1:20 AM
Does anyone have the quick-and-dirty to choose what fields are displayed in a biblatex fullcite?
(Or a manual page? ;))
Just trying to get rid of DOIs and such without altering my bib file (which I share with my thesis)

2 hours later…
3:12 AM
please guys can you help me? tex.stackexchange.com/questions/174119/… i have report tomorrow to submit :(

@mrSuperEvening It would help if you could provide some sort of MWE

3 hours later…
5:44 AM
@StefanKottwitz Can i really edit after such a long time? Maybe @JosephWright can edit chat.stackexchange.com/transcript/message/15220375#15220375

2 hours later…
7:33 AM
@Johannes_B I could edit, I just don't know which spelling was meant (deities is correct, or? Additional s to Thanks?

8:32 AM
(./sample2e.tex
LaTeX2e <2014/05/01>
3

@DavidCarlisle Cool
@DavidCarlisle Will have the generic loader for expl3 done today: most of the mech is sorted, just have to do driver code and testing

@JosephWright Planning to advertise the zip files on dropbox on latex-l in a bit. let people have a day or so before we push to ctan and tl2014 test.....

@DavidCarlisle Cool

@DavidCarlisle Nothing on the team list :-(

8:37 AM
@JosephWright in a minute:-) just logged in, Frank and I had a quick "are we done yet" conversation last night:-)

@DavidCarlisle Ah

@DavidCarlisle :-)

@DavidCarlisle can you elaborate on your comment about not being able to create a vbox environment.

@DavidCarlisle Pretty much everything can be a verb actually, I've googled that. :)

8:43 AM
\documentclass{article}

\begin{document}

\begin{vbox} hello \end{vbox}

\end{document}

I'm slow today.

The above makes
\$ pdftex \&./latex vb
This is pdfTeX, Version 3.1415926-2.5-1.40.14 (TeX Live 2013)
restricted \write18 enabled.
entering extended mode
(./vb.tex
LaTeX2e <2014/05/01>
Babel <3.9k> and hyphenation patterns for 54 languages loaded.
(./article.cls
Document Class: article 2007/10/19 v1.4h Standard LaTeX document class
(./size10.clo))
No file vb.aux.
! Missing { inserted.
h
l.5 \begin{vbox} h
ello \end{vbox}

@SeanAllred does \usepackage[backend=biber, doi=false]{biblatex} do the trick?

@StrongBad even if you have an older latex than that. And if you use \renewenvironment{vbox} to makke an environment that doesn't d that you will redefine \vbox and break everything.

@DavidCarlisle so it is the fact that redefining \vboxwill break everything.

8:47 AM
@StrongBad well that and the existing definition can not be used as an environment. The comment just couldn't be more wrong really.

@DavidCarlisle Format working here with some 'real world' documents :-)

@JosephWright wow you have real latex documents?
2

@DavidCarlisle Yes: tested with a grant proposal

@DavidCarlisle thanks

@DavidCarlisle Have a paper in LaTeX too: a first for me as it's been written by physicists and they, unlike us chemists, don't use Word

8:52 AM
@JosephWright I wish I were a quantum physicist working on multiverse stuff, so in any universe my paper is already written and accepted.
:)

@PauloCereda you'd probably be in the alternative universe where it was already rejected.

@DavidCarlisle That's true at the moment. sigh :(

@PauloCereda oops sorry:(

@DavidCarlisle Nah don't worry, I'm used to that. :)
@David: for the first time, the reviewers liked my English. Something is very wrong. :)

@PauloCereda you've been chatting to too many Italians and improved your English grammar probably.

8:58 AM
@DavidCarlisle LOL don't forget the Englishmen. :)

@PauloCereda Never learn grammar from an Englishman.

@DavidCarlisle :)
@David: speaking of Englishmen, I had to defend your honour yesterday. :) Someone was complaining about LaTeX tables, so I had to prove him wrong. :) I managed somehow to properly pronounce your last name correctly. That's my greatest achievement since how should I pronounce Mittelbach the German way. :)

@DavidCarlisle Though I think the claim that no contributed package could define such a thing is too strong. A simple patch to \begin and \end should do nicely.

