TeX, LaTeX and Friends

yo'

00:18

Gotta love some SE sites :-)

samcarter

@LoopSpace A good opportunity for something that is probably not said often enough: tikzmark is a fantastic and very useful library!

marmot

02:27

@egreg I am referring to everything but Hopf wheat beer.

Harald Hanche-Olsen

07:15

@LoopSpace You never know why someone downvotes you. Many years ago I got a downvote on one of my answers on MO. This was puzzling, since my answer was clearly correct and had already got a number of upvotes. So I asked in a comment, and got a reply from the downvoter, saying that indeed, my answer was perfectly fine, but another answer was better, and he didn't think my answer should have more upvotes than the better one (which it did at the time).

yo'

08:19

@marmot There's the lawsuit between Czech Budvar and American Budweiser over the trademark. After the session is over, everyone goes for a meal. The Americans order: "Four Budweisers!" and the Czechs: "Two cokes!". The Americans looked shocked so the Czechs explain: "Well, if you don't drink beer..."

David Carlisle

@UlrikeFischer quite nice short page deriving that number (and suggesting one with better error behaviour) at tinaja.com/glib/ellipse4.pdf

Joseph Wright

08:46

@DavidCarlisle Once I have l3draw done, I'm wondering about doing xpicture (picture mode using l3draw as the base layer) :)

mickep

@JosephWright l3draw will be "the new" pgf?

Joseph Wright

@mickep Who knows :)

@mickep At present, I'm working on this as we know that the long-term L3 approach will require code that is targetted at the format as we rename all of the primitives

@mickep I think it's reasonable for an extended kernel to provide a full 'base layer' of drawing support, which is more-or-less pgf, even if one still expects complex user functions to be third-party

@mickep The next phase will be to try to 'swap out' pgf for l3draw: TikZ is set up so this should be (I hope) doable

mickep

@JosephWright Oh, I see. Do you have something up and running? And I agree that it is nice if there is a base layer of drawing support!

@JosephWright Cool!

yo'

@JosephWright well, I'm pretty sure it will be quite painful, but it should be doable.

Joseph Wright

@mickep If you follow the L3 checkins, I have some basics done but not yet enough to use without having to be familiar with the code. I need path usage working: probably a couple of weeks.

@yo' Yes: may or may not come off

@yo' Largely of course an experiment: if we can get it working, we've likely got the base layer right

@mickep What we are interested in really is seeing how the FPU does here: it should avoid the 'Dimen too big' stuff that can come up in pgf (the L3 FPU is essentially IEE754-compliant)

yo'

08:56

@JosephWright that's an alibistic approach, which I approve :-)

Joseph Wright

@yo', @mickep This is largely @egreg's fault: he posted an answer using the driver-level code I've already done, prodding me to get on with the base layer :)

mickep

Haha

Joseph Wright

I've already seen in a few tests that we come out quite different from pgf as the dimen rounding can be quite severe

yo'

Feb 9 at 11:33, by David Carlisle

@WillRobertson always have a person to blame in case of emergency:-)

Paulo Cereda

@LoopSpace oh no

@LoopSpace I am a \tikzmark fan! <3

Ulrike Fischer

09:00

@DavidCarlisle I like the word "the point sets .. the enthusiasm ..." ;-)

@JosephWright wouldn't it be useful to have a "\cs_if_exist_or_relax:cTF" in expl3?

Joseph Wright

09:22

@UlrikeFischer I'd need a use case: we've not had one to-date

Harald Hanche-Olsen

@UlrikeFischer “enthuasiasm” [sic]!

Ulrike Fischer

@JosephWright on c.t.t. there was just the case that french.ldf broke as a previous package used \csname XeTeXrevision\endcsname in a test and so \ifdefined\XeTeXrevision broke. etoolbox has for this \ifundef and the KOMA classes \scr@ifundefinedorrelax so I was wondering about the kernel solution.

Joseph Wright

@UlrikeFischer We are looking at changing \@ifundefined for exactly this case (@DavidCarlisle)

@UlrikeFischer At the expl3 level we'd expect 'native' code to always have a 'known' state

Paulo Cereda

@JosephWright I will not define silly L3 commands for a pound.

(no one mentioned this MP sketch so far)

David Carlisle

@UlrikeFischer github.com/latex3/latex2e/commit/…

typo in changes.txt I see :(

Paulo Cereda

09:37

@DavidCarlisle oh the changes should go to metachanges.txt

Harald Hanche-Olsen

@DavidCarlisle not your fault?

