

@DavidCarlisle Not 0.3, please, see above

@DavidCarlisle Why \raisebox? Isn't that graphicx?

@egreg or should it be using \charht in some way...

@WillRobertson no kernel:-)

@DavidCarlisle The top picture is with -.5ex

@DavidCarlisle There you go. The things you learn!

@WillRobertson Doesn't need any driver code :)

I suggestion \raisebox{\fp_eval:n{pi/7}}{*}

@JosephWright Makes total sense -- I just always lumped it in my head with \scalebox and friends.

@WillRobertson :)

@WillRobertson Being serious, we are in a way moving toward that sort of case: expl3 is getting reasonably wide use

More seriously, in a sense, with both XeTeX and LuaTeX we could query the glyph bounds of the asterisk and dynamically calculate how much to shift it. If we were doing that we'd also want to scale it up (but probably using font commands rather than graphicx ones)

@JosephWright Thanks to you in very large part :)

@WillRobertson but if doing that doing it in \textasciicentered is presumably the wrong place and you could just fix up a mapping or whatever so the 2217 slot worked?

@DavidCarlisle In LuaTeX I guess so. In XeTeX not so much.

@WillRobertson get harfbuzz working in luatex....

@PauloCereda In Italian "papera" (= duck) also means lapsus, gaffe; whereas "oca" (= goose) also means "stupid woman". We don't have double meeting for "anatra" (= duck, too), instead.

@CarLaTeX the promised pictures of penguins in normal life are made right now. Stay tuned for the results at the end of day :-)

Hi anybody. I have the following code:

I interpret it as follows:

First, the catcodes for % and * are changed. % is made an other sign, * is made a comment sign.

Therefore the next line "\gdef\processLineX%#1{" would normally be written with a "%" at the end of the line. Now the "" is the comment sign and will prevent the white spaces from acting on the following code or being printed out.

@WillRobertson scary thing is that in principle we could end up with every command defined in tuenc being guarded by an \iffontchar test :(

It will globally define the macro \processLineX. TeX has to search for the argument which has to begin with a percent sign. The next token will be stored in #1.

@DavidCarlisle Nah, TU should just be "sensible default"...

@WillRobertson but I thought 2217 was a sensible default asterisk, but apparently not sensible enough....

The expansion for the macro is in the {} and is mainly an ifcase.

\ifcase should check a given macro for integer values beginning from 0 up to ...

@Jan can not follow your comments in this form. First, the catcodes for % and * are changed. % is made an other sign, * is made a comment sign. is true, after that I have no idea what you are asking.

@DavidCarlisle I'd say these are emergency measures

@Jan Yay!

@DavidCarlisle David, I beg your pardon.

@WillRobertson do you want to reply to my team mail suggesting the font size change version?

@Jan It's a simple test really: 'What is the nature of the char grabbed?' followed by 'Do X depending on what it is'

I am asking for your (or others) help to tell me, if I understood the macro code correct.

@Jan no need to apologise, I just meant I could not work out what you were asking

@Jan It is just handling the difference between % and %% processing in docstrip

@Jan It tests if #1 is a %, a <, a ! or 'something else', and chooses the code path appropriately

@JosephWright Yes. Thats what I understand.

a % (that is %% in total) is \putMetaComment, a < is \checkOption and ! means \expandVariable, correct?

@DavidCarlisle Okay — I wasn't sure if you'd take it seriously or not :)

@CarLaTeX Pato, in Portuguese, might also be a slang for naïve. And the expression pagar o pato (literally pay the duck) means something that you get responsible for, even if it's not your fault.

Sorry for the delay, I was in the mass. :)

@WillRobertson partly related to previous comment about how far this should go, it's better in some sense but we could also test for composites and use the composite only if it is there and and \accent if it is not (which may reduce the need to use the tuvariant encodings feature) but ....

@PauloCereda like Joseph and beamer?

@DavidCarlisle Yes, a nice example. :)

@Jan yes

A normal \ifcase would be

but the tokens %, <, and ! are not integer values.

@Jan well not exactly as it is testing characters not numbers and you'd need some \expandafter`

Therefore, the \ifx needs to convert the tokes into integers.

Sorry, if my question was sluggish. Thats what I meant. ifx gets the tokens and returns (expands) them to integers.

@Jan oh sorry misread your comment yes nested in an ifcase here so the inner ifx are returning integers

@DavidCarlisle :-)

@DavidCarlisle Thank you very much David!

