Good maen

@ChristianHupfer Moin.

@Johannes_B: I am going through unanswered stuff in my favorite tags (either VtC or trying to answer)

@ChristianHupfer I am avoiding reading the bunch of 15 fb messages that piled up over the last two days.

@Johannes_B Being active here will prevent them there ...

@ChristianHupfer 15

@Johannes_B Yes, I've read the number -- scary

@ChristianHupfer reminds me of youtube.com/watch?v=ogci09kq_u4

@Johannes_B: I see you answered the 'annex' question ;-)

@ChristianHupfer i wanted to do this two days ago, but got distracted. The answer was already written.

And it is a crosspost. :-(

@Johannes_B: I hope the O.P. will appreciate this. He's not a member of the accepting squad ;-)

@ChristianHupfer One of my favourite members.

@DavidCarlisle Colour is different (there's a clear technical advantage to using attributes), though we'd need to think about a functionality split there

@Johannes_B: Some guys (or at least one bloke) tried to breach my WP site, hacking the superuser account and password ;-)

@ChristianHupfer Was it your alter ego?

@Johannes_B You mean the account name?

@ChristianHupfer yes.

@Johannes_B No. Stupid WP installations use admin as superuser account name -- Of course, I've not done this way. My superuser account is of course named 'administrator' and the password is 12345678 ;-)

@ChristianHupfer :-)

@Johannes_B: I have blocked a wider IP range, but this is of course not totally secure

So I guess we can expect lots of messages to get starred from 1200GMT today

@JosephWright Why?

@ChristianHupfer Hat

@JosephWright Ah, forgot about that. I've already plenty of them (and I think all by accident ;-))

@ChristianHupfer Same here (and I'm not even taking part)

@JosephWright: The community has been pretty active in closing questions during the last days

@ChristianHupfer Yes

@ChristianHupfer I do keep saying about it!

@DavidCarlisle <3

@PauloCereda Good morning to Brazil

@PauloCereda Hello

@ChristianHupfer Good morning!

@JosephWright Hello! :)

Arrrrrrrrrr... users /shaking head ...

shaking my fist ...

@ChristianHupfer Anger issues. :)

@PauloCereda I give in to my anger :-(

@ChristianHupfer <3

@PauloCereda What have users ever done for us :-P

Morning @DavidCarlisle

@DavidCarlisle Are you going to take on the luatex.def work initially?

Ok, not really a palindrome ... ;-)

@JosephWright yes can do

@JosephWright you've got mail. :)

@ChristianHupfer what the actual hell? looks like somebody upvoted nearly all my answers, too

@JosephWright should we check this in, or should we normalise away line numbers in lua errors in l3build?

@Rico No, I worked against the voices in my head that say click on the green reputation box ;-)

@ChristianHupfer oh ok :D

@JosephWright in tlb3476 l3build seems to be too regressive with normalisation in luatex show\mv@normal makes this which is more or less same as etex

> \mv@normal=macro:
->\getanddefine@fonts \symoperators \OT1/cmr/m/n \getanddefine@fonts \symletter
s \OML/cmm/m/it \getanddefine@fonts \symsymbols \OMS/cmsy/m/n \getanddefine@fon
ts \symlargesymbols \OMX/cmex/m/n \getanddefine@fonts {8}\OT1/cmss/m/n \install
@mathalphabet \mathsf  {\relax \ifmmode \else \non@alpherr \mathsf  \fi \use@ma
thgroup \M@OT1 {8}}\install@mathalphabet \mathit  {\select@group \mathit  \M@OT
1 \OT1/cmr/m/it }\getanddefine@fonts {7}\OT1/cmtt/m/n \install@mathalphabet \ma

but normalised log just shows

> \mv@normal=macro:
}}\getanddefine@fonts \symboldoperators \OT1/cmr/bx/n \getanddefine@fonts \symb
oldletters \OML/cmm/b/it \getanddefine@fonts \symboldsymbols \OMS/cmsy/b/n .
l. ...\makeatletter\show\mv@normal

@JosephWright with first part gone. I'll try to debug later but thought you may be able to read this more quickly:-)

s/regressive/aggressive/ in the above:-)

@ChristianHupfer I always do that to see which answer got an upvote.