David Carlisle

@HaraldHanche-Olsen I suspect that it's the same as this chat, that my perfect typing gets mysteriously corrupted while being posted to the internet

Harald Hanche-Olsen

@DavidCarlisle The Internet is full of trolls. You could blame them.

David Carlisle

@HaraldHanche-Olsen good plan

Paulo Cereda

@DavidCarlisle you need some sort of design services

David Carlisle

09:46

@PauloCereda I tried that, but the results were very old fashioned and not that good

Paulo Cereda

@DavidCarlisle :)

Ulrike Fischer

@JosephWright Yes I remember. I was more interested on the expl3 layer (I'm currently rewritting some code in expl3 and constantly think about how to transpose things ...).

Joseph Wright

@UlrikeFischer Sure, but unless it's '2e glue' the problem shouldn't arise, and for the '2e glue' case a nested (explicit) test seems best

Ulrike Fischer

10:17

@JosephWright either I'm blind or \cs_if_exist:N actually gives false for \relax:

\documentclass{article}

\usepackage{expl3}

\begin{document}
\ExplSyntaxOn
\cs_if_exist:NTF \XeTeXversion {exists}{doesn't~exists}
\par

\csname XeTeXversion\endcsname
\cs_if_exist:NTF \XeTeXversion {exists}{doesn't~exists}
\par

\let\XeTeXversion\relax
\cs_if_exist:NTF \XeTeXversion {exists}{doesn't~exists}
\par

\def\XeTeXversion{blub}
\cs_if_exist:NTF \XeTeXversion {exists}{doesn't~exists}
\ExplSyntaxOff
\end{document}

mickep

@UlrikeFischer Nice poem!

Ulrike Fischer

@mickep well the s at the end should be removed (I have a tendency to overdo them ;-))

Paulo Cereda

@mickep Roses are red, violets are blue, wait a minute, shouldn't violets be violet?!

2

Because oranges are orange, so lemons should be green. :)

We ducks are very good at words.

yo'

@PauloCereda well, lemons are citrons in French, and guess what, citron vert is a lime :-)

Paulo Cereda

@yo' oooooh

Le canard!

yo'

10:31

@PauloCereda magret de canard!

Ulrike Fischer

@PauloCereda I'm not sure it is a good idea to mention canard and orange so near by each other.

Paulo Cereda

@yo' oh non

@UlrikeFischer they are mean, I mean, the French :)

yo'

@UlrikeFischer :)

@PauloCereda les frogs!

Paulo Cereda

@yo' ooh

yo'

anyway, time to run away! (seriously, I should be in a hurry, I have a meeting downtown in 50 minutes)

Paulo Cereda

10:43

@yo' Have a nice meeting, Tom!

d125q

11:26

Currently, there is a babel-macedonian package on CTAN (which I have contributed a little bit to) to support Macedonian in Babel. However, it does not play well with fontspec because of Unicode. I would like to extend it by adding Unicode support and also hyphenation patterns. Can anyone point me to some directions? Where do I look for references?

Loop Space

@HaraldHanche-Olsen Yeah, I know. I really just wanted to whinge about it. It seemed so ... unnecessary ... given that it was an old answer that I'd just updated. Anyway, thanks for the cancelling upvote!

Paulo Cereda

@LoopSpace ooh time travel

Loop Space

@PauloCereda Timey-wimey wibbly-wobbly

Paulo Cereda

@LoopSpace yes :)

Loop Space

@PauloCereda "Doctor Who and the \tikzmark of Doom"?

Paulo Cereda

11:36

@LoopSpace I like it. :)

Just don't go with \usedoctorwhocompanion{Clara Oswald} and everything will be fine. :)

David Carlisle

@d125q don't your hyphenation patters for pdftex already assume utf8 (most do and are converted to T1 encoding versions automatically)

d125q

@DavidCarlisle There aren't any hyphenation patterns for Macedonian at the moment. I would like to write some myself but don't really know where to begin.

David Carlisle

@d125q ah.

@d125q there was someone in this chat last year getting help setting up hyphenation for (if I remember correctly) albanian, which is possibly related?

@d125q for hyphenation patterns there are really two sorts of languages, some have very regular rules for breaking into syllables and adding hyphens and people have written patters that directly match those rules. Or ther eare languages like english which follow no rules and there the usual mechanism is to get an existing dictionary of hyphenation points from a traditional publisher and use the patgen program which automatically tries to compress the list of hyphenated words into a patterns

d125q

11:51

@DavidCarlisle Not related at all I'm afraid. I would probably get most help from looking at Russian/Bulgarian/Serbian hyphenation patterns and the process of writing them (e.g., things like automatic conversion from UTF-8 to other encodings, and so on).

