@DavidCarlisle Did I do something wrong?

@WillRobertson no, just seeing if you were around:-)

I can't find amssymb on CTAN but texdoc amssymb gives me its documentation. What am I doing wrong?

@Kurzd first hit for ctan amssymb on google is ctan.org/tex-archive/fonts/amsfonts

@DavidCarlisle There's no page for it, but the documentation is in /doc, yet loading amsfonts doesn't load amssymb.

@Kurzd yes probably they'd make a ctan.org/pkg/amssymb entry if you pinged ctan@ctan.org and asked them:-) (@barbarabeeton <<<<<)

Hmm. The readme mentions a amssym.tex; amssymb.dtx, amssym.tex, amssym.def are in the /source folder. I am so confused.

@DavidCarlisle Is pinging Barbara enough of a ping or is the email necessary?

@Kurzd you need \usepackage{amssymb} ctan traditionally had a problem cataloguing things when a collection includes a package of the same name, that's why we changed "graphics" to latex-graphics (for the collection) leaving the entry "graphics" free for graphics.sty

@Kurzd no, but she may have a view on what any wording should be:-)

@Kurzd I'm not sure what you are confused about, the documentation organisation, or how to use the packages, or both?

@DavidCarlisle Hadn't noticed that. Is pkg/latex-amsmath the complete collection, while pkg/amsmath hosts amsmath mainly?

Confused about having two things with such similar names, having two different amssym.*, and having amssymb.dtx in the same folder. And the meaning of these extensions.

@Kurzd until last year it was just amsmath which was for the collection and the amsmath package itself did not have an entry. That is essentially the state that amsfonts still is in, the amsfonts catalogue entry is for the collection,a amsfonts, amssymb, ... not just for the amsfonts package itself.

@DavidCarlisle In that case, a quick thought: with fontspec+2e updates, there's no point me adding code to check if the 2e update hasn't happened because then I need to re-add the TU files in some messy way. I guess we need to check old fontspec with new 2e, though?

@Kurzd amssymb.dtx is tehe documented source of amssymb.sty, the latex definitions, amssymb.tex is a plain tex interface to the fonts (note this is in the generic fonts area of ctan not under latex)

@WillRobertson in practice I'd just release both at same time and not support anything else, although I think (since I managed to set a document or two) the current fontspec works with a new 2e format (even if some promised encoding variant interface isn't in the old version I think that's OK)

@DavidCarlisle great, that's enough for me

And it also means I can be slightly late without too much problem :)

@WillRobertson just don't break it with this evening's updates:-)

@WillRobertson hmm actually I should check the old tuenc.def would also work, I probably removed that, if we release before you update fontenc there will be two tuenc.def in the tree which would be sub-optimal, it's probably best to wait until there is a fontspec update that at least removes that?

anyway bed time, good night..

@DavidCarlisle Okay, I think this works — so we update fontspec first, which has a check IfFileExists{tuenc.def} and falls back to the old approach. good

@DavidCarlisle good night :)

@DavidCarlisle Sounds like amsfonts is going to get the same treatment amsmath had. Why is CTAN changing things around anyway?

Uh. Good night?

@DavidCarlisle Funny, that's actually what I do say about my code when I hand it off to students to work on :)

@Kurzd isn't that what you were asking for? ctan changed amsmath bundle name to latex-amsmath when I asked them to when amsfonts hasn't changed for years.

@DavidCarlisle will do in the afternoon

@DavidCarlisle Thank you. I deleted my answer.

Why did the duck cross the road? The chicken was busy!

Wondering: is there any reason at all to use split and not just aligned? aligned has the benefit of not being "full width" (which split is) so equation+aligned saves space over equation+split

@daleif the point about split is that it (can be) part of the outer alignment, compare:

\documentclass{article}

\usepackage{amsmath}

\begin{document}

split
\begin{align}
  a&=2\\
  \begin{split}
    b&=y\\
     &+z
  \end{split}\\
  c&=3
\end{align}

aligned
\begin{align}
  a&=2\\
  \begin{aligned}
    b&=y\\
     &+z
  \end{aligned}\\
  c&=3
\end{align}
\end{document}

@daleif aligned just makes a box so there is no way to align a= and b= but split does align these.

@PauloCereda yummy dinner last night, it involved pancakes and hoisin sauce.....