@StephanLehmke I think with the accuracy expected of a comment on tex.sx saying you cant define a vbox envioroment means you cant do \renewcommand{vbox}{} without bad things happening. Yes you could patch \begin or tex-the-program instead but....
@StephanLehmke one thing I have learnt is that there is no "simple patch" to latex. If you add \relax (or {}) or ... to any command anywhere some document somewhere will have been patching the old definition to add \relax and will break.
3

9:15 AM
@DavidCarlisle So you're saying such a contributed package could be incompatible with another one? Shocking ;-)

9:36 AM
Heya
Man, I still can not get my citations correct. "Minimal" here
\usepackage[sort,numbers,square]{natbib}
\bibliographystyle{plainnat}
\bibliography{Refferences}
I could post a true MWE, but I feel there must be something silly and minor I am missing :p

@N3buchadnezzar presumably the bib file is wrong for [4] it looks like you have title="title=...

I posted the bib file in the pastebin link, at least for the two relevant citations

@N3buchadnezzar which key is [4] ?

Laplace1

@N3buchadnezzar looks ok perhaps one above had a missing { or " and threw the parsing out? also can you do Poincaré in bibtex these days? rather than {\'e}

9:50 AM
@DavidCarlisle Did not think about it, but seems like it.
I only avoid using æøå in biblatex, and change the rest if I do get an error.

@N3buchadnezzar actually if it's utf8 it'll probably just pass through, if you used it in a sort key ordering would mess up but there it's just text

Yeah, I have no idea what messes up the bibliography. I checked all my other citations twice. None seemed to be missing any " { or }. No additional title = "title = ... " either.
Perhaps I am simply haunted

@N3buchadnezzar r-order your bib file see if it's always that one that's wrong or if it changes

3 hours later…
1:02 PM

1:22 PM
@SeanAllred I don't get a DOI with
\documentclass{article}
\usepackage[backend=biber, doi=false]{biblatex}
\begin{document}
\fullcite{kastenholz}
\end{document}

2:00 PM
@StrongBad Yeah—you're right. It would seem then that \printbibliography doesn't internally use \fullcite

2:10 PM
@SeanAllred I am not sure what you mean? tex.stackexchange.com/questions/43196/…

2:42 PM
Test material for LaTeX2e release: dropbox.com/sh/ms53wpmpo36ycpb/xLSAYzzhs2
3

@JosephWright did it appear on latex-l ?

Anyone interested please try (e.g. @egreg, @heikooberdiek)
2
@DavidCarlisle Not yet
@DavidCarlisle Will repost if it doesn't pop up over the next couple of hours

@JosephWright hmm perhaps latex-l doesn't like me, wasn't a good sign that I couldnt see the archives:-)

@DavidCarlisle Other people have had issues, I know

@JosephWright wish me luck: driving to london: the tube strike won't make it busy will it?

2:44 PM
@DavidCarlisle Have fun!
@DavidCarlisle Has popped up on LaTeX-L :-)

@StrongBad I suppose that's the same concept—

@JosephWright @egreg will say it's not music but classicfm.com/concerts-events/classic-fm-live
@JosephWright ah so it does like me after all

@SeanAllred So what exactly is the problem you are trying to solve? Do you have an MWE?

@StrongBad I'll get one up soon as a proper Q and let you know :)

@DavidCarlisle Will they bring and fire guns for the 1812 Ouverture?

3:37 PM
@DavidCarlisle It did now. :)
OMG guys, @Heiko is here! /we are not worthy

@JosephWright It is intentional that many source files (I have not checked all) contain CR/LF line ends instead of LF?

@HeikoOberdiek That's odd
@HeikoOberdiek Frank did the modifications of the L3 build scripts: the latter normalise line endings
@HeikoOberdiek Looks like Frank didn't retain the appropriate flag for the zipping step: I'll check with him

3:53 PM
@JosephWright Another hint for us to write a build tool? :)
Preferably, one that sends emails to @David when something goes wrong. :)

@PauloCereda The L3 scripts work pretty well: I can't help it if someone alters things

@JosephWright That means it's not intentional and will be fixed? I am asking because of latex-tds. I can add a normalization of the line ends there, but if it will be fixed, then there is no need for it.

@HeikoOberdiek Give me time to check with Frank: he's not got -ll or -X, and I'd say we should have both
@HeikoOberdiek The whole point of testing is to address this sort of thing

@JosephWright Hey let me dream of writing complicated stuff. :)

@HeikoOberdiek This is the first time we've used the current set of build scripts (cons is a pain for non-Unix people, which currently seems to mean Frank and David, and me on my work system)
Ah, @FrankMittelbach :-)

4:02 PM
@JosephWright have to be in some calls for the next hours ... feel free to update the make

@FrankMittelbach Am doing now :-)

Firefox with Australis. Hm.