It took me half a day, to resemble that code and understand, what it is doing exactly.

I am glad, I understood it now. I will be glad, if I am able, to write such code myself :-)

@CarLaTeX Don Carlos short after breakfast.

@Jan Un pengüino :-)

A mexican penguin. But don't tell Donald. He is an immigrant ;-)

@DavidCarlisle If we were to provide a standard solution along these lines, I would prefer to do it when first loading the font. And in LuaTeX this can be done seamlessly by constructing a virtual font on the fly. So for the time being I think what we have is a good approach... after all, it's only providing an equivalent to xunicode, which no-one has really complained about for 10-odd years :)

ah dear, late again. see you all around

@WillRobertson yes agreed, if we were luatex only I think fixing up a virtual font so all needed slots are filled would be better than fixing up csname definitions but where we are now i agree that a definition as above is best we can do

@WillRobertson good night:-)

@Jan Hahaha fantastic! And it's huge!

Spotted on c.t.t. (thanks JF):

confluence.cornell.edu/display/arxivpub/2017/02/10/…

@DavidCarlisle @WillRobertson the fun seems to be over but my two cents: imho some fallback code like the one from Will in such emergency case is quite ok but I wonder if one could add some debug code so that one can easier check where the fallbacks are used, e.g. through a parallel TU-debug encoding ...

@JosephWright Wow! From TL 2011 to TL 2016 in just one big jump!

@egreg I had some email from them about that earlier in the year they've been doing pixel for pixel comparisons of rendered pages, they picked up a few things...

@WillRobertson -0.75ex would be better

@egreg I assumed he'd done .5*1.2

@DavidCarlisle Apparently not. ;-)

x\textsuperscript{\textasteriskcentered}%
{\termes\textsuperscript{\textasteriskcentered}}%

\textasteriskcentered{\termes\textasteriskcentered}

@barbarabeeton Do you know why the search in the various mailing lists archive disappeared? E.g. tug.org/pipermail/xetex Is there somewhere a replacement?

@egreg oh above it's scaled by 1.2 but still shifted by .5ex, in a version by mail the shift was .6ex (hence my .5*1.2 comment)

@UlrikeFischer All there for me

@UlrikeFischer Oh, the search: ask Karl

@JosephWright -- this year, there's another reason to get names in early for tug membership. it's an election year, and only members can vote. in fact, the slate of candidates has been posted (just yesterday), and voting is open.

@barbarabeeton Will need to really read the statements now to work out how to case my votes!

@JasperLoy -- somehow, earlier editions of "more math" got overlooked. but there is a review of the 5th edition, added january 19. "guide to latex" is also there, although it's not near the top of the books page. you just have to keep looking.

@JosephWright -- you're not the only one. i know all these people, and it's still a hard choice!

@barbarabeeton For me, there are some 'definite yes' people but that makes the choice from the rest harder still ...

@WillRobertson -- actually, that's more expected of you than a tie!

@WillRobertson @egreg guess I need to use \setbox/\raise after all, hard to lose (just) the depth using \raisebox

123\fbox{*}\fbox{\textasteriskcentered}abc

\DeclareTextCommand{\textasteriskcentered}\UnicodeEncodingName{%
  \iffontchar\font"2217 \char"2217 \else
    \begingroup
      \fontsize
       {\the\dimexpr1.2\dimexpr\f@size pt\relax}%
       {\f@baselineskip}%
      \selectfont
      \setbox\z@\hbox{\raise-.6ex\hbox{*}}%
      \dp0\z@
      \leavevmode\box\z@
    \endgroup
  \fi
}

@UlrikeFischer -- hmmm. very interesting, and not explainable by me. i will inquire.

@DavidCarlisle \raisebox has optional arguments for setting the apparent height and depth, I thought you knew. :P

@egreg yes but not sure it's any easier, I don't want to set the height just the depth (which is the second optional) I was going to do that but ... On the other hand probably should in a kernel file

@MaestroGlanz not so different from the standard \@defaultunits using .001\textwidth rather than 1pt as the default?

@egreg \raisebox{-0.6ex}[\dimexpr\height-0.6ex][0pt]{*}% I suppose?

@DavidCarlisle Exactly

@DavidCarlisle Left Times New Roman, right TeX Gyre Termes

@egreg i suppose it is better, even if it does cause tex to go round 5 sides of a square to get to where it needs to get:(

@DavidCarlisle If you want to be fussy, \raisebox{-.6ex}[\dimexpr\height-.6ex][\dimexpr\depth-.6ex]{*}

@MaestroGlanz can not change the actual units that are accepted such as pt from the macro layer, that would have to be done in the engine (which has no knowledge of \textwidth)

@egreg depth of 0 seems safer.