@DavidCarlisle Macedonian has some pretty nice rules in that regard, but also some edge cases which would have to be taken care of manually. (And would probably have to be taken care of over a longer period of time.)

Loop Space

@PauloCereda \keys_define:nn {doctor / who} { companion .tl_set:N \g__companion_tl, companion .default:n = Donna\space Noble }

Paulo Cereda

@LoopSpace YES!

David Carlisle

@d125q the conversion would "just work" if you have the patterns in utf-8, you can see what is needed by looking at /usr/local/texlive/2017/texmf-dist/tex/generic/hyph-utf8/

@d125q hyph-pt for example looks like a regular set of patterns someone has written for portuguese (@PauloCereda) whereas hyph-uk.tex is much bigger and less regular (and is derived from some lists made available from oxford university press's oxford english hyphenation dictionary)

Paulo Cereda

@DavidCarlisle \o/

Ulrike Fischer

@DavidCarlisle there is a discussion about the typewriter package on the context list: "I guess it would be fun to have in ConTeXt" and "But the fun of the package is the added irregularity (while it still misses the typical super-bold dots)." <--- looks like another update is needed ;-)

d125q

12:10

@DavidCarlisle Thanks a lot! I will be having a look today and during the weekend.

David Carlisle

@d125q typ above the -uk ones are Ukrainian not for the UK, but you'll see. It is these days easiest to use lualatex to test these things as you can load \patterns into normal lualatex document, without having to use initex to build a special format. Once the patters are working in luatex they can be set up to work in pdftex and xetex (as the needed changes are the same for all languages)

@d125q see this example here

Oct 17 '17 at 21:10, by David Carlisle

\documentclass{article}

%\iffalse % use no patterns for albanian
\iftrue % use new patterns

\makeatletter
\newlanguage\l@albanian
\language\l@albanian
\lefthyphenmin2
\righthyphenmin2
\patterns{
e5rgji
}
\makeatother

\fi
\setlength\textwidth{6cm}
\usepackage[albanian]{babel}
\begin{document}


\showhyphens{ gjergj  gjergji}
\end{document}

It doesn't really matter at that point whether or not macedonian is like albanian, the language could be called foo the above just declares a new language and adds (only) a pattern influencing the hyphen after e in a string ergj so you just need a similar setup, with more patterns....

marmot

12:26

@PauloCereda Answer is no, see here

Paulo Cereda

oh

Keks Dose

13:39

Have "you" seen this question: tex.stackexchange.com/q/415604/4736 . I checked the md5 and the sha-1 of the package cm-super.tar.xz from the source given in the comments. The are different from the checksum printed in the log by the user372565. But I have no idea about checksums in texlive, maybe they are not md5.

David Carlisle

@UlrikeFischer ooh Hans and I can get into a my typewriter is worse than yours competition:-)

egreg

@KeksDose I've had the same problem with bidi yesterday

2018-02-15 22:13:33 +0000 Warning tlu_ipctask[27870] TLUtils::check_file: removing /var/folders/8b/hx9v58ln1txgzv4v32_vpbr40000gp/T/kuv8KuayCx/QW3tWZL9qT/bidi.doc.‌tar.xz, checksums differ:

2018-02-15 22:13:33 +0000 Warning tlu_ipctask[27870]	TLUtils::check_file:   TL=5db25b9467def8f963c5cdec267e946d44f0489998507862c4910d1970c27359b385a28cfa7a0d1d48248297a3c4dcb9dc935c1461af208bd5a0e1869090931f, arg=b8f9b8e1de58a1ad05b44a187f9ce1a972b5076a58a37f1677799f4458f88f6956a18a67f711c8058aad549520d42abc4953bf36d4bd1bb6df37c1455202ac55

Ulrike Fischer

@DavidCarlisle currently he seems rather reluctant "also, i don't expect this kind of hackery / show-off to be used". He obviously doesn't know the egreg design service ;-)

Paulo Cereda

@UlrikeFischer :D

Aug 23 '12 at 14:09, by David Carlisle

@AndrewStacey nah nah na nah nah: my pool is bigger than your pool

Keks Dose

Hm, how do we call Norbert tex.stackexchange.com/users/10829/norbert ? Failing checksums are something making me wince.

Paulo Cereda

13:44

@KeksDose I believe it's a mirror problem, perhaps trying another mirror?