@HeikoOberdiek I think there you have the answer: I'll alter the script to normalise line endings

@JosephWright Many thanks.

4:22 PM
Hi, has anyone of you run into sudden problems related to (apparently) encoding problems due to the files being exchanged between Linux and Win 7 (all via git)?

@henry 'via Git'?

Yup. I use my remote repo on my desktop (xubuntu 14.04) and on my laptop (currently win 7).

@HeikoOberdiek Fix for that checked in
@henry How do you actually transfer the data?

5:01 PM
@henry Are you sure that both linux and win7 editors use the same encoding by default? I may assume linux -> utf-8 and for win7 some other, which may cause you troubles.

@JosephWright I use git pull and git push but on Windows mostly the GUI
@s__C Yes absolutely, I can say that with certainty since I use Texstudio on both OSs. It's set to UTF-8 on both.

@henry does the encoding problem also appear when you open the file on github (online)?

@s__C No, it does not not appear.
@s__C (it's on bitbucket btw although this isn't really relevent)

@henry: sorry for the "dumb" question: why do you suspect it's an encoding problem?

@PauloCereda imgur.com/ApGNhnF
@PauloCereda And this is the line: \addplot [sharp plot, no marks, x=Wegnormiert] table [col sep=tab] {K04/Messdaten/Wegmessung-Kalibrierung-PunkteWolke.txt} coordinate [pos=0.5] (A) coordinate [pos=0.6] (B);
\draw (A) -| (B);
That popped up all of a sudden. Suddenly the machine ("it") doesn't seem to know the 'coordinate' key in TikZ anymore. Something is going on there.

5:42 PM
@henry do you have an MWE? as it stands it should normally work I think

6:14 PM
does anyone know if there is a way to get RAL colours in latex?

If they are spot colors, then the spotcolor package is your friend.

not sure what spotcolors are en.wikipedia.org/wiki/RAL_colour_standard but I'll look into it, for some reason google thinks I mean paint when combining latex and RAL in a search request :P

@JosephWright @FrankMittelbach base/README contains 17 bytes with link 00readme.txt. This automatically gets resolved, when it is uploaded to CTAN?

@HeikoOberdiek I fear not (I thought I removed the README)

@FrankMittelbach Probably poke Robin?

6:28 PM
@JosephWright @FrankMittelbach Would it be possible to remove the space character at line ends, too? I do not know the build process and how complicate this feature request is.

Hi, is this answer the recommended best way for resizing tikzpicure? It's the highest rated answer.
60

You could try using the command \resizebox. You enclose the tikzpicture environment in one of those and then in is scaled. Like this: \resizebox{<horizontal size>}{<vertical size>}{% \begin{tikzpicture} \end{tikzpicture} } If you want the image to be scaled proportionally, you can giv...

6:50 PM
@FaheemMitha Depends on what you're after I suppose. The OP specifically wanted to resize the text as well, in which case I guess it's fine. If you don't want to resize the text, then you're better off with scale=<factor>, or setting the x and y unit vectors.

@TorbjørnT. Yes, I think I do need to resize the text.
If I don't, the picture won't contract enough.
Does anyone happen to know what \tikzsetnextfilename expands to?

7:12 PM
@JosephWright @FrankMittelbach base/utf8ienc.dtx contains a comment äöü that is typeset in environment macrocode. Therefore the doc driver needs \usepackage[utf8]{inputenc}.

@HeikoOberdiek I'll take a look
@HeikoOberdiek Also will look at this (specific files?)

7:28 PM
Anyone got an idea why this doesn't compile if you remove },% ? pastie.org/pastes/9125069/text
I'm getting The key 'fontspec/BoldFont' is unknown and is being ignored.
Google returns a single result and the answer was apparently "works for me after upgrading to tl 2011"
Tried both xelatex and lualatex
Fwiw I use Mac OS X 10.9.
and TL 2013.

@JosephWright base/utf8ienc.dtx is a little more complicate. If LuaTeX/XeTeX is used, then inputenc cannot be used, then it would need \usepackage{fontspec}.

