« first day (2658 days earlier)      last day (2290 days later) » 

6:21 AM
@ShreevatsaR Wow!
@egreg Wow, wow!
 
6:45 AM
```
\newcommand*{\eulerfont}{\fontfamily{ccr}\selectfont}
\newcommand*{\fnn}[1]{\ensuremath{\eulerfont #1}}
```
this doesn't change the font when i do something like $\fnn{x}$ -- any ideas?
 
7:39 AM
@user89 they are text font settings, math is more complicated as there is not a single "current font" each character is (by default) assigned to one of 16 (or 256 in xetex/luatex) math groups (fonts)
you want to define \fnn by \DeclareMathAlphabet\fnn{OT1}{crr}{m}{n} or some such
 
8:01 AM
@WillRobertson vvvvv
\documentclass{article}

\ifx\Umathchardef\undefined\else\usepackage{unicode-math}\fi

\ifx\directlua\undefined\else
\ExplSyntaxOn\makeatletter
\cs_set:Nn \__um_setmathfont:nn
  {

    \__um_init:n {#2}
    \cs_set_eq:NN \glb@currsize \scan_stop:
    \cs_if_exist:cF { S@ \f@size } { \calculate@math@sizes }
    \use:c { S@ \f@size }

    \keys_set_known:nnN {unicode-math} {#1} \l__um_unknown_keys_clist

    \bool_if:NT \l__um_init_bool
      {
        \__um_fontspec_trial_font:
        \__um_declare_math_sizes:
@HaraldHanche-Olsen ^^^^^^^^^
@WillRobertson @HaraldHanche-Olsen not sure I understand fully what's going on there but tracing the differences between xetex (which was OK) and luatex (which added a space) leads to the lines marked %HERE and if they are removed the extra group goes and the aftergroup does not trigger the global restore in the wrong place. But I don't know what the commands do or what other effects commenting them out will have.....
 
@DavidCarlisle $$\relax$$ is the most relaxing math display I ever didn't see.
 
@HaraldHanche-Olsen I copied it from @PauloCereda's thesis
 
@DavidCarlisle Another thing I never saw!
 
@HaraldHanche-Olsen you have seen it all though
 
@DavidCarlisle Now I've seen everything.
 
8:07 AM
@HaraldHanche-Olsen I am not convinced that that is a completely valid conclusion, but then I don't know the logical framework that you are working under
 
@DavidCarlisle You expect logic, of all things, in this chat?
3
@DavidCarlisle I didn't know that LaTeX contains ExplSyntax without loading a package. Since when did that happen? I don't see it in latex.ltx.
 
8:35 AM
@DavidCarlisle oi
HAPPY PIZZA DAY
2
 
Uh, silly me – it's the unicode-math package that pulls in expl3, of course. 🤪
 
@HaraldHanche-Olsen The chat is in fact very logic, in terms of a duck logical algebra ... or pizza logics ;-)
@PauloCereda: Duck logical algebra is not be confused with duck counting strategies: zero, one, many ;-)
 
@ChristianHupfer I think you just proved my point.
@ChristianHupfer Or counting the pages in a thesis: Zero, zero, zero, …
 
@HaraldHanche-Olsen Yes, it is a weird logic
@HaraldHanche-Olsen :D:D... or is that not meaning that there pages that are just numbered with zero?
 
@ChristianHupfer it's all very simple. However don't forget the pure duck calculus is untyped. :)
@HaraldHanche-Olsen OI
 
8:45 AM
@HaraldHanche-Olsen it is loading a package (unicode-math)
 
@DavidCarlisle Yes, I already figured it out (see above). It happened in a sudden attack of logic. Not likely to be repeated anytime soon.
 
@HaraldHanche-Olsen from Scandinavia I expect constructive type theory to be honest:-)
 
@DavidCarlisle ooh
 
\documentclass{empty}

\author{Some Duck}

\title{Counting strategies of ducks}

\begin{document}
% The thesis content -- finished.
\end{document}
2
@PauloCereda: Citing Monty Python: Just in order to annoy you ;-)
 
@ChristianHupfer you naughty German
 
9:39 AM
[4/7, 00:07/00:14] auto-install: typewriter (46574) [88k] ... done
 
@DavidCarlisle ooh
 
10:00 AM
@PauloCereda I think the most informative and well typeset package documentation ever.
 