Keks Dose

@PauloCereda And find out in three weeks that the cm-super.tar.xz carries some more data than cm-super on some mirrors?

Paulo Cereda

@KeksDose Sorry, I just suggested it.

David Carlisle

@PauloCereda how do you find these things

Paulo Cereda

@DavidCarlisle <3

David Carlisle

@PauloCereda speaking of finding things

Jan 12 '16 at 12:15, by David Carlisle

@PauloCereda finished your thesis?

Harald Hanche-Olsen

13:58

Jan 17 at 14:43, by David Carlisle

Jan 12 at 13:02, by Harald Hanche-Olsen

Nov 8 '17 at 10:40, by Harald Hanche-Olsen

Oct 17 at 6:51, by David Carlisle

24 hours ago, by Harald Hanche-Olsen

Oct 11 at 16:46, by David Carlisle

Sep 10 at 12:03, by David Carlisle

@HaraldHanche-Olsen do you ever get a feeling of déjà vu in this chat room?

Paulo Cereda

@DavidCarlisle oh no

Joseph Wright

14:36

uk.tug.org/tug2018

Paulo Cereda

@JosephWright awesome!

@LoopSpace ^^ will you come, please? :)

COTO

14:59

You gentlemen were so helpful the other day, I'm back for more. :) I have the following 3x3 matrix:

\[G\left( t\right) = \left[ {\begin{array}{*{20}{r}}
	{G_{aa}}               & {G_{ab}}               & {G_{a\times}} \\
	{-G_{ab}^\dagger}      & {G_{bb}}               & {G_{b\times}} \\
	{-G_{a\times}^\dagger} & {-G_{b\times}^\dagger} & {G_{22}}
\end{array}} \right]\]

As can be seen, the elements are all over the place. What I would like is for all three matrix columns to be the same width, with each column aligned on the right side of the `G`. Is this possible to do without inserting invisible hand-tuned shims everywhere? (In particular, the `a`, `b`, and `\times` subscripts may change in future, and it would be nice to not have to redo all the shims.)

Is it a big and novel enough task that I should ask it as a question on the main site?

Your TeXpertise would again be most appreciated.

David Carlisle

@COTO it would be easier to see if you had posted a full document but as an initial comment why specify 20 columns ? *{20}{r}

@COTO also i wouldn't do \left( t\right) you will get better spacing from (t)

daleif

@COTO Is this being generated by some program? Hardly anyone would write latex like that. Especially when you know about the bmatrix env from amsmath

barbara beeton

@yo' -- most u.s. states now do require photo identification to vote, although it's relatively recent (about 6 years for rhode island). but the rules aren't written clearly -- according to the most recent formulation, i could presumably use my brown alumni library card (it's from a recognized local university, and it has my photo). the person teaching the session for pollworkers blanched when i showed it to him and asked.

COTO

This is how it comes out of my equation editor GUI.

David Carlisle

@barbarabeeton we don't need any identification at all....

@COTO don't mention GUIs here:-)

COTO

15:09

Will the bmatrix package solve my alignment problem?

barbara beeton

@PauloCereda -- i don't mind having to get a visa. i do object to some of the requirements, though. why is it necessary to present a copy of one's most recent bank statement?

David Carlisle

@COTO no but any math in latex should use the amsmath package and that defines bmatrix which is like \left[\begin{array}... but with better spacing

Paulo Cereda

@barbarabeeton Oh I have updates regarding visa, hold on.

daleif

@COTO I haven't even looked at the problem yet. Perhaps you should post a proper question on the site with a full minimal example.

COTO

@DavidCarlisle I see. I'll look into converting my document.

@daleif: I'll post it as a question. Just wanted to make sure it wasn't a ridiculously simple thing.

Paulo Cereda

15:12

@barbarabeeton: vfsglobal.com/brazil-evisa/index.html now you can request an e-visa which is way easier than the traditional visa. The TUG2018 page is updated regarding this new feature.

barbara beeton

@yo' -- that's quite the opposite of my being asked, in hamburg, by a british couple, for directions. (never figured that one out.)

egreg

@barbarabeeton At one election, a voter handed in his “Permesso di condurre veicoli a motore a scoppio” (license to drive combustion engine vehicles), issued in the 20s. ;-) But I knew him personally, he was a neighbor and the father of a colleague of my mother's. He died several years later, aged more than 100.

Paulo Cereda

@barbarabeeton Brazil has wacky requirements.