7:45 PM
@StefanKottwitz Yes, the additional s was what i meant. :-)

and fontspec 2013/05/20 v2.3c

@JosephWright @FrankMittelbach @DavidCarlisle First serious bug: If someone uses \usepackage[ascii]{inputenc}, then he probably want to ensure that the source is 7-bit clean. However, now LaTeX throws an error message, if XeTeX or LuaTeX is used. (The same problem I will have with x-ascii.def of package inputenx).

Is there anything obviously wrong with this function/macro? I'm getting errors.
\newcommand{\makefigure}[6]{%
% \par
% \mbox{}
% \par
\begin{figure}
\centering
\tikzsetnextfilename{#2}
\resizebox{#3 cm}{#4 cm}{
\input{#5}
}
\caption{#6}
\label{#1}
\end{figure}
}
I'm invoking this as follows:
\makefigure{cdf}{Figure4-cdf}{8.5}{8.5}{CDF}{%
some stuff
%
}

8:00 PM
I guess SizeFeatures literally just doesn't do BoldFont, ItalicFont, etc.
How goofy

@JosephWright I have made a list with files and line numbers: latex-spaces.txt. 199 files with 5274 lines.

Welp I guess that's it for using fontspec

PS/PDF experts: I got an old PS file that, when converted to PDF through ps2pdf, everything looks fine, but the text is all messy when copied. I believe it's because of the PS conversion?

@PauloCereda Try first ps2ps

@egreg Ah thanks, let me check.

8:14 PM

@PauloCereda poke robin to do what, manually add links? not good I'd say

@FrankMittelbach I meant replace the deployment file, I guess. :)

@HeikoOberdiek I wouldn't call that a serious bug, but something to fix, I agree thanks
@PauloCereda well that I should be able to do myself

@FrankMittelbach oopsie, sorry.
@egreg: I got a better result, thanks. :) But apparently the PS file isn't salvageable at all. :)

@HeikoOberdiek what do you mean by removing space characters? a) normally we would try not to have any, but if they crept in they wouldn't get removed in the past and they are not removed now

8:20 PM
@PauloCereda Copy-pasting from PS wasn't even dreamt of in the olden times!

@egreg ooh good to know, at least I have an excuse now. :)
@egreg: By the way, you got mentioned in a Apache mailing list: mail-archives.apache.org/mod_mbox/commons-user/201404.mbox/… :)

@PauloCereda Found it. ;-)

@egreg :)

@FrankMittelbach I meant trailing spaces at line ends, but it is not an important issue. It could be done automatically, when correcting the line ends. Or I can do it in latex-tds' if you do not mind.

@HeikoOberdiek basically we check out from svn to native, eg when going onto windows we get windows line endings so when I produce the zips there we got the windos line endings. that is now changed simply by zipping them differently. We could have done this differently in which case we could have also looked at stray spaces at line ends, but as I said, if they are there now they have been there in the past most likely
so if you want to do this at latex-tds level, fine
@HeikoOberdiek @David any suggestion with respect to handling inputenc[ascii]?

8:38 PM
@FrankMittelbach They tend to 'build up' over time (certainly in expl3)

@JosephWright they sure do, but the don't do harm within TeX, do they?

@FrankMittelbach No: Bruno and I just get annoyed by them

@JosephWright then you shouldn't put them in :-)
but I know ... some cleanup perl would be in order
anyway, your zip change gives some interesting "info" now when building the distrib

@FrankMittelbach Yes?

@JosephWright any idea what that means
zip info: local extra (114 bytes) != central extra (17 bytes): distrib/base/
zip info: local extra (114 bytes) != central extra (17 bytes): distrib/base/00readme.txt
zip info: local extra (114 bytes) != central extra (17 bytes): distrib/base/alltt.dtx
etc

8:45 PM
@FrankMittelbach How do you get that?

when running the make ctan
it happens on all zips the second time around when the pdfs are added

@FrankMittelbach Ok, then I will remove the trailing spaces in latex-tds.

@JosephWright maybe just because of the missing -X
@HeikoOberdiek this time around it is probably the simplest solution

@FrankMittelbach I decided against -X: reading the specs, I think it's only useful on Unix
@FrankMittelbach Hmm, struggling to reproduce this. What do you get for zip --help?

zip -help
Type 'zip "-L"' for software license.