Dear friends, we come here because of TeX, but we stay because of the community. Thank you very much for the wonderful friendship and great moments! 2016 will be a fantastic year!

@PauloCereda It's just like vim: once you are in, you can't get out

@DavidCarlisle :-)

@DavidCarlisle I'll take a look this afternoon

@barbarabeeton Received TUGboat this morning: excellent editorial

@JosephWright tools pass apart from that, cyrillic passes, I'm going to attack graphics after lunch, babel I have decided to leave:

  Check failed with difference files
  - ../../build/test/tlb-cite.xetex.diff
  - ../../build/test/tlb-french-hyphen.xetex.diff
  - ../../build/test/tlb-frhyph.xetex.diff
  - ../../build/test/tlb-hyphenc.xetex.diff
  - ../../build/test/tlb-natbib.xetex.diff
  - ../../build/test/tlb-showkeys.xetex.diff
  - ../../build/test/tlb-switch.xetex.diff
  - ../../build/test/tlb0735.xetex.diff
  - ../../build/test/tlb0868.xetex.diff
  - ../../build/test/tlb0977.xetex.diff
  - ../../build/test/tlb1048.xetex.diff

@JosephWright probably I should just configure the babel part of the test suite just to use etex. Not sure it's worth checking in that many etex (and then luatex) tlg files mostly for nonsense encoding errors for tests assuming inputenc

@DavidCarlisle babel tests use pdftex I think

@JosephWright yes looks that way (there are a three pdftex.diff in the above list, we should probably at least fix those, but eat first:-)

@DavidCarlisle LOL

@esdd: Happy New Year to you too!

@PauloCereda Star catcher? Hat catcher? ;-)

@ChristianHupfer I was being sincere. :)

@ChristianHupfer: ^^

I put my old blue hat back

@clemens Reading your chemistry article, I notice one thing I'd query. In the IUPAC naming, ' should end up as a prime not a quote mark. Have you considered handling that?

So is everyone busy reading TUGboat?

@JosephWright mine just arrived as we started lunch, not opened it yet:-)

@JosephWright I am eating toast. :)

And I need to break my piggy bank in order to pay the UK-TUG subscription.

@JosephWright did you know about the gmoa thing (just read first para so far)

@PauloCereda Two hours of sincerity :-P

@ChristianHupfer Quack! :)

Chapters made in LyX and TeXMaker calls the chapters? ... Wait? What????

@ChristianHupfer My thought as well.

@TorbjørnT. Very strange indeed -- I am no LyX or TeXMaker user, but this sounds .... weird

@DavidCarlisle Yes, Karl asked me to review it (and to write something actually accessible about the \romannumeral trick)

any German speakers around? Frue-h-auf-ste-her or Frue-hauf-ste-her (pdftex gives first xetex and luatex the second) I thought the utf8 hyphenation files would be more compatible than that....

@JosephWright ^^^^ tlb2101 :)

@DavidCarlisle There are two sets of patterns for German: one 'traditional' and one newer set. I think that the split is reflected in the engine choice.

@DavidCarlisle I think I said to Karl something like 'well it's bonkers'

Happy New Year, everyone!

@JosephWright I knew there were two sets of hyphenation patterns but why different defaults at engine level?

@DavidCarlisle The discussion was somewhat complex: I'd need to dig through the set up to work it out

@JoelReyesNoche happy new year (well we still have over 8 hours to go here:-)

@DavidCarlisle I'll start on the tests in a second

@JoelReyesNoche Happy new year!