David Carlisle

@COTO you might want to use G_{\mathrlap{...}} (mathtools package, an extension of amsmath) then the width of all the sub and subscripts will be hidden so everything will line up on the G, then you just need to make sure that the column sep is enough so the subscripts do not over-print

COTO

@DavidCarlisle: That's an excellent idea. Could I also use \smash{}? I learned about that recently, and it doesn't seem to require any special packages.

barbara beeton

15:16

@egreg -- well, i'd be happy to accept that as identification. one thing that has cheered me up, in my capacity of "warden" at polling places in elections over the past ten years, is the number of first-timers who are so excited to be able to take part in the process. i really hope they can hang onto that enthusiasm.

David Carlisle

@COTO nope \smash is vertical.. you could use \rlap with no package but you'd need \rlap{$\scriptstyle... so it's easier to use \mathrlap from mathtools

egreg

@COTO I would simply go with bmatrix, but here's how you can do

\documentclass{article}
\usepackage{amsmath}

\begin{document}

\[
G(t) =
\begin{bmatrix}
\begin{aligned}
 G&_{aa}                & G&_{ab}                & G&_{a{\times}} \\
-G&_{ab}^\dagger        & G&_{bb}                & G&_{b{\times}} \\
-G&_{a{\times}}^\dagger &-G&_{b{\times}}^\dagger & G&_{22}
\end{aligned}
\end{bmatrix}
\]

\end{document}

@COTO You have too many useless braces and miss the important ones. ;-) Here \times is used as ordinary, so bracing it will override its nature of binary operation symbol. Not really necessary here, but it could bite you somewhere else.

COTO

Looks excellent. I shall use that. But I'll remember your tip about \mathrlap too, David. That will surely come in useful.

Paulo Cereda

@barbarabeeton: is the e-visa information useful?

daleif

@COTO here is a different alternative

\documentclass[a4paper]{article}
\usepackage{mathtools}
\usepackage{xparse}
\NewDocumentCommand\G{ e{_^} } {
G
\IfNoValueF{#1}{% there is _ data
\sb{\mathrlap{#1}\hphantom{99}}
}
\IfNoValueF{#2}{
\sp{#2}
}
}
\begin{document}
\[
G( t) =
\begin{bmatrix*}[r]
\G_{aa} & \G_{ab} & \G_{a{\times}}
\\
-\G_{ab}^\dagger & \G_{bb} & \G_{b{\times}}
\\
-\G_{a{\times}}^\dagger & -\G_{b{\times}}^\dagger & \G_{22}
\end{bmatrix*}
\]
\end{document}

COTO

15:26

One last question. The matrix elements themselves, such as G_{aa} are actually defined in macros, e.g. \newcommand{\Gaa}{G_{aa}} since they're used in several places and I expect the subscripts will change. Will this still work properly if I use e.g. \newcommand{\Gaa}{G&_{aa}}? Should I perhaps define two versions of each macro: the first for when the element appears in an aligned matrix of this nature and the second for when the element appears by itself, e.g. $\Gaa = 1$ ?

It looks like @PauloCereda's solution obviates this issue, so never mind.

barbara beeton

@PauloCereda -- i'll let you know later. (right now, i actually have to do some work.)

Paulo Cereda

@barbarabeeton Thanks. According to PN, it's very simple and easy to apply. They are suggesting e-visa for ICM atendees.

egreg

@COTO A different strategy:

\documentclass{article}
\usepackage{amsmath}

\newlength{\cellwidth}
\newcommand{\MC}[1]{\makebox[\cellwidth][l]{$\displaystyle#1$}}

\begin{document}

\[
G(t) =
\settowidth{\cellwidth}{$G_{a{\times}}$}
\begin{bmatrix}
\begin{array}{@{}rrr@{}}
 \MC{G_{aa}}                & \MC{G_{ab}}                & \MC{G_{a{\times}}} \\
-\MC{G_{ab}^\dagger}        & \MC{G_{bb}}                & \MC{G_{b{\times}}} \\
-\MC{G_{a{\times}}^\dagger} &-\MC{G_{b{\times}}^\dagger} & \MC{G_{22}}
\end{array}
\end{bmatrix}

COTO

I'll give it a whirl in a moment. I'm compiling @daleif's suggestion as we speak.

Ulrike Fischer

3

@PauloCereda ^^^

Paulo Cereda

15:38

@UlrikeFischer LOLOLOL

COTO

@daleif: Your suggestion works beautifully, including with the macros. Thank you gentlemen one and all for your assistance.

Christian Hupfer