@JosephWright advance notice of a question for you to answer if it gets asked:-)

So it looks like today's target badge is something El Capitan related

By the way, I have 2¹¹ bronze badges

@egreg All undeserved of course :-P

@egreg 211 would be more than you deserve!

@DavidCarlisle Always so kind!

@egreg I do my best!

I've found a nice "feature" of tabularx :-)

@yo' it only has nice features

@DavidCarlisle yeah, like not supporting too narrow columns (or too many columns, who knows...)

@yo' It has lots of documentation about how to control narrow columns, not sure what you mean by "not support" ?

@DavidCarlisle well, I didn't read the documentation of course :)

@yo' It has lots of code to support narrow columns, you could read that instead (more reliable anyway)

@DavidCarlisle ok, I solved the problem another way, but I will anyway

strange, I can't reproduce the problem in a MWE...

This gives zero-width columns:

\documentclass[a4paper,10pt]{article}

\usepackage[margin=0.5in]{geometry}

\usepackage{tabularx}

\begin{document}

\tabcolsep0pt

\noindent
\begin{tabularx}{\linewidth}{|l|*{31}{@{}X@{}|}}
TEST &&&&&&&&&&&&&&&&&&&&&&&&&&&&\leavevmode\rlap{\the\hsize}
\end{tabularx}

\end{document}

Changing 31 to 30 gives \hsize of 17.37259pt. No warning from tabularx itself, just an underfull \hbox.

as posted it gives:

Package tabularx Warning: Target width: \linewidth  = 525.23788pt..

(tabularx)     Table Width    Column Width    X Columns
(tabularx)     14745.4662pt     525.23788pt      32
(tabularx) Don't exceed \maxdimen
(tabularx) Reached minimum width, backing up.

Underfull \hbox (badness 10000) in alignment at lines 14--14
[][][][][][][][][][][][][][][][][][][][][][][][][][][][][][][][]

changing to 30 it gives:

Package tabularx Warning: Target width: \linewidth  = 525.23788pt..

(tabularx)     Table Width    Column Width    X Columns
(tabularx)     14745.4662pt     525.23788pt      31
(tabularx)      1473.25343pt     51.23029pt      30
(tabularx)       557.92824pt     18.5401pt       29
(tabularx)       525.23793pt     17.37259pt      28
(tabularx) Reached target.

@yo' ^^ what do you expect it to do?

Where?

*geometry* driver: auto-detecting
*geometry* detected driver: pdftex
[1{/usr/local/texlive/2015/texmf-var/fonts/map/pdftex/updmap/pdftex.map}]
(./tmp.fXtkdb08Wi.aux) )</usr/local/texlive/2015/texmf-dist/fonts/type1/public/
amsfonts/cm/cmr10.pfb>
Output written on tmp.fXtkdb08Wi.pdf (1 page, 12585 bytes).
SyncTeX written on tmp.fXtkdb08Wi.synctex.gz.
Transcript written on tmp.fXtkdb08Wi.log.

@yo' \tracingtabularx

@DavidCarlisle aha!

@yo' It takes a lot more care lining up the columns in that tracing output than it does making the columns in the output.

I still probably don't understand it, but OK. I don't think I exceed maxdimen, but fine, I can live with that :)

@yo' tx doesn't let a column go below 1em, and then you have tabcolsep either side, so if you have 31 columns you are getting wide even before you start adding data

@DavidCarlisle with \tabcolsep0pt? Or with @{}X@{}?

the column with is 17pt for 30 columns, so it would be no less than 16pt for 31 columns! (Actually, replacing X with p{16pt} works.)

@yo' well you get tabcolsep still, but it's 0 (in the first case) you still get at least p{1em} though

@DavidCarlisle well, I get p{0pt}

@yo' so you do > 0.0pt. sorry I believed the documentation.

@DavidCarlisle yep, there's enough space. The problem is that at the beginning of the calculation, there is 31\linewidth, which is more than \maxdimen.

@yo' ah the clue is (tabularx) Don't exceed \maxdimen which is a new extra check added in 2012, before that you would have had a low level tex arithmetic error. It avoids the error and just forces things to fit but does mean that some of the constraints (like not going less than 1em) seem to be dropped in that case)

@DavidCarlisle yep, that's the point it seems :) (I still don't understand, do you need to start with assuming each column is as wide as the table? ...)