@DavidCarlisle Frueh-auf-ste-her -- the first h-letter can't stand alone (in my point of view)

@ChristianHupfer so you just volunteered to maintain/edit the german hyphenation patterns:-)

@JosephWright interesting discussion with Will and @UlrikeFischer here tex.stackexchange.com/questions/285337/… could possibly do something for 2016/01/01??

@DavidCarlisle You're so mean... I feel like being tricked to enroll to the Army :-P

@ChristianHupfer "It'll all be over by Christmas"

@DavidCarlisle As @JosephWright says:

> pdflatex
This is pdfTeX, Version 3.14159265-2.6-1.40.16 (TeX Live 2015) (preloaded format=pdflatex)
 restricted \write18 enabled.
**\makeatletter\language=\csname l@ngerman-x-latest\endcsname
entering extended mode
LaTeX2e <2015/10/01> patch level 2
Babel <3.9m> and hyphenation patterns for 79 languages loaded.

*\showhyphens{Fruehaufsteher}

Underfull \hbox (badness 10000) detected at line 0
[] \OT1/cmr/m/n/10 Frue-hauf-ste-her

@DavidCarlisle: Please engage this site: -P

silbentrennung24.de/?term=Fr%C3%BChaufsteher

But I was right (and I am a stupid Physicist only)

@egreg ^^ (with pdflatex or xelatex)

@JoelReyesNoche Happy new year!

@ChristianHupfer but you cheated and used the german dotty letter thing, if you use ue then that site agrees with xelatex and gives Frue•hauf•ste•her

@DavidCarlisle BTW, what 'internal pressure' are you expecting re. the 2e repo?

@DavidCarlisle: Do you really mean german?

@ChristianHupfer I mean use ue rather than ü

@DavidCarlisle True, but this isn't correct German spelling then. If a font does not provide Umlauts (yeah, we got them!!!) it's alright to write them this way but the pattern is wrong anyway

@ChristianHupfer the actual test file had the version with and without the umlaut checking the babel "u shortcut enabled hyphenation, but it just hit my screen today as the "no accent" base test is giving different results in xelatex

@ChristianHupfer But ü is just a ligature for ue

@JosephWright is it?!

@PauloCereda Well OK it's historical, but ü for German developed from the blackletter representation of ue

@JosephWright I was worried for a moment, since ü is not a ligature for ue in Portuguese. :)

@JosephWright If it was just suggestions to switch from svn to git easy to ignore but if the entire website is already on github, and everything checked in to svn is mirrored to github then the advantages of maintaining a separate more or less private svn rep are somewhat small and the extra maintenance hard to justify, really.

@JosephWright That's true, but I think Frue-h- is wrong, since h can't stand alone in hyphenation (in German rules)

@DavidCarlisle The chosen language is 30 in all cases; looking at pdflatex.log and xelatex.log this should correspond to the same hyphenation patterns. Of course, in the xelatex case there's the UTF-8 translation.

@ChristianHupfer We (the team) don't control the hyphenation patterns, we can only make sure out code uses them correctly

@egreg but you wouldn't expect the utf8 translation to be doing much to an all ascii word....

@DavidCarlisle Well, in xelatex.log I see this

\l@german=\language30

(/usr/local/texlive/2015/texmf-dist/tex/generic/hyph-utf8/loadhyph/loadhyph-de-
1901.tex UTF-8 German hyphenation patterns (traditional orthography)
(/usr/local/texlive/2015/texmf-dist/tex/generic/hyph-utf8/patterns/tex/hyph-de-
1901.tex
German Hyphenation Patterns (Traditional Orthography) `dehypht-x' 2014-05-21 (W
L)))

@DavidCarlisle It seems that the loaded file is the more modern one.

@egreg yes as Joseph hinted, got to go, back later...

@DavidCarlisle I can see that for L3 (as it is mirrored) but 2e isn't so ...

