6:33 AM
Door label template. What the hell?

6:53 AM
@Johannes_B I guess “a set of macros to ease input of data” for producing door labels.
@Johannes_B I find the term “template” quite appropriate in this case.

@egreg \tabular and \scshape and fontsize and the rest from the kernel and the classes?
@egreg I don't. What would a possible macro name be?
@egreg If this is needed a hundred times, the term template would be appropriate. But i don't see that from the question.

@DavidCarlisle One more log change question on the LuaTeX list: other than those three, I think everything else can be normalised or arises in cases that use a .luatex.tlg anyway (so have to be updated as the \discretionary changes aren't normalised in those)

@JosephWright ok thanks going to work now but will run test suite later see what happens here:-)

1 hour later…
8:07 AM
@DavidCarlisle I guess I could try to either normalise out \discretionary after a - line generally or 'add back in' for LuaTeX: not sure this is a 'good idea'

@JosephWright if you are down to a reasonable number of luatex.tlg I'd say stop, if the normalisation gets too inventive then likely to generate false results or be harder to debug the results later. It's a shame as you can I think reliably convert the luatex logging of a discretionary to tex format but I don't think you can reliably know when they will display unused discretionaries.

 Check failed with difference files
- ../build/test/tlb-callbacks-001.luatex.diff
- ../build/test/tlb0435.luatex.diff
- ../build/test/tlb0881.luatex.diff
- ../build/test/vlatex01.luatex.diff
- ../build/test/vlatex07.luatex.diff
- ../build/test/vtl2e01.luatex.diff
- ../build/test/vtl2e07.luatex.diff
@DavidCarlisle ^^^
@DavidCarlisle (Excluding changed but existing .luatex.tlg)

@JosephWright there was a luatex bug (last year?) about hyphens between digits, just failed to find it now.
@JosephWright call it done:-)

@DavidCarlisle Of the above, one is I think a new callback, one is the hyphenation business, one is the zero belowdisplayskip and the rest are the digit business
@DavidCarlisle I'll wait to see what I get back on the LuaTeX list

@JosephWright yes
@JosephWright I see you pre-answered Hans' predicted response:-)

8:14 AM
@DavidCarlisle Indeed: the one that worries me is the hyphenation change. The other two may just be logging.

0

Since you gave no clue of what format, font you wanted, here is something for you to start with: The \employee command scales the name to cover the whole text width. To go on adding more names just follow the lead. Font choosing is of your picking, there is the nice Latex Font Catalogue or you c...

@Johannes_B, don't be so harsh on the guy... He's only lazy to research, I gave him something to start with.

@DavidCarlisle There is at least one discussion around here tug.org/pipermail/luatex/2016-May/005927.html. Does my example in the thread later on now work as expected?
\parindent0pt \hsize=0.1cm \relax
12-34-56-78-12-34-56-78 \par

@UlrikeFischer thanks will test : @JosephWright Hans' option (3) in the mail Ulrike links to would also explain the hyphenation test failure if he's switched that (again)
@ulrike oh it's reversed! You said there that it breaks after the first and not the rest, now it only doesn't break after the first

8:32 AM
@DavidCarlisle Ah, could be

@JosephWright but the above result on Ulrike's test looks "unexpected" (do you get the same?)

@DavidCarlisle Indeed: another one to report I guess
@DavidCarlisle Yes, see the same

@JosephWright do you want to followup or should I ? (Ulrike gives the previous behaviour here tug.org/pipermail/luatex/2016-May/005928.html) not sure what 0.95 does?

@DavidCarlisle ;-))))))). With 0.95 I get the old behaviour

one step forward, two steps back....

8:42 AM
@DavidCarlisle Reported

@DavidCarlisle and it is not some hyphenmin side effect. 99999999-99999999-99999999-99999999 behaves similar (I also tried in context where I have 1.0.1)

9:11 AM
@DavidCarlisle More new primitives, I see: have to add to expl3 (and we probably need to take a position for the 2e kernel)

9:38 AM
@JosephWright do we have a test file that lists them all, like we do for the callbacks? so extra ones always show up?
@JosephWright ah just saw Hans' reply, should that be set in the format for compatability?

10:16 AM
@DavidCarlisle Yes
@DavidCarlisle Yes, probably

@JosephWright and plain?

@DavidCarlisle Hmm, can add to the .ini files I guess

@JosephWright that's what I was wondering
Although arguably in other cases (for both plain and latex) we've opted not to tweak the format for compatibility (eg name changes) and just document the incompatibility and make the test suite cope..