@DavidCarlisle :)
 
10:22 AM
@DavidCarlisle How could we do without it?
 
10:41 AM
@egreg ^^
 
@PauloCereda Should I file a lawsuit?
 
@egreg oh no
 
10:57 AM
@egreg you may get a nice little side career on the strength of that recommendation
 
11:08 AM
@DavidCarlisle logging of lua-files looks messy ;-). luaotfload mentions the path to the fontloader and luaotfload-names.luc in the log, the .fls-files shows the lualibs-lua-files, luaotfload-characters.lua and the .luc-files but e.g. fontspec.lua is missing.
 
@UlrikeFischer does the --recorder thing log lua (never tried)
 
@DavidCarlisle Only in part (see my comment about "fls-files").
 
@UlrikeFischer ah Ok, not sure I have ever actually used recorder:-)
 
@DavidCarlisle only trombones?
@WillRobertson HAPPY PIZZA DAY
 
@WillRobertson I should probably have put that as a github issue (could still if that would be useful) but it is s stripped down version of an example @HaraldHanche-Olsen posted here
23 hours ago, by David Carlisle
\documentclass{article}


\begin{document}

We study \ldots
\tracingcommands=2 \tracingmacros=2
$$\partial_tu-A\varphi(u)=0.\aftergroup\begingroup\aftergroup\endgroup$$
These equations \newline bla bla
\end{document}
 
11:23 AM
:4277029 You are very brave trying to run \tracingall there. The code in that which you commented out that causes the problem is \check@mathfonts. Do you suspect that's incorrect, and I need to dig into this code again?

The short explanation of what is happening here is that LuaTeX uses the LAST maths font that is loaded to initialise all the (equivalent of what it uses instead of) fontdimens. But in unicode-math I need it to be the FIRST maths font.
To be honest I've forgotten the exact reason why \check@mathfonts is there. I think I borrowed that code from Ulrike so maybe she knows...
 
@WillRobertson always have a person to blame in case of emergency:-)
2
 
Dec 2 '17 at 22:30, by Joseph Wright
@DavidCarlisle Blame Joseph?
 
@DavidCarlisle The internals of NFSS maths font loading is one of those things I only understand when I'm in the middle of writing the code around it, and then next time I look at the code I just have to trust what I had done...
 
@WillRobertson Not got a lot of time this morning perhaps I'll put it in github and point Frank at it, as the example above shows if you aftergroup a \begingroup\endgroup out of a math display then you lose the implicit \unskip so that gives space even in pdftex, but I failed yesterday to get pdftex or xetex to trigger the \aftergroup\glb@settings in the wrong place
@WillRobertson although it looks like that could happen (which would then arguably be a latex format issue) I only actually got it to happen with unicode-math luatex
 
11:40 AM
Holy cow, arara is insanely popular amongst German TeX users!
 
It looks to me like one shouldn't be calling \check@mathfonts in a general context; all the cases in the kernel have it being executed inside a box or math mode or similar. If I replace it by

\ifx \glb@currsize \f@size \else
\glb@settings
\init@restore@glb@settings
\fi

It appears to continue to do the job it's supposed to be doing, without the side-effect.
 
@WillRobertson er I trust you (and we can blame @HaraldHanche-Olsen if something breaks)
 
@DavidCarlisle Good plan!
 
@WillRobertson I have no idea which code you mean but there is a high probability that I used \check@mathfonts only as a ad-hoc way to force rechecking of the math in some example.
 
@UlrikeFischer And now I wonder whether that snippet above can be replaced entirely by just \glb@settings... it claims on the box that it only reloads maths fonts when necessary, so the currsize check would seem to be redundant...
 