@DavidCarlisle Just for information: in pdflatex.log we have

\l@german=\language30

(/usr/local/texlive/2015/texmf-dist/tex/generic/hyph-utf8/loadhyph/loadhyph-de-
1901.tex EC German hyphenation patterns (traditional orthography)
(/usr/local/texlive/2015/texmf-dist/tex/generic/hyphen/dehypht.tex
German Traditional Hyphenation Patterns `dehypht' Version 3.2a <1999/03/03>
(Formerly known under the name `ghyph31' and `ghyphen'.)))

@egreg Ah, that sounds about right

@JosephWright It should explain the difference

@egreg I meant 'that sounds familiar from the discussion on the mailing list'

@ChristianHupfer @DavidCarlisle @JosephWright You sound so unsure. Naturally Frue-h-auf-ste-her and Frue-hauf-ste-her is wrong, only Frueh-auf-ste-her is ok.

@JosephWright Following the patterns is good, but in this case I think the pattern is wrong. Fruehaufsteher contains the prefix "Frueh" which is actually not hyphenable and the rest of word where no problems occur

@UlrikeFischer: Well, I was a little bit unsure.

@UlrikeFischer I wouldn't make any statements on the 'correct' hyphenation in English, let alone a language of which a native speaker

@ChristianHupfer There was some discussion about what is and isn't allowable in terms of changing the existing patterns. The argument was that for pdfTeX there are far too many documents already in existence to risk changing the patterns even if they can be improved.

@JosephWright: German hyphenation rules follow speakable or meaningful syllables -- this is not the case for frue-hauf-steher

@JosephWright Well, a user can override the hyphenation anyway

@ChristianHupfer Yes, but if one has an existing document, changing the hyphenation patterns at all may alter line breaks, ....

@JosephWright Of course, but is that an issue for a thesis or an article really?

@JosephWright In this case hyphenation is clear for a german, the h belongs to früh/frueh, duden.de/rechtschreibung/Fruehaufsteher. There is really no doubts.

@UlrikeFischer Remember my native language is English, so the idea of a 'correct' hyphenation or indeed spelling is not really something I feel is real (it's a question of current consensus)

@ChristianHupfer Perhaps not, but for a book that's not the same

@JosephWright Sure, if the hyphenation pattern for particular exceptions has to be applied quite a lot. I don't think such a word like "Frühaufsteher" will occur a lot, even for a German book ;-)

I am off for a while. Happy New Year in advance to all!

@ChristianHupfer Sure but this is just an example from a long list of improvements/changes in the pattern files

@JosephWright ;-) There are also in german hypenation points that are only "current consensus" but as there are so many composed words the places between word parts are main hyphenation points, "früh" is a word and "aufsteher" too.

@UlrikeFischer @ChristianHupfer @JosephWright Anyway, lualatex does Früh-auf-ste-her

@ChristianHupfer tex.stackexchange.com/review/suggested-edits/86029

@egreg Yes, newer patterns

@egreg I'm not actually sure the hyphenation people have this right (not changing for pdfTeX) but that's their call

@JosephWright pdflatex does the same

\documentclass{article}
\usepackage[T1]{fontenc}
\usepackage[utf8]{inputenc}
\usepackage[german]{babel}
\begin{document}

\message{\language=\the\language}
\showhyphens{Frühaufsteher}

\stop

Underfull \hbox (badness 10000) in paragraph at lines 8--8
[][] \T1/cmr/m/n/10 Früh-auf-ste-her

@JosephWright Using the wrong spelling is not guaranteed to produce sensible hyphenation.

@egreg Like I said, ü is ue in German so I don't see this is 'spelled incorrectly'

@JosephWright I'm not sure German grammars would agree with you.

@egreg Possibly not, but the history of the umlaut is quite clear

@egreg In any case, what I'm not sure about is the idea that the hyphenation should be engine-dependent here

@egreg @JosephWright @ChristianHupfer Yes, I thought that with an ü it would be better, the pattern "üha" is more obvious then "eha". But german hyphenation can not be done with patterns only anyway, e.g. "Stau-becken" or "Staub-ecken" depends on the meaning. So I always expect some faults and check the hyphenation.