@DavidCarlisle Yes, not trivial to decide what is best here
@DavidCarlisle Also reminds me we should probably do something about latex.ltx and e-TeX (and remove all the TeX2, etc., stuff)
@DavidCarlisle This is though a setting that the user can decide to alter in their document: I guess the question is whether we alter the 'out of the box' value (we do after all for e.g. Unicode code point set up)

10:51 AM
@JosephWright but obviously not for fonts to match:-) apart from compatibility it's not obviously a bad thing that long words abutting an hbox hyphenate. I'd be tempted to leave this (like leaving first word of a para hyphenating)

11:10 AM
@GuilhermeZ.Santos but I get tikz points just for adding an edef not actual tikz code so it's OK
@GuilhermeZ.Santos I seemed to have slipped with my timings but getting closer:
Jan 2 '13 at 16:46, by David Carlisle
@FaheemMitha The new years resolution is to get a tikz gold badge by answering tikz question with tex primitives, picture mode, or if really pushed pstricks, doing quite well so far this year.

@DavidCarlisle Sounds reasonable to me: a case where one can argue a change is sensible (also fonts, but that's another story)

@DavidCarlisle Still far behind the gold badge, though.

@egreg That's because @DavidCarlisle, @UlrikeFischer and I are testing LuaTeX :)

@egreg there's this Italian bloke keeps stealing ticks that I deserve

Sep 5 at 15:04, by David Carlisle
Nov 10 '15 at 17:18, by egreg
@DavidCarlisle Excuses, excuses.

11:22 AM
@DavidCarlisle See latest L3 checkins for new LuaTeX primitives :)

@DavidCarlisle Now that I know this, I cannot believe I voted up for the airplane thing...
@egreg not that far no, 38 tikz questions to go and he'll have it

\$ svn diff -rPREV l3kernel/l3names.dtx
Index: l3kernel/l3names.dtx
===================================================================
--- l3kernel/l3names.dtx        (revision 6721)
+++ l3kernel/l3names.dtx        (working copy)
@@ -976,6 +976,8 @@
\__kernel_primitive:NN \Umathlimitbelowbgap         \utex_limitbelowbgap:D
\__kernel_primitive:NN \Umathlimitbelowkern         \utex_limitbelowkern:D
\__kernel_primitive:NN \Umathlimitbelowvgap         \utex_limitbelowvgap:D
+  \__kernel_primitive:NN \Umathnolimitsubfactor       \utex_nolimitsubfactor:D

11:48 AM
@GuilhermeZ.Santos I got it much earlier

@DavidCarlisle Indeed: aren't these the two for the 'missing' parameters in Unicode math fonts (from a TeX POV)

@egreg I sense you only posted this to rub in someone's face... feel free correct me if I'm wrong ;P

@GuilhermeZ.Santos Nope. :)

@JosephWright don't think so, at least not the ones that came up before (looking at the names) I'd guess it's to do with subscripts rather than atopwithdelims worth adding a

@DavidCarlisle me too. Imho compability doesn't work anyway (even if luatex stays put for a while ...). But some document which demonstrates how hyphenation behaves and can be altered would be usefull. Btw: there is a question on the context list "Changes in hyphenations with 2016-10-12 beta".

11:56 AM
@egreg proving my earlier assertion

@PauloCereda Ohh darn! I thought I was on to something... xD

@GuilhermeZ.Santos he meant no, you are not wrong

@DavidCarlisle I did?

@PauloCereda yes

@DavidCarlisle ooh

11:58 AM
Anyone singing nobel prize for literature songs?

@DavidCarlisle hm?

@PauloCereda see this year's laureate

@DavidCarlisle The Times They Are a-Changin'

@egreg my thoughts exactly

@DavidCarlisle I got the irony, the "Oh darn!" was supposed to give it away (Who says 'Oh darn'???)

12:02 PM
@GuilhermeZ.Santos hot diggity dog!

12:57 PM
@DavidCarlisle Jokerman

Anyone know which colormodel xcolor uses by default under pdflatex? I need to find some rgb colors that work when the printing office convert our images to CMYK (I cannot use CMYK as the colormodel as I suspect that Tikz still have a tendency to force RGB in some circumstances). If I explicitly give the rgb option to xcolor the colors changes (as seen in evince) compared to without the rgb option.

@daleif tikz has no preference AFAIK, but you shall stick to user-defined colors as the predefined colors need not to convert properly (but I may be wrong).

@daleif if it's like color it will just use whatever model is defined to use the colour in each case without conversion to a common format.

1:15 PM
@yo' at least earlier for some of the shadings, colors was forcefully converted into rgb even if you used CMYK, so and edge mighh end up in one color, and the shading would not match.
@DavidCarlisle then what is the default when we have not specified anything? In color proofs (test) I got bach all my blue was purple.

@daleif ah shading ... well, there it's possible, this would lie in the pgf layer...