11:55 AM
@DavidCarlisle The reason I asked is that this question: tex.stackexchange.com/questions/414102/…. It boils down to the problem that in context require 'pgf.gd.control' doesn't load pgf/gd/control.lua/ but pgf/gd/force/Control.lua` (remark the additional "force" in the path) and I was wondering if this could happen in latex too. (I have to go now and won't be back for a few hours).
@WillRobertson I think you should ask Frank ;-)
 
@UlrikeFischer Eyeballing the code it seems safe! Famous last words when it comes to the NFSS. I need to get to bed I'm afraid...
 
@WillRobertson what could possibly go wrong? :)
 
@PauloCereda your thesis could fail to typeset
 
@DavidCarlisle oh no
 
This is all around setting and resetting fontdimens (or rather math params) in LuaTeX :)

I haven't forgotten the request to add some interface here... I just for the life of me can't get the time to concentrate hard enough to figure out how to do that nicely. (The code isn't the hard part, the interface is. Especially since XeTeX is involved too.)
 
12:17 PM
@DavidCarlisle: I could go to a Decathlon store and buy a cricket kit for us to play in TUG 2018
 
@PauloCereda or you could work on your th.....
 
@DavidCarlisle oh no
 
 
2 hours later…
2:09 PM
Is there any way, when running luatex in TeX mode, to drop into lua interaction mode, as if one were running texlua instead? It seems it could be useful for debugging purposes.
 
2:19 PM
Hi mr. @marmot! How was the talk?
 
@PauloCereda Thanks for asking! It will be in almost 6 hours.
 
@marmot ooh
 
@ChristianHupfer -- oh, i think it's "zero, one, two, several, many". a much richer system, really.
 
@PauloCereda Since I don't have the TeX code for the nice pizza day animation, this might be the highlight:
user image
3
 
@marmot yay!
@marmot @UlrikeFischer: ^^ our resident marmot would like the pizza code. :)
 
3:09 PM
@TeXnician: would you poke me by mail? :)
 
@PauloCereda Which mail address should I use?
 
@TeXnician good question: try cereda at users dot sf dot net
 
@PauloCereda You should have mail.
@PauloCereda Btw: The python code on your home page is a very interesting way to distribute that address ;)
 
3:24 PM
@TeXnician oh no :)
@TeXnician oh no, nothing arrived so far. I blame SF!
@TeXnician Perhaps trying cereda dot paulo at gmail dot com might be better...
 
@PauloCereda You may also blame my mail provider which may have some difficulties.
 
The first should be an alias to that...
@TeXnician we could blame everybody. :)
 
@PauloCereda Oh no, gmail. They always mark my mails as junk.
 
@TeXnician Ah hold on. Let me check.
@TeXnician No you are not in spam. :)
 
@PauloCereda Well, maybe it has some delay in delivery. I hope it will come through within the next minutes.
 
3:30 PM
@TeXnician Let us wait. Want me to poke you instead? :)
 
@PauloCereda Well, I would prefer to wait as I currently do not communicate my e-mail address in public (due to previous junk).
 
@TeXnician Ah I completely understand. :) Sorry for the trouble, but I am sure you will enjoy some news. :)
 
@PauloCereda Then I will wait a bit longer.
 
3:46 PM
@TeXnician: It arrived!
 
@PauloCereda Good to know.
 
@TeXnician Replied. :) Now let us see if it arrives. :)
 
@PauloCereda It arrived and I just replied. On another note: Which technology do you plan to use for ArTeXmis? Java with Swing?
 
@TeXnician Yes, or perhaps Groovy. I am still deciding. I thought of JavaFX, but I want better coverage.
 
@PauloCereda I would love to see JavaFX (maybe even Kotlin which makes life easier). Then I'd be able to contribute some more.
 
3:54 PM
@TeXnician So far I only managed to run JavaFX with Oracle, and that's definitely a downside for me. I really enjoy OpenJDK. If I manage to make things work with opensource alternatives, I might look into it.
@ShreevatsaR: Hi mr. velociraptor, happy pizza day!
 
@PauloCereda the same to you
 
@PauloCereda I'm running OpenJDK with JavaFX here (on Arch Linux). Unfortunately it is still version 8 (so old ;) ). But I've had problems with it before, so completely understand you.
 
@TeXnician Thanks for the comprehension. :) I really want to cover as much scenarios as possible. You probably see how wacky I am by making arara still Java 5 compliant. :)
Even with version 4!
 