8:54 PM
@FrankMittelbach Could be a version issue then: I have
Copyright (c) 1990-2008 Info-ZIP - Type 'zip "-L"' for software license.
Zip 3.0 (July 5th 2008). Usage:

zip --help

zip error: Invalid command arguments (no such option: -)
@JosephWright you reproduce on what OS?
I guess you have to do it on win

@FrankMittelbach Activating all Unicode characters except the ASCII ones does not make much sense, because there are too many of them. The easiest patch would be ignoring ascii and x-ascii. I think notifying via \PackageInfo is enough in this case, may be a \PackageWarning, but no error message.

@HeikoOberdiek @DavidCarlisle we can certainly just accept the options with a warning/info that it is not really checking for anything on this engine

@FrankMittelbach Win7, 64-bit, Info-ZIP as part of TeX Live (Note: this is using Parallels 'inside' my Mac set up)
@FrankMittelbach I've tested a couple of the zip file builds at the command prompt: no sign of odd warnings

@JosephWright I'm fairly sure it is not an issue so yes could be a version issue
not seen that before and it shows up only now after you made the split

9:04 PM
@FrankMittelbach Like I said, I've never had an issue just with -ll and PDFs, but perhaps that's version dependent too ...

@JosephWright or luck

@FrankMittelbach Should I rebuild the ZIP files and drop into DropBox?

@JosephWright you could - we fixed most of what Heiko found but not the inputenc package yet
@JosephWright actually I call it a day ... was on the phone until half an hour ago .

@FrankMittelbach OK
@HeikoOberdiek How do you find all of those trailing spaces?

@JosephWright could be that that too is version dependent, after all the zip identifies those as binary files. but could also be that just none of your examples broke. in the distrib the first files I looked at worked but then some didnt

9:10 PM
@JosephWright Of course not manually ;-) I have written a script check-eolspaces.pl. Then I run find to collect the files: find . -not -name \*.pdf -exec ~/work/latex-tds/lib/check-eolspaces.pl {} \;

@HeikoOberdiek Ah, right

Hello everybody! Anybody else having issues logging in to Meta?

@HenriMenke No problem here in calling the Meta site as logged in user.

@HenriMenke No issue here

Is it not possible to turn the frenchlinks option of hyperref on/off after the \usepackage?
For example: \documentclass{article}\usepackage[frenchlinks=false]{hyperref}\begin{document}‌​\href{http://foo}{bar}\hypersetup{frenchlinks=true}\href{http://foo}{bar}\end{doc‌​ument} or \documentclass{article}\usepackage[frenchlinks=true]{hyperref}\begin{document}\‌​href{http://foo}{bar}\hypersetup{frenchlinks=false}\href{http://foo}{bar}\end{doc‌​ument}
In all cases I get: Package hyperref Warning: Option 'frenchlinks' has already been used,(hyperref) setting the option has no effect on input line 5.

9:22 PM
@henry sorry, had to run (some miscommunication, appointment wasn't postponed after all..), but now I'm too tired to really delve into that problem, only thing I found was this link that might give a clue somehow (also looking at the strange edef error in your log which has to be read vertically as the letters are not defined in nullfont)

@hftf No it is not possible. It can be indirectly turned off together with colorlinks and ocgcolorlinks by option hidelinks.

@HeikoOberdiek Sweet! I'm 3 for 3 today on silly things latex packages do.

@FrankMittelbach I've checked the zip files generated on Windows using the Mac-native version of zip: all pass OK
@HeikoOberdiek You will soon have your desired generic loader for expl3 :-)

I attempted to put an ! as one of the fields of \resizebox, as per
61

You could try using the command \resizebox. You enclose the tikzpicture environment in one of those and then in is scaled. Like this: \resizebox{<horizontal size>}{<vertical size>}{% \begin{tikzpicture} \end{tikzpicture} } If you want the image to be scaled proportionally, you can giv...

but i got an error. Can anyone confirm this is correct syntax?
Oh, I see. I have it in a macro with cm after it.

@JosephWright Sorry, I haven't answered the pdftexcmds/\pdfstrcmp request. My suggestion would be to copy more or less the code needed for strcmp. Then expl3 would not need the external package pdftexcmds. The real work would be catching errors (old LuaTeX versions, ...).

9:32 PM
@HeikoOberdiek I'd considered that approach: one reason we use pdftexcmds is that in package mode it's 'reasonable' to use pre-existing solutions if available and robust. We already have the code for format mode, of course.
@HeikoOberdiek Note we don't support LuaTeX < 0.40, so the version issue is less of a problem!