@DavidCarlisle Of course, if a font realizes the asterisk not as high as usual…

@egreg which is one reason I stuck to .6 rather than .7 or more as the shift, the more you shift by default the more it's likely to do something silly on unknown fonts

@DavidCarlisle The .75ex was due to a different timing in the shift, I guess.

@MaestroGlanz you mean like this?

\makeatletter

\def\mgunits#1#2{\@defaultunits#1#2\dimexpr.001\textwidth\relax\@nnil}

\mgunits\dimen@{2in}\showthe\dimen@
\mgunits\dimen@{2cm}\showthe\dimen@
\mgunits\dimen@{2}\showthe\dimen@

\stop

It didnt use the tabs. This might be a bit unreadable sorry.

@barbarabeeton Thanks! I'm not sure if every archive had a search but I'm sure regarding the xetex list.

@MaestroGlanz well that's pgf specific of course using pgfparsing just to default the units would be too much but if ##1 is anyway ` pgf math expression fiar enough. the ifx##1 looks suspicious though, what is ##1 here?

@DavidCarlisle Based on \BooleanTrue looks like it's from xparse, but of course not really being used correctly ...

@JosephWright that was my guess:-)

@barbarabeeton Wow, that's good to know.

Every time the PGF manual is updated I want to see how many pages it has increased by.

Wow, I finally got a star in this room. This means a miracle is about to happen in my life.

@JasperLoy perhaps related to the fact that you seem to have become the moon

Everyone getting on with the important vote? tug.org/election/2017/candidates.html

@MaestroGlanz yes clearly it's inside another macro hence ## but I just meant that in general ifx\BooleanTrue##1 is safer than ifx##1\booleanTrue unless you have very close control over the possible values of '##1

@DavidCarlisle And the correct test here is \IfBoolean(TF) anyway ...

@JosephWright that was part two:-)

@DavidCarlisle I do. It is from DeclareDocumentCommand

stared version vs non-stared

@MaestroGlanz Yes sure, but the official way this should be used is testing with \IfBoolean(TF) and (ideally) that appearing directly in the code implementing your document command

@MaestroGlanz then as Joseph says there are defined conditionals for that, you shouldn't need ifx tests

Good point. I put it on my todo list.

@JosephWright @DavidCarlisle Has something changed in, possibly, xparse? My code has suddenly broken.

\NewDocumentCommand\newfref{ O {english} m m }{%
  \fancyrefaddcaptions{english}{%
	\expandafter\NewDocumentCommand\csname Fref#2name\endcsname{}{#3}%
	\expandafter\NewDocumentCommand\csname fref#2name\endcsname{}{%
	  \expandafter\MakeLowercase{#3}}%
  }%
  \expandafter\NewDocumentCommand\csname fancyref#2labelprefix\endcsname{}{#2}%
}

@cfr quite a bit changed, but what exactly broke? given that a high level description of xparse is "tex programming without \expandafter " a bit sad to see so many \expandafter there:-)

Given that the last time I touched this file was in 2015 and I load it in almost every document I compile, something else seems to have changed. Or I'm missing something.

@cfr this doesn't give an error ?

@DavidCarlisle Yes, well. I probably wouldn't write it this way now. But this is not at all recent. (The 2015 file touch date is unlikely to have concerned this code. I probably just added a new prefix.)

\documentclass{article}

\usepackage{xparse}



\NewDocumentCommand\newfref{ O {english} m m }{%
  \fancyrefaddcaptions{english}{%
	\expandafter\NewDocumentCommand\csname Fref#2name\endcsname{}{#3}%
	\expandafter\NewDocumentCommand\csname fref#2name\endcsname{}{%
	  \expandafter\MakeLowercase{#3}}%
  }%
  \expandafter\NewDocumentCommand\csname fancyref#2labelprefix\endcsname{}{#2}%
}
\begin{document}


\end{document}

@DavidCarlisle Well obviously \NewDocumentCS is missing ;-)