@UlrikeFischer Sure, you get the same in English: 'record' is probably the most famous example

For crying out loud, why? Why? Why are users not able to apply code markup?

@Johannes_B ?

@Johannes_B Good use of 'for crying out loud' there ;-)

@JosephWright BbCode is too confusing, some use backslashes to close, some disable BBCode. Some do'nt getthe hang of Markdown evn after months of asking questions. @stefan Can we change the tags for latex and mwe on LC? I have to edit constantly.

@Johannes_B Ah: I do find that a bit tricky myself (the BBcode business)

@JosephWright Yes, the naming is confusing.

A quickie TikZ question: How do I put a text at the beginning of the line (\draw)? I want it to start at the beginning of the line, above it.

The text
--------------------------------

@LazarLjubenović A test file?

Ah, after 20 minutes of searching through the site, as usually, I find the answer right after asking it. The "near start" and "very near start" is what I was looking for.

@UlrikeFischer Yes, I stated some time earlier, that writing ue is 'wrong' and using äöü is better.

@Johannes_B Too late for me.

@ChristianHupfer One of my favourite users.

@Johannes_B Did you notify him to keep away from such stupid edits?

@JosephWright -- urk. looks like we missed a glitch. sorry.

@barbarabeeton Well perhaps but I wonder if \iupac should deal with this anyway (not that many people will pick it up themselves)

@DavidCarlisle --both of these are wrong. frueh-auf-steher. (yes, i've seen the other comments.). happy new year from quebec city.

@ChristianHupfer It is allowed when no ü is available, also it is not really the point: There are other example where the patterns can't really decide if a consonant belongs to the first word or the next. E.g. pdflatex hyphenates "Zie-h-au-to". Newer patterns are sometimes better, but none are foolproof.

@ChristianHupfer The edit rejected provided me to flag the edit.

@Werner good number!

@DavidCarlisle A nested palindrome.

...if such a thing exists.

@barbarabeeton actually I don't care about whether they are wrong or right. But I'd be happier if they were the same.

@JosephWright, @egreg so the German patterns don't use the converted ut8 ones, they just load a separate set of patterns? OK I suppose that's just the way it is...

@DavidCarlisle That sounds familiar

@JosephWright been doing other stuff but just had a quick look at graphics/luatex.def, trickiest if we want to avoid adding all the pdfxx aliases is probably going to be the eps to pdf conversion which currently pulls in half of Heiko's stack...

@DavidCarlisle Hmm

@DavidCarlisle Can we work round that by using the Lua escape route?

@JosephWright possibly, I'll have a look exactly what it uses. A simple shell escape to the conversion script would work, depends if we want to keep all the options for just doing it the first time or whatever, which is tiresome to get exactly the same if we simply re-implement rather than share code. It is somewhat over-engineered, graphics.cfg has rather more requirepackage than when I last looked:-)

@DavidCarlisle Well had the kernel had some updates things might have been different ;-)

@JosephWright :-)

@DavidCarlisle Of course, part of this is because HO was left to do the work and took the path best for him (fair enough)

@JosephWright yes I'm thankful that he did, I had to drop out for a while, and it's good someone kept on top of pdftex developments. Just means that there's an awful lot of history in that file:-)

@DavidCarlisle I'm working on that l3build issue

@DavidCarlisle Indeed

@DavidCarlisle Shame I didn't start using LaTeX during my PhD ;-)

Or indeed my degree

@JosephWright but then you'd have had too many diversions and never got the thing written:-)

@DavidCarlisle Took 10 week-ish, I had something to concentrate my mind (job starting)

@JosephWright hmm my degree was 1982 so I suppose in principle I could have used tex.