So why can't I do \begingroup\makeatletter\def\Hy@colorlink#1{\begingroup}\makeatother ... a link ... \endgroup
It complains that Use of \Hy@colorlink doesn't match its definition.

@hftf Perhaps you are using it inside the definition of another command and you get the top level argument as the parameter text. Then you have to double # or use \renewcommand*{\Hy@colorlink}[1]{\begingroup}.

@HeikoOberdiek Okay that works for me in a pristine document so now I will search for the next culprit
thanks :)
Oh. it was pandoc being silly

9:50 PM
@JosephWright: check for a game named Skullgirls Encore on PSN. :)

\def\removefrenchlinks{\expandafter\def\csname Hy@colorlink\endcsname##1{\begingroup}}
Thanks all

10:35 PM
Should \resizebox work with table? or do I need to apply it directly to tabular?

11:06 PM
@FaheemMitha table is a floating object that cannot be squeezed inside an horizontal box. Apply \resizebox to tabular.

@egreg not guns exactly but smoke lights a lot of fireworks and rockets and the band of the royal air force all making a lot of noise:-)
@HeikoOberdiek @HeikoOberdiek we could certainly make ascii act like utf8 in xetex although I must say I don't really see the use case yet (but I just got back from London and I'm tired so perhaps I'm missing something) although the behaviour of ascii in xetex and pdftex would then be so different is it really useful?

11:24 PM
@DavidCarlisle inputenx.dtx is using it. And what is wrong with portable documents, which restricts themselves to 7-bit characters and which can be processed by different TeX engines? Throwing an error message will get you more sooner than later a LaTeX bug report.

@DavidCarlisle The only good thing in the 1812 Ouverture are the guns!

@HeikoOberdiek yes I don't object to changing it if its being used, just musing that it doesn't actually ensure portable documents.
@egreg well in the round in the Albert hall it makes a lot of people jump. Clarinet was good too and gives M something to aim for if he's going to become rich and famous and support his father.

@DavidCarlisle I do not see that character codes >127 can efficiently be supported by making them active for throwing error messages if used. But the 7-bit control characters covered by inputenc can be supported also with LuaTeX/XeTeX at least in the case of ascii.def and x-ascii.def.

@HeikoOberdiek yes agreed, I guess we'll change it, no point in a preview release if we don't change it when people comment:-)

@DavidCarlisle Thanks. :-)

11:35 PM
@HeikoOberdiek That's what I thought. Thanks for the clarification.

@HeikoOberdiek given that utf8 is using PackageWarning so far what would you suggest the wording be for utf8 and ascii, current wording is
\PackageWarning
{inputenc}%
{inputenc not required for xetex or luatex.\MessageBreak
utf8 assumed}%

@DavidCarlisle You are the native English speaker ;-)

@HeikoOberdiek I'm going to sleep, I'll sketch something in the morning but you basically want on xe/luatex that utf8 and ascii are both accepted and make the package do nothing other than warning?

@DavidCarlisle Case asscii/x-ascii: (a) the old behavior for character codes <128, Thus 7-bit control characters should throw an error as before. Character with codes > 127 cannot be handled with LuaTeX/XeTeX in a reasonable manner. Thus these can be ignored IMHO and \inputencoding issues a warning/info that says that character codes >127 are not supported (or whatever wording).
@DavidCarlisle Sleep well.

@HeikoOberdiek do I need to do x-ascii (what does it do:-) I wondered about utf8 utf8x but only did utf8 on the grounds that (a) that was the one in the core distrib and more importantly xetex only has one behaviour so supporting multiple encoding options seemed logically wrong and just supporting utf8 seemed a justifiable choice but you are pushing back on that I think...

11:51 PM
@DavidCarlisle x-ascii.def comes from package inputenx. If it is used as \usepackage[x-ascii]{inputenc}, then there is no chance for inputenx to catch the error, because the error is already thrown by inputenc before any code of inputenx is called and in this case never called.

@HeikoOberdiek grrrr :-) I really am going to sleep now:-) If you have any more thoughts drop them here (I'll scroll back the log in the morning) or drop a mail to us or latex-l or whatever. Otherwise I'll draft something in the morning. Good night @FrankMittelbach ^^^^

@DavidCarlisle Yes, I think, package inputenc should also be aware of utf8x, the same reasons for utf8 are also valid for utf8x.
@DavidCarlisle I think that the warning messages for utf8(x) should mention fontspec`. Otherwise the user gets a warning and probably many missing characters without knowing, what to do.