@yo' that's the algorithm it uses, start by assuming X is some width set the table, see how wide it is and then do some arithmetic to work out the next guess of X width. The guesses only ever get smaller so it needs a reasonably big guess to start with....

@DavidCarlisle ah ok, and could it start with table width / #X ?

@yo' possibly, possibly not. tricky cases to make sure the algorithm converges are all to do with \multicolumn if you specify three X columns but two of them are always hidden by \multicolum{..}{c} entries then there is effectively only one X and that has to be wide enough to pad out the table and you don't know how wide that is (or if there are any \multicolumn) until you try....

@JosephWright ltluatex.tex (not sty) unless you changed it at last minute?

@DavidCarlisle Don't actually remember: we were messing about with different versions

@JosephWright yep but seems like we ended with:

format.ins:          \from{ltluatex.dtx}{2ekernel}
format.ins:\generate{\file{ltluatex.tex}{\from{ltluatex.dtx}{tex,plain}}}
format.ins:\generate{\file{ltluatex.lua}{\from{ltluatex.dtx}{lua}}}

@DavidCarlisle Send a correction then :-)

@JosephWright OK I was just giving you a chance, but I'll do it:-)

@DavidCarlisle ah ok, I see.

@egreg that's the one:-)

Quack!

El Capitan on queue right now. :)

@PauloCereda Can you run TeX programs?

@egreg I will find out later. :)

Good MAEN.

@ChristianHupfer: ^^ :P

@PauloCereda GEMA will sue you for stealing my line :-P

:)

@PauloCereda :D

@PauloCereda: By the way, El Penguin isn't online very often recently ??

@ChristianHupfer No. :( I miss @Johannes. :(

@PauloCereda Perhaps you should make a github repository full of Liver recipes ... he will rush in and be an avid follower ;-)

@DavidCarlisle No wailing yet

@JosephWright no one has it yet:-)

@DavidCarlisle Nothing to stop them doing svn co http://latex-project.org/svnroot/latex2e-public/trunk && cd trunk && texlua build.lua install && fmtutil --all ;-)

@DavidCarlisle @UlrikeFischer has ;-)

@DavidCarlisle More seriously, I think the luatexbase part will be fine, it's the \luatex... bit that will doubtless give us headaches

@JosephWright yep

@DavidCarlisle I also think it's the right approach long-term :-)

Interesting thing with the Travis set up, I get to see every file change in a basic TL install

@PauloCereda Oh <3

@PauloCereda @ChristianHupfer A bit busy helping out on nother places for the last few weeks.

@Johannes_B Sorry, can't follow: Filling my "Versetzungsantrag" :D

@Johannes_B yaayyy

@Johannes_B: Versetzungsantrag erledigt und abgeschickt... jetzt Daumen drücken ...

If @DavidCarlisle were German. ^^

@PauloCereda Sind Sie sicher, ich bin nicht auf Deutsch?

@DavidCarlisle oh nein! :)

@DavidCarlisle "Sind Sie sicher, daß ich kein Deutscher bin?";-) :D

@ChristianHupfer Moving to another town?

@PauloCereda: @DavidCarlisle is still looking for the Deutschen Klassenzimmer :D

@ChristianHupfer @DavidCarlisle Ich dachte wir dutzen uns.

@Johannes_B Nope, staying here in my home 'village', just another school, no longer Offenburg

@ChristianHupfer What kind of school is it now?

@PauloCereda Wir Deutschen verstehen, die alles über Cricket , nicht in der mag diese portugiesischen Fußball Kerl Sie wurden gestern im Gespräch .

@DavidCarlisle :D :D :D :D Oh my god.

@Johannes_B Technisches Gymnasium... I want to return to my previous work place

@DavidCarlisle oooh

@ChristianHupfer :P

@ChristianHupfer Oh, cross fingers then.

@Johannes_B Thanks! I need it !

@ChristianHupfer If you need any help with the German language, just let me know.

@DavidCarlisle I let you know ... :D In the meantime I teach Portuguese to @PauloCereda, Danish to @daleif, Norwegian to @TorbjørnT., Italian to @egreg and Welsh to @cfr ;-)

@ChristianHupfer Penso di essere un buono studente, che ne dici?

@egreg: Certo che si!

@PauloCereda (or anyone) have you used github pages gh-pages branch thing?

@ChristianHupfer :)

@DavidCarlisle I did!

@ChristianHupfer Sono italiano de la foresta nera. :)

@egreg @DavidCarlisle Is there any way to extract the information if a specific counter is in the reset list of another counter: I tried \IfSubStr{\csname cl@chapter\endcsname}{{section}}{...}{...} which works only with the explicit {section} but not with section alone (\IfSubStr is from xstring)

@ChristianHupfer \IfSubStr doesn't look inside groups, by default.

@PauloCereda “della”, not “de la”

@egreg oh sorry. :)

@PauloCereda Italian is difficult!

@egreg <3

 Conference programme

9:00 Meeting opening

9:15 I calcoli matematici in pdftex. Claudio Beccari

9:45 Il pacchetto ECG. Ezio Aimé e Marco Scavino

10:15 Canzonieri online con LaTeX. Francesco Endrici e Matteo Conci

10:45 Coffe break (30 min)

11:15 Creare un database bibliografico da un file di testo. Matteo Fadini

11:45 Zbl-build: a GUI interface for Biblatex. Guido Milanese

12:15 MlBibTeX 1.4: the New Version. Jean-Michel Hufflen

12:45 Lunch break (1 h 45 min

14:30 Generare documenti LaTeX con diversi linguaggi di programmazione. Roberto Giacomelli e Gianluca Pignalberi

^^ yaaaaaaaaaaaaaaaaaaaay

@egreg Ok, I will try that

@ChristianHupfer On the other hand, l3regex is agnostic with respect to groups.

@egreg: I will come back to it later on ... now cooking (Indian menu) @Johannes_B ;-)

@ChristianHupfer VVVV

\documentclass{book}
\usepackage{xparse,l3regex,chngcntr}

\ExplSyntaxOn
\NewDocumentCommand{\CounterIsInResetListTF}{mmmm}
 {
  \regex_match:nvTF { \b #1 \b } { cl@#2 } { #3 } { #4 }
 }
\cs_generate_variant:Nn \regex_match:nnTF { nv }
\ExplSyntaxOff

\begin{document}

\CounterIsInResetListTF{section}{chapter}{IS}{IS NOT}

\CounterIsInResetListTF{subsection}{chapter}{IS}{IS NOT}

\counterwithin{subsection}{chapter}
\CounterIsInResetListTF{subsection}{chapter}{IS}{IS NOT}

\counterwithout{section}{chapter}

@ChristianHupfer Note that thanks to \b, the macro is not confused by the fact that subsection is still in the reset list for chapter.

@ChristianHupfer With \IfSubStr you can't do that, AFAIK. With it I'd look for the braced name.

@ChristianHupfer no need for ifsubst, just execute the reset list after having defined \@elt that's the point of those lists they are designed to be executed not searched as data

\documentclass{report}

\begin{document}

\makeatletter

\def\zzz{section}
\def\@elt#1{\def\zz{#1}\ifx\zz\zzz\@tempswatrue\fi}

\@tempswafalse\cl@chapter\if@tempswa section is in chapter \else section is not in chapter\fi

\@tempswafalse\cl@footnote\if@tempswa section is in footnote \else section is not in footnote\fi

\end{document}

@ChristianHupfer (@egreg) ^^^^

@ChristianHupfer You can have home-made food delivered to work.

@DavidCarlisle Too simple. :P

@egreg let's compare \tracingall logs with that against the l3regex version:-)

@Johannes_B: Where's the fun in delivered food? Cooking is more fun

@egreg Thanks. But a I am not that deep into (La)TeX core: What does \b do???

@DavidCarlisle Ok, it looks more hackish than \ExplSyntaxOn ;-)

@ChristianHupfer Did you follow the link? Your fiancee would cook the meal for you.

@Johannes_B: Sorry, no I did not follow the link. And believe, it's better if I cook instead of my fiancée ;-)

@ChristianHupfer that's not latex core that's regex (expl3 version)

@ChristianHupfer Oh, not a good chef?

@Johannes_B Hmm... hüllen wir den Mantel des Schweigens darüber ;-)

@DavidCarlisle I should look into regex then. Thanks!

@ChristianHupfer not a hack at all in this case it is why the lists are the way they are. DEK has a section about the technique in the texbook comparing a list stored as say {section}{footnote}{theorem} which you would need to parse/search compared to one stored as \do{section}\do{footnote}\do{theorem} where you just define \do then execute the list. LaTeX uses this technique all over the place.

@ChristianHupfer not for this, no you shouldn't!

@ChristianHupfer all the latex reset mechanism works by defining \@elt locally then executing the list, see the basic stepcounter than needs to reset all the counters in the list:

\def\stepcounter#1{%
  \addtocounter{#1}\@ne
  \begingroup
    \let\@elt\@stpelt
    \csname cl@#1\endcsname

locally define \@elt to reset a counter, then execute cl@#1 you don't need to do any kind of "for-each" over the list as the list can be simply executed

@DavidCarlisle: I've understood that but I've to break if the relevant counter name is found (i.e. if it's in the list)

@ChristianHupfer why?

@ChristianHupfer no need to break, the list is in no particular order so you can skip over the ones at the end same as the ones at the start, see above with the "is section in the list test": \@elt does nothing in all cases but one, so you can execute the full list

@DavidCarlisle The \@elt list is executed to pursue all counters (in \stepcounter) That's fine there, but I want to know if section is reset by chapter, so I have to stop this execution.

@ChristianHupfer no, you don't see the above document which reports that section is in the reset list for chapter and not in the reset list for footnote

@DavidCarlisle I've to dig into this...

@ChristianHupfer You've finished teaching your grandmother to suck eggs, I guess ;).

@cfr Hi from the hwyaden! :)

@ChristianHupfer It's a standard regex operator, it denotes a word boundary.

@PauloCereda Listening to symphony n. 5. No, not Dvořák's. ;-)

@egreg Oh I got scared for a moment. :)

@PauloCereda Bruckner's, of course. :P

@egreg :)

@egreg: You are mentioned in my GuIT talk. :)

@PauloCereda Like how I mentioned @DavidCarlisle in mine at TUG?

@egreg ^^ in a good way. :)

@PauloCereda :) With Ale Del Piero photo?

@egreg Oh my, I forgot about him!

Anyone got a Linux set up with TL2015 available?

@JosephWright Yes: cardiff. :)

@PauloCereda Could you run

%
% Copyright (C) 2015 Liam Huang
%

\documentclass{article}
\input{regression-test}
\AUTHOR{Liam Huang}

\usepackage{expl3}

\START

\OMIT
\usepackage[UTF8, fontset = fandol]{ctex}
\TIMO

% Skip the font loading stuff
\OMIT
\begin{document}
\TIMO

\showoutput

我爱 Anyi。

\end{document}

with LuaLaTeX and let me know if you get any missing font warnings?

@JosephWright Sure, let me grab my laptop. :)

@egreg 211 is far more rep than you deserve

Trying to set up a new github repository (or at least access one that was set up for me....)

@PauloCereda Cwac ;) I assume 'hi' to be a translation from the original Hwyadeg?

@cfr ooh I have no idea. :)

@DavidCarlisle I'm having fun with pull requests

dspdfviewer.danny-edel.de

@JosephWright gimme a minute, updating my TeX distro.

@JosephWright I saw you got one:-) On the above in cygwin I get no missing font warnings but it ends with

Missing character: There is no 我 (U+6211) in font "file:lmroman10-regular:scri
pt=latn;+trep;+tlig;"!
Missing character: There is no 爱 (U+7231) in font "file:lmroman10-regular:scri
pt=latn;+trep;+tlig;"!
{/usr/local/texlive/2015/texmf-var/fonts/map/pdftex/updmap/pdftex.map}
Missing character: There is no 。 (U+3002) in font "file:lmroman10-regular:scri
pt=latn;+trep;+tlig;"!
(./aaaa.aux)
END-TEST-LOG

@JosephWright github.com/w3c/xml-entities

@DavidCarlisle Envy? :P

@DavidCarlisle Full TL install?

But seems like I can't push to it yet:-)

@DavidCarlisle Yes, that's the warning I mean: missing glyphs in the font (you'll have one earlier telling you the font has no glyphs in it, which it clearly does)

@JosephWright running it right now.

@JosephWright but why is it using lmroman not the fandol thing?

@DavidCarlisle Ah, I'd missed that. I get

*************************************************
* ctex warning: "fntef-not-available"
*
* Functions of `CJKfntef' is not available in LuaLaTeX.
*************************************************

Missing character: There is no 我 (U+6211) in font "[FandolSong-Regular.otf]:mo
de=node;script=latn;language=DFLT;jfm=quanjiao"!
Missing character: There is no 爱 (U+7231) in font "[FandolSong-Regular.otf]:mo
de=node;script=latn;language=DFLT;jfm=quanjiao"!
{/tmp/texlive/texmf-var/fonts/map/pdftex/updmap/pdftex.map}
Missing character: There is no 。 (U+3002) in font "[FandolSong-Regular.otf]:mo
de=node;script=latn;language=DFLT;jfm=quanjiao"!
(./CTeX001.aux)
END-TEST-LOG

on a VM equivalent to the Travis rig

@PauloCereda I got that but ignored it as I didn't know what it meant

@DavidCarlisle Neither do I.

@DavidCarlisle I see that too, but that's not important!

What's up is that on my native set up (Mac) I do get the fonts

I wasn't sure if it's due to the VM being a cut-down install: obviously not

@JosephWright would you like my log?

@PauloCereda Please

Hello, has anyone tried Texlipse?

@JosephWright fpaste.org/273840/14437336

@user2692669 I used it once in order to write an answer. It was the most annoying thing of my whole life. :)

Ouch

I want to write my thesis but I need something efficient.

@DavidCarlisle Once again looks like we've found a LuaTeX (related) bug

@user2692669 What's wrong with TeXworks ;-)

@PauloCereda tell @user2692669 to use emacs

@DavidCarlisle Now you are just having fun :P

@DavidCarlisle Could you send me your log too?

@user2692669 @David and I having fun on each other's expense because of emacs and vim? No way. :)

@user2692669: tell us more about your contact with TeX and friends, we might help. :)

@PauloCereda Looks like your one is OK: what system is that?

@JosephWright Fedora 22 64bits, last TL update.

@PauloCereda EXPLICIT CONTENT AHEAD: I don't have any solid experience, I wrote some reports 2 years ago but that was it (mostly). My guilty pleasure is that afterwards I discoverd Garamond font (all of it's variations) and rolled back to MS word (guilty as charged).

@DavidCarlisle, @PauloCereda So three Linux(-like) systems, three different outcomes, apparently with the same files!

I spent about 15 mins working out that ctex does a conditional load of everyhook, but that's nothing to do with it

@DavidCarlisle Could you check what you get with the 2015/01/01 format?

@JosephWright hmm no:-) paths seem messed up will sort out but I am a bit distracted at present

@DavidCarlisle Looks like something is up with the updated format

@JosephWright yes I did wonder, I'll try to build a 2015/01/01 clean format in a bit..

@DavidCarlisle Patch level 2048?

@JosephWright hmm 2015/01/01 works...

@JosephWright hmmm we need to inspect the paths.

@user2692669 dickimaw-books.com/latex/thesis

^^ Have fun. :)

@PauloCereda I suspect the custom whatsit allocator

@DavidCarlisle ooh a whatsit!

@PauloCereda 146 pages? You are killing me. Every day is precious...

@PauloCereda One minor detail: I must use xelatex or anything that supports Greek (unicode is the best fit i guess) (also I'm undergraduate)

@user2692669 depends what you mean by support: you can typeset Greek with pdftex, xetex or luatex, details differ though depending which you use

@DavidCarlisle 90% or more

@user2692669 90% of what?

Sorry I misread the semantics (I meant 90% use of greek)

@user2692669 no idea what that means:-)

@DavidCarlisle Just: the text will be in Greek.

@user2692669 yes so you can use any of the engines

@DavidCarlisle All support unicode?

@user2692669 enough to do greek, if you are doing right to left chinese in an otherwise arabic document with Hebrew quotations, then xetex might be better at unicode than pdftex but Greek is OK for pdftex (mostly:-)

@DavidCarlisle so I should prefer pdftex as the best solution, is that right?

@user2692669 not necessarily, as I said, it depends. Not least on which fonts you want to use, if you have some requirements to use specific opentype fonts then pdftex can't use opentype, so luatex or xetex is your choice, otherwise it depends on what packages you need and how well they work with the various engines. but basically just start writing, you should be able to make it work with any of the systems with just a few changes to the preamble

@DavidCarlisle Wise words, thank you.