@DavidCarlisle uh oh

@DavidCarlisle noted, I'll write that down somewhere. My recommendation would then just be if your goal is to just have a single eq no, then use aligned, if part of a larger structure, consider split

That should be use equation (or gather) + split, larger structures, consider align+split

@barbarabeeton there should probably be an example like ^^^ in amsldoc :-)

@daleif Yes: the tag position is adjustable with split: vertically centered or top/bottom (which depending on the side where the numbers are printed); with aligned you only get center (or top with [t] or bottom with [b]).

@egreg I've very rarely needed it. I haven't updated my LaTeX book for a number of years, so perhaps I should re-evaluate some of the recommendations in it. At least I've changed some opinions on certain constructions.

BTW: here is the latest book project I've been involved in: en.unipress.dk/udgivelser/s/speciel-relativitetsteori

Probably the most comprehensible book I've made so far, we had control of almost everything, figures, line widths, colors etc. The cover is made by a professional (aka not me), the image is by Torben Ulrich, farther of Lars Ulrich of Metallica fame.

Bah. LuaLaTeX. This MWE:

\documentclass[paper=A6]{scrartcl}
\usepackage{fontspec}
\defaultfontfeatures{Ligatures=TeX,Scale=MatchLowercase}
\setmainfont[]{Andika New Basic}
\title{Říkanky}
\date{}

\begin{document}

Čistím zoubky řízy, řízy,\\
ať jsou bílé jako břízy.\\
Vlevo, vpravo, vpředu, vzadu,\\
horní a pak dolní řadu,\\
čistím zoubky řízy, říz,\\
ať má každý radost z nich.

\end{document}

LuaLaTeX does not render the 'ť' on the second and last line in 'ať'.

XeLaTeX has no problem.

vs

I should probably make this a TeX.SE post.

Well, I cannot delete the wall of text above. Feel free to do that. I have posted on TeX.SE instead.

@wilx ť should render with a "v" accent? (looking at your image) it looks like t' in the font I'm seeing in this chat

@DavidCarlisle Yes, in this case, in this font. This is for children to read. The usual adult ť is with apostrophe.

@wilx oooh I'm grown up!

@DavidCarlisle Girl, you'll be a woman, soon!

@wilx I found a Andika-R.ttf on the internet but using that both xe and lua latex render it as t'

oh that was wrong font...

@wilx A source for the font?

@egreg zip at software.sil.org/andika/download (you want the new one lower down)

@wilx I get the same with luatex 1.0.2 you could ask on luatex list

@wilx luatex says

Missing character: There is no ť (U+0165) in font [AndikaNewBasic-R]:mode=node;
script=latn;language=DFLT;+tlig;+tlig;!
Missing character: There is no ť (U+0165) in font [AndikaNewBasic-R]:mode=node;
script=latn;language=DFLT;+tlig;+tlig;!

Perhaps @WillRobertson may have more experience of such differences and whether specifying different scripts might help

@wilx if you look at the font with fontforge it has a T with caron but not t, presumably the glyph is hidden behind one of the font tables that luatex can't read. fontforge also warns that is ignoring several graphite tables.

@DavidCarlisle Yeah, I have noticed that only after I posted.

@wilx fontforge says there is no character in slot 0165 and no glyph named tcaron so it's probably not surprising that luatex can't find it as its code is partly derived from fontforge. I have no idea what xetex did

@DavidCarlisle I see.

@DavidCarlisle @wilx The survey also says that the t-caron is not in the font: software.sil.org/wp-content/uploads/sites/19/2015/12/…

@UlrikeFischer @wilx oh I think I can guess. it's something I saw Kahled mention. If harfbuzz (so xetex) can not find the character but it can decompose the character and find the accent and base it tries that.

@DavidCarlisle @wilx xelatex is producing a t + accent in the pdf. So it is a normalization effect.

@UlrikeFischer yes or rather its last attempt fallback in this case (it doesn't generally NFD normalise to split up composites)

Ah, thank you both. I did not realise the Andika New Basic missed the character.

I guess I will change the font to the plain Andika version then.

@wilx @egreg to the rescue! (for once:-)

@DavidCarlisle That's what I meant. But it is naturally quite a neat method -- you could ask Hans if luatex can do it too ...

\documentclass[paper=A6]{scrartcl}
\usepackage{fontspec}
\defaultfontfeatures{Ligatures=TeX,Scale=MatchLowercase}
%\setmainfont[]{Andika New Basic}
\setmainfont[]{AndikaNewBasic-R.ttf}
\usepackage{newunicodechar}
\newunicodechar{ť}{\v{t}}
\title{Říkanky}
\date{}

\begin{document}

Čistím zoubky řízy, řízy,\\
ať jsou bílé jako břízy.\\
Vlevo, vpravo, vpředu, vzadu,\\
horní a pak dolní řadu,\\
čistím zoubky řízy, říz,\\
ať má každý radost z nich.