@DavidCarlisle LaTeX was mentioned to me during my PhD but I didn't explore it at the time

@DavidCarlisle That l3build thing seems to be a Lua(TeX) issue, not us: if I simply write the formatted data to the terminal its fine

@JosephWright well, I'm shocked:-)

@DavidCarlisle Actually, I may just be looking in the wrong place: am continuing to trace the issue

@JosephWright almost got a plausible luatex.def, mismatched if somewhere from removing some of the pdftex version tests:-) but it includes some images and rotates some things (not looked at eps support yet)

@DavidCarlisle Great

@JosephWright I can do string replace in emacs:-)

@DavidCarlisle Think I've tracked down the issue

@DavidCarlisle Will fix then make sure it doesn't affect L3/L2e

@JosephWright just wondering if I should do an update to \ProvidesPackage{trig}[1999/03/16 seems a bit daft that pdftex.def goes to some lengths to patch it at end of package, seems I may as well just change trig rather than make luatex.def do same. (I already updated dvips.def to tell it dvips can handle compressed images without shell escape to save graphics.cfg patching the dvips rules)

@DavidCarlisle Does it?

@JosephWright apparently texlive's graphics.cfg does:

% Fix for dvips and xdvi versions that can uncompress
% graphic files without an explicite call of gunzip.
% (The fix is not applied for miniltx.tex, because
% \AtEndOfPackage is not available in plainTeX.)
\begingroup\expandafter\expandafter\expandafter\endgroup
\expandafter\ifx\csname AtEndOfPackage\endcsname\relax
\else
  \AtEndOfPackage{%
    \begingroup
    \@ifundefined{DeclareGraphicsRule}{%
      \endgroup
    }{%
      \def\x{dvips.def}%
      \expandafter\endgroup
      \ifx\x\Gin@driver

@DavidCarlisle Ah, that's handy to know

@JosephWright but all of the above is equivalent to me removing three instances of `gunzip -c in drivers.dtx

@DavidCarlisle Yes I see that but worth me knowing for the idea of writing some L3 image inclusion drivers

@JosephWright yes so I guess:-)

@JosephWright I realize you have a secret plan to make all my work obsolete!

@DavidCarlisle No, I have a non-secret plan to rip-off leverage your work into updated structures

@JosephWright food! back in a bit

@DavidCarlisle Issue with l3build is the part where we have to undo the different line wrapping in LuaTeX. I need to think again about the logic.

Happy New Year, everyone. Did I say that already? It's hard to keep track.

Anyone else want something starred?

Anybody likes the choice of the bibliographic references?

@egreg -- yes, like the choice, but i think you're being naughty. what's that new year's resolution?

@barbarabeeton I never make new year's resolutions, particularly if they involve @DavidCarlisle.

@JosephWright I did.

@StefanKottwitz :-)

@StefanKottwitz I'm not even doing the hat thing

@JosephWright who writes a summary of the spectacular event for blog and meta? :-)

Could people please help by voting to undelete the following answer?

tex.stackexchange.com/a/283645

@cfr I undeleted

It has been misidentified as something which should be a question edit but it is an answer. I've edited this to make it clearer and added a screenshot of the output, but I can only vote to undelete once. [This is not an author deletion and the author left a comment pointing out that they posted a solution but it got deleted.)

@StefanKottwitz Thank you!

@cfr Don't worry, @StefanKottwitz and I get infinite (un)delete votes

@JosephWright Thanks.

@PauloCereda I wish you a happy new year

and everyone else too ..

@UlrikeFischer A quackshop!

How can I set the style for Python docstrings in lstinputlisting?

I tried setting commentstyle but that doesn't work

@user1667423 perhaps post this question on the site

@DavidCarlisle I've worked out where the issue is in l3build. I need to rework some of the whitespace normalisation before fixing it, and that will mean regenerating a reasonable number of testfiles. I'll do that tomorrow.

@UlrikeFischer Happy New Year to you too, Ulrike! <3

@UlrikeFischer OOH I WANT THIS

@egreg This is the best store ever. :)

Happy new year again! :)

Although it is still 10PM. :)