\documentclass[british]{article}
\usepackage{babel,fancyref,xparse}
\NewDocumentCommand\newfref{ O {english} m m }{%
  \fancyrefaddcaptions{english}{%
    \expandafter\NewDocumentCommand\csname Fref#2name\endcsname{}{#3}%
    \expandafter\NewDocumentCommand\csname fref#2name\endcsname{}{%
      \expandafter\MakeLowercase{#3}}%
  }%
  \expandafter\NewDocumentCommand\csname fancyref#2labelprefix\endcsname{}{#2}%
}
\newfref{claim}{Claim}
    \frefformat{vario}{\fancyrefclaimlabelprefix}{{\frefclaimname}\fancyrefdefaultspacing#1}

^^ @DavidCarlisle

! Missing \endcsname inserted.
<to be read again>
                   \tex_let:D
l.159 ...fformat{vario}{\fancyrefclaimlabelprefix}
                                                  {{\frefclaimname}\fancyref...

?

@UlrikeFischer How am I meant to do it?

@cfr It was a joke.

@UlrikeFischer I figured. But if this is the wrong way, I'd still like to know the right way. Don't like to disappoint @DavidCarlisle.

@cfr There is no better way in xparse as it hasn't commands to define csnames.

@UlrikeFischer Thanks. I work on the assumption I don't know the obvious. Very occasionally, the assumption is false ;).

I can't figure out how to do this with \newcommand either. Possibly why I used \NewDocumentCommand originally. Damn.

I'll have a look, more @JosephWright and Bruno's area these days really:-)

@DavidCarlisle I pinged Joseph, too, I think. The following works:

\NewDocumentCommand\newfref{ O {english} m m }{%
  \fancyrefaddcaptions{english}{%
    \expandafter\newcommand\csname Fref#2name\endcsname{#3}%
    \expandafter\newcommand\csname fref#2name\endcsname{%
      \expandafter\MakeLowercase{#3}}%
  }%
  \expandafter\newcommand\csname fancyref#2labelprefix\endcsname{#2}%
}

but I'm wondering what else is going to be broken ....

@DavidCarlisle Thanks.

@cfr you just removed the {} on the last line?

@DavidCarlisle No. I abandoned \NewDocumentCommand. So I remove the {} because the syntax of \newcommand is different.

@cfr @DavidCarlisle @JosephWright As far as I can see if I compare it with my miktex (which still has the older xparse) the handling of definition without arguments has changed. \fancyrefclaimlabelprefix is no longer a "simple" variable. But I wouldn't call it a bug. Imho it is not correct to use \NewDocumentCommand for such things. \newcommand or \def or \tl_set:n are more sensible. Btw: You don't seem to use the optional argument #1.

@cfr oh yes sorry was just skippming over it

@UlrikeFischer Because of the need for \expandafter? Presumably not just because of the no arguments.

@cfr No. You get the same problem with plain command names.

\documentclass{article}
\usepackage{xparse}
\NewDocumentCommand\test{}{blub}%
\show\test
\begin{document}
  something
\end{document}

In miktex (old)

> \test=\protected macro:
->blub.
l.11 \show\test

In TeXLive (new)

> \test=\protected macro:
->\__xparse_start:nNNnnn {}\test  \test code {}{}{}.

@cfr You can (ab)use xparse

\usepackage{xparse}
\ExplSyntaxOn
\NewDocumentCommand{\newnamecommand}{sm}
 {
  \IfBooleanTF{#1}
   {
    \exp_args:NNc \newcommand*{#2}
   }
   {
    \exp_args:Nc \newcommand{#2}
   }
 }
\ExplSyntaxOff
\newnamecommand{foo}{}\show\foo
\newnamecommand*{baz}{}\show\baz

\NewDocumentCommand\newfref{ O {english} m m }{%
  \fancyrefaddcaptions{english}{%
    \newnamecommand{Fref#2name}{#3}%
    \newnamecommand{fref#2name}{\MakeLowercase{#3}}%
  }%
  \newnamecommand{fancyref#2labelprefix}{#2}%
}

@UlrikeFischer @cfr as I recall part of a move to keep the argument specification with the command (even in the no argument case) rather than storing them separately but iIdidn't follow the discussions that carefully;-)

@cfr But it is not related to the number of argument, { m m } gives a "simple" macro too in miktex, but as soon as optional arguments are involved the macro is more complicated.

@JosephWright (@barbarabeeton) has there been any announcement or email about that?

@DavidCarlisle I checked and there's no dowvoting

@egreg I have no idea what my username or password are so no upvoting either:-)

@DavidCarlisle Mail TUG, they're fast.