\end{document}

@WillRobertson cf ongong TU discussion :-) ^^^^

@wilx just saw you asked the same on site so I answered there:-)

@DavidCarlisle Well, yes. I thought after the fact that it was better suited for the main site:

@wilx oh yes:-)

@egreg can you try \usepackage[tuenc]{fontspec} with \v{t} ? (I'm using the latest version that defaults to TU)

@DavidCarlisle I was just trying, stay tuned

@DavidCarlisle No, missing character

grrr OK I'll add a comment to my answer saying that the form you suggested might be needed, thanks

@DavidCarlisle Nice package newunicodechar.

@egreg in this case it's only \catcode'#1\active\def#1{#2} :-) But I'm confused now why declaring it as \v{t} works (@WillRobertson) tuenc defines \v{t} as a composite to U+0165 so if you declare the active ť to expand to \v{t} which expands to a non-active ť how come we don't end up back at the missing glyph? Strange (but it certainly works....)

@Kurzd -- amssymb is an add-on for amsfonts; it loads amsfonts rather than the reverse. on ctan, you will get a search result if you enable the "file names" option. there are many results (too many to be immediately useful), but if you look in the list for amssymb.pdf (the documentation), this will identify the underlying package, which is amsfonts.

@DavidCarlisle Remember I'm using the distributed version of tuenc.def

@DavidCarlisle I have \DeclareUnicodeComposite{\v}{t}{"0165}

@egreg yes but they both have \DeclareUnicodeComposite{\v}{t}{"0165} so I don't see why it worked for me.

@Kurzd -- amssym.tex and friends are for ams-tex, based on plain tex, not latex. related, but neither identical nor interchangeable. you'd get good information if you refer to the amsfonts documentation: texdoc amsfonts.

@egreg same here, or at least \DeclareTextComposite{\v}{TU}{t}{"0165}

@DavidCarlisle Maybe this is the point to investigate on

@egreg we have a team call tomorrow (I think) to finalise the default-to-tu arrangements so we can ship latex 2017/01/01 ......

@DavidCarlisle Something to do after dinner. ;-)

@DavidCarlisle Where is the new version?

@UlrikeFischer in an email Will sent to me overnight

@UlrikeFischer but the "new" tuenc.def is in the base latex sources in our public svn, but fontspec needed a few changes to fit in

@UlrikeFischer latex-project.org//svnroot/latex2e-public/trunk/doc/…

@DavidCarlisle -- will make a note. thanks.

@DavidCarlisle I can't find tuenc.def there. But \DeclareTextComposite{\v}{TU}{t}{"0165} is really the active definition of \v{t} it really doesn't make sense ...

@UlrikeFischer it's gone into the dtx alongside t1enc.def etc, but I need to run tracingall this evening and find out what it's doing:-)

@DavidCarlisle @egreg Do you know a good document to test math/text setups?

@UlrikeFischer The example file for amsmath?

@egreg that's an idea. I will check.

\edef\z{\v{t}}\show\z @egreg if you have a minute what do you get for <<

@JosephWright are we on for tomorrow or thursday now?

@DavidCarlisle Tomorrow: I'll make that clear on the list

@JosephWright OK, trying to understand @wilx 's example here tex.stackexchange.com/questions/349148/… which seems like a real world example of interesting composite usage with tuenc ...

Hi everyone! Why does \includegraphics[angle=-90, height=100mm]{tiger.pdf} issue a division by 0?

@AndreasMatthias because you rotated it to have height 0 then specified to scale to height 100mm so it works out as a scale factor of 100mm/0mm ....

@AndreasMatthiasuse totalheight= not height= (that is height+depth which isn't zero)

@egreg @JosephWright @WillRobertson think I understand the \v{t} weirdness \v{xxx} is unconditionally doing xxx\char"030C ignoring all the declared composites.... \DeclareUnicodeAccent

@DavidCarlisle Oh, I see... Haven't thought about that. Thanks.

@AndreasMatthias after over 20 years of answering that question I could have answered it even without seeing the \includegraphics example:-)

@DavidCarlisle I get (with \usepackage[tuenc]{fontspec} and miktex)

> \z=macro:
->ť.

@UlrikeFischer hmm (that's what you are supposed to get but I think I broke Will's code:-) as I get t\char "030C\relax need to back up a bit and see where it went wrong...

@DavidCarlisle ť

@DavidCarlisle Also

i\expandafter\show\csname TU\string\v\endcsname
> \TU\v=macro:
#1->\@text@composite \TU\v #1\@empty \@text@composite {\add@unicode@accent {"030C}{#1}}.

@egreg thanks, and @UlrikeFischer.

@DavidCarlisle I was afraid it would be like that. ;) But I couldn't figure it out myself. Having been absent from TeX for too long...

@AndreasMatthias well it's not like I thought of adding totalheight= originally, only added it after I tested rotating by -90 and early test files failed:-)

@DavidCarlisle Perhaps it is simply a question of the order: The \DeclareTextComposite are before the \DeclareUnicodeAccent. What should happen it such a case? Which one should win?

@UlrikeFischer no I messed up \v itself, compare @egreg's \show with

> \TU\v=\long macro:
->\add@unicode@accent {"030C}.
<recently read> \TU\v

l.27 \expandafter\show\csname TU\string\v\endcsname

@DavidCarlisle Please, add also \newcommand\showencodingcommand[2]{\expandafter\show\csname #1\string#2\endcsname}

Or \newcommand\showencodingcommand[2][?]{\expandafter\show\csname #1\string#2\endcsname}

Maybe with \encodingdefault instead of ? in the second version

@DavidCarlisle Well in my tuenc.def \DeclareUnicodeAccent{\v}{"030C} is before the \DeclareUnicodeComposite{\v}{t}{"0165} and if I add \DeclareUnicodeAccent{\v}{"030C} to my document then it wins and your answer to wilx example work and I get the tcaron.

@UlrikeFischer ahhh maybe you are right:-) thanks

@UlrikeFischer thanks that's it. I blame unicode for making me put the commands in the wrong order

(freddo intenso today -- and much wind)

$ svn commit -m "accents must be declared before the composites that use them" ltoutenc.dtx
Sending        ltoutenc.dtx
Transmitting file data .done
Committing transaction...
Committed revision 1340.

@UlrikeFischer mountain looks more interesting than Unicode accents but thanks for the fix!

@UlrikeFischer There was cold wind also here: bora from Trieste

@egreg My next project will be near Trieste, in Monfalcone

@StefanKottwitz Beware of the wind!

@egreg The Paul & Shark shop in Trieste will protect me!

@egreg is it even colder wind than you get in Barcelona in the summer?

@DavidCarlisle Quite. This is in today's news

ilpiccolo.gelocal.it/trieste/cronaca/2017/01/17/news/…

This can give you an idea

@egreg oops

@egreg Efff. That is a crazy strong wind. That said, I'm from inland, so...

MY PAPER SHOULD BE READY TO PUBLISH END OF THIS WEEK

EEEEEEEEEEEeeeeeeeeeeeeeeeee

@Canageek Yay!

@DavidCarlisle This strong wind is rare in the summer, but I'm always a bit scary when I do a certain road around Trieste, because it goes downhill all on a viaduct.

@DavidCarlisle More or less this is the height the road goes down to town from:

@StefanKottwitz The lights on the opposite coast are those of Monfalcone (Falkenberg)

@StefanKottwitz In between you see Miramare castle.

By the way, if any of you are interested my Dad and I are selling off somewhere between 1000-2000 fiction books. Most for ~$1. Shipping is from Canada, but our dollar is VERY weak right now. ($0.75USD last I checked)

docs.google.com/spreadsheets/d/… is our spreadsheet

@Canageek I love SF! Seeing A (Adams, Anderson, Anthony) and B (Bradbury, Bradley, Brooks) that's amazing! Even before continuing through that list.

@StefanKottwitz Those would mostly be my Dad's. He has been collecting since well before I was born, but is retiring to the other side of the country as soon as the house sells.

I've also got a seperate list of RPGs and board games if anyone wants a look

PLus a few hundred Star Wars toys for sale

@Canageek Now I'm at C! Wasn't Arthur C. Clarke the writer of the story with the moving forest?

@StefanKottwitz We may have sold all of our Clark. Some people came along and just bought entire authours. I know that is why there is no Moorcock left: I had a bunch of rare editions from the 60s and someone bought them all

@StefanKottwitz He might have been? I was never a big Clark fan.

@Canageek I love books, especially SF. I just cannot afford to have a house like a library :-D

@StefanKottwitz Eh, we picked them up slowly second hand. Not for a dollar each