@PauloCereda Oh my. All of my applications are at minimum language level 7, but I do not have that big projects (and control over the updates on my institution's computers). That will be an interesting project (especially with JavaFX) ;)
 
@TeXnician Java 6 or 7 is surely the minimum requirements nowadays. :)
 
4:04 PM
@PauloCereda Well, starting off a new (and bigger) project I always use the newest version available for Debian stable (which is currently 8). Because I'm mostly sure that it will be everywhere when the program is in a usable state. Some call me pessimistic ;)
 
@TeXnician Oh if I write something now, it will definitely be version 7 or 8. :)
 
@PauloCereda My private projects are all in language level 8 as I love those lambda expressions (and JavaFX has some nice additions).
 
@PauloCereda you have it and can forward it.
 
@UlrikeFischer Thakns :)
 
@PauloCereda It is no secret. I will upload it to GitHub this week end (in my fork of the Christmas ducks).
 
4:14 PM
@UlrikeFischer Ah cool. :)
 
yo'
ok, my (para)medic training starts in an hour, I gotta go :-)
 
@PauloCereda @UlrikeFischer Thanks a lot!
 
@yo' ooh
 
yo'
@PauloCereda I have to go through it every 4 years in order to be a scout camp medic. It's only like 10 hours or training and then exams, but it's fairly useful!
 
@yo' Ah!
 
4:29 PM
@PauloCereda Sadly, I cannot make it work in beamer because of some option clash.... really sad....
 
@marmot oh no! @UlrikeFischer ^^ we need you!
Save the marmot!
 
@PauloCereda TEX.SE is a great site. I was able to fix the clash, and learned something ;-)
 
@marmot yaaaaaaaaay
 
4:59 PM
@marmot xcolor? It is a pain that there is no sensible interface to load the colors later. Imho there is no reason.
 
@PauloCereda simpler to blame @JosephWright surely?
 
@DavidCarlisle /sigh Probably
 
@UlrikeFischer Yes, but luckily I could fix it with the help of this site.
 
@ChristianHupfer yes, "destination unreachable" or "request timed out"
 
@JosephWright you could add a colo u r option:-)
 
5:08 PM
@JosephWright more xcolor's fault. I suggested years ago that they should be a \xcolorsetup command to set most package options later. This would make it much easier to load the package.
 
@UlrikeFischer Amongst other stuff, I'm working on colour support ...
 
Oops. I naïvely said \DeclarePairedDelimiter{\angles}\langle\rangle, but it was silently overwritten during \begin{document}, presumably by unicode-math. Not very cool!
 
@UlrikeFischer Probably over the next six months or so I'll have (x)color, graphics and a basic form of pgf implemented in expl3
 
@UlrikeFischer @PauloCereda Thanks so much! It all works very smoothly!
 
@JosephWright I know and this is good. It is simply a pain that with a "kernel" package like xcolor one either has to use package options or use some internal stuff.
 
5:18 PM
@UlrikeFischer xcolor is not the team's ...
 
user image
5
@DavidCarlisle ^^
@DavidCarlisle savethesounds.info
 
5:35 PM
@JosephWright I know. If it were I'm sure the situation would be better ;-).
 
5:57 PM
Hi mr. @Mooncrater, happy pizza day!
 
6:16 PM
Absolutely Sir. Paulo!
@PauloCereda
 
@Mooncrater woo
 
@DavidCarlisle Are you starting to support also good packages?
 
@egreg starting to regret doing that, yes.
 
@DavidCarlisle @egreg: you two behave. :)
 
@PauloCereda why?
 
6:30 PM
@DavidCarlisle no reason. :)
 
View from our building today.
 
@AlanMunn ooh snow
@AlanMunn ^^ I hate MS sense of humour.
 
 
4 hours later…
10:50 PM
@UlrikeFischer @PauloCereda The students loved the animation. Thanks so much for sharing it with me. And it is soooo much better than the stuff they were declaring pizza. ;-)
2
 
yo'
11:09 PM
@PauloCereda you should definitely come here one winter to see some snow.
 

« first day (2658 days earlier)      last day (2290 days later) »