@egreg yes will do that:-)

@DavidCarlisle -- not yet -- there's a question about who is to send it. regarding your username and password, send a message to the tug office to get that info.

i'm about to go offline -- theater tickets. if it isn't called off on account of the snow ... which has already deposited about a cm. (prognosis -- 10-25cm, but it keeps changing. blizzard watch on again.)

@barbarabeeton I wonder how many of your countrymen would report snow depth in cm without being prompted:-)

@DavidCarlisle -- probably nobody, but i've gotten used to communicating with you guys. 'bye now.

@barbarabeeton bye. Enjoy the theatre

@JosephWright thanks for the reminder, voted!

@UlrikeFischer That was always an implementation detail

@cfr As others have noted, looks like you were relying on how we were implementing commands created by \NewDocumentCommand with 'simple' arg specs: that was never supported

@egreg Did you see JF's c.t.t. post?

I created an Instagram account. I am cool now, right?

@PauloCereda unless you put yourself in a freezer, you're not cold cool. (Oh, a frozen duck...)

@PauloCereda I did that: I can't find a point ...

@yo' ooh :)

@JosephWright I follow you, I was expecting a more positive feedback. :)

@JosephWright: as it happened with WhatsApp, I suffered a female peer pressure.

@PauloCereda My sister uses WhatsApp so there I do see a (small) point

@JosephWright I have some good contacts too. :)

@JosephWright: github.com/gshslatexintro They are using L3!

@DavidCarlisle Will's photo is fake! There's no mug.

@JosephWright Have you been watching? Ryan Day made the finish very interesting, getting from 9:3 to 9:7, the final is 10:7 and Barry Hawkins got the title.

@JosephWright No. Link?

@JosephWright Hello :) I see some changes to the xparse package, are they just the e argument type and the “self-referential” possibility in the optional arguments?

Hou could one define something like the old \NewDocumentCommand \integral { k_ k^ } { \int \IfValueT{#1}{_{#1}} \IfValueT{#2}{^{#2}} }?

@egreg I'll look one up: I read by traditional Usenet so not so easy ...

@Manuel Currently needs an auxiliary

Personally, I think the k arguments should be different from a “multiple-option single-token” argument. So, first, I don't entirely understand what e does; plus in my opinion there was another type missing in the past. And now I see that a mix of those is implemented, so I would like to understand (if it's fast) how it works :)

@egreg groups.google.com/forum/#!topic/comp.text.tex/q2kLbHvW7d4

@Manuel Probably one for LaTeX-L, really (so other members of the team can contribute)

@Manuel Broadly, the issue is that we are more-or-less simulating the primitive _/^ syntax but viewed as arguments: they are freely-ordered

@Manuel We (team) have argued quite a bit about what makes most sense and I don't know that we are 'done' yet: this is experimental

@JosephWright Still, I'm interested in how would that example work, just to put it to use, in case it's easy.

@Manuel Notably, almost all document commands fit into the s o m framework so this is something special

@JosephWright Yes, I've been “out” of TeX for some time. Sorry to bother here.

@JosephWright Okey, I can imagine something like { \int \IfValueEq{#1}{_}{_{#1}} \IfValueEq{#1}{^{#1}} } (not working as of now, but something in that direction may be?)

\documentclass{article}
\usepackage{xparse}
\NewDocumentCommand\integral{e{_^}}{%
  \integralaux#1%
}
\protected\def\integralaux#1#2{%
  \int
    \IfValueT{#1}{_{#1}}%
    \IfValueT{#2}{^{#2}}%
}
\begin{document}

$
\integral
\integral^a
\integral_b
\integral^c_d
\integral_e^f
$

\end{document}

or { \int \IfTokenCase{#1}{ {_}{_{#1}} {^}{^{#1}} } }?

@JosephWright Okey, understood. Did you at least leave k available? (even if it's going to be deprecated in favour of this, it might be useful)

@Manuel No: the whole point of 'experimental' here is we need some 'space' to explore ideas. If we try to keep everything, even on a deprecated basis, we'll get bogged down entirely.

@Manuel How do you see k being useful in place of e^ e_ if you want a fixed order?

@JosephWright Okey, understood.

@Manuel @DavidCarlisle was keener on going for something like e_ e^ implying a variable ordering: it's tricky as I can see arguments both ways

@Manuel Really this is experimental: I would strongly recommend against using it in code you expect others to rely on!

@JosephWright Why exactly is the auxiliar function needed? Why not \IfValueT inside the main function? Sorry, I just thought about that :)