@daleif you can't not specify anything, \definecolor takes a mandatory model argument
@GuilhermeZ.Santos I see you are impressed by my tikz knowledge

@DavidCarlisle then which colormodel is used when you ask for one of the predefined colors?
\documentclass[a4paper]{memoir}
\usepackage{xcolor}
\colorlet{A}{cyan}
\definecolor{B}{rgb}{0,1,1}
\begin{document}

\color{A}

\rule{2cm}{2cm}

\color{B}

\rule{2cm}{2cm}
\end{document}

@DavidCarlisle Indeed quite amazing, for someone drawing with tex primitives, you are quite the tikz wizard!

As far as I can see xcolor defines it in both cmyk, rgb etc
rgb cyan looks bad, cmyk cyan is good

1:22 PM
@daleif it depends what defined them, in the original color package they are different depending on whatever worked it ends:
\ifx\color@gray\@undefined
\ifx\color@rgb\@undefined
\else
\definecolor{black}{rgb}{0,0,0}
\definecolor{white}{rgb}{1,1,1}
\fi
\else
\definecolor{black}{gray}{0}
\definecolor{white}{gray}{1}
\fi
\ifx\color@rgb\@undefined\else
\definecolor{red}{rgb}{1,0,0}
\definecolor{green}{rgb}{0,1,0}
\definecolor{blue}{rgb}{0,0,1}
\fi
\ifx\color@cmyk\@undefined\else
\definecolor{cyan}{cmyk}{1,0,0,0}
\definecolor{magenta}{cmyk}{0,1,0,0}
\definecolor{yellow}{cmyk}{0,0,1,0}
\fi
@GuilhermeZ.Santos it takes years of reading the tikz manual carefully to recognise an .ist extension.

@DavidCarlisle in xcolor:
\definecolorset{cmyk/rgb/hsb/gray}{}{}%
{cyan,1,0,0,0/0,1,1/.5,1,1/.7;%
magenta,0,1,0,0/1,0,1/.83333,1,1/.41;%
yellow,0,0,1,0/1,1,0/.16667,1,1/.89;%
olive,0,0,1,.5/.5,.5,0/.16667,1,.5/.39}

@DavidCarlisle Very much carefully, I believe if not so careful one might see himself at manual's very last page and still unaware of such file extension...

@daleif do you load xcolor manually before any other graphics package? Maybe even using \RequirePackage before \documentclass?

@yo' should that make any difference? In the sample doc I'm preparing there is no gfx so no such packages.

@daleif ok wasn't sure. It shouldn't make a difference, but you know, it's a LaTeX package so ...

1:31 PM
@yo' yes. This color model stuff is a bunch of trouble. At the moment there are some of our images they cannot process and so far they haven't told us exactly which, just pointed to some pages (which contains more than one image).
As for the colors, I'll just prepare a bunch of color series and have them convert and print it in their end, then we'll choose the colors that match our vision the best

@daleif that's a common thing (remember I typeset a journal, working with external source images all the time. It knows to be painful, and ImageMagick is my friend.)

@yo' deoes ImageMagick do cmyk? because gimp does not

@daleif yep, it's called colorspace, just google "ImageMagick rgb to cmyk" this shall be enough I gotta go now
but remember that it's always raster, never vector

1:48 PM
@DavidCarlisle, tex.stackexchange.com/a/333974/114143 much better than your answer... ^.^

@GuilhermeZ.Santos references the manual, which is cheating
5

Mar 26 '12 at 19:37, by David Carlisle
@Canageek moral of the story: never read the documentation, bad things happen

You guys are unbelieveble... hahaha

4 hours later…
6:18 PM
@DavidCarlisle What did you conclude from Hans' mail about hyphenation around numbers/-?

@JosephWright I'm not sure he addressed the issue of why it doesn't pick that one up

@DavidCarlisle :)
@DavidCarlisle That's the usual answer from Hans!

@JosephWright same rules as apply here: don't put two issues in one mail:-)

6:37 PM
@DavidCarlisle Hmm

2 hours later…
8:43 PM
 _______
< David >
-------
\   ^__^
\  (oo)\_______
(__)\       )\/\
||----w |
||     ||

@egreg that was first thought but I don't have cowsay on here:-)

 ___________________________________
/ David is using a broken operating \
\ system, if we can call it an OS   /
-----------------------------------
\   ^__^
\  (oo)\_______
(__)\       )\/\
||----w |
||     ||
@DavidCarlisle ^^^^

8:59 PM
 ___________________________
< dont be rude dont be rude >
---------------------------
\                             .       .
\                           / .   .' "
\                  .---.  <    > <    >  .---.
\                 |    \  \ - ~ ~ - /  /    |
_____          ..-~             ~-..-~
|     |   \~~~\.'                    ./~~~/
---------   \__/                        \__/
.'  O    \     /               /       \  "
(_____,    ._.'               |         }  \/~~~/
@egreg an animal from your youth

@DavidCarlisle I don't remember the hat

@egreg sign of old age, being forgetful

@DavidCarlisle It used to wear this one
I met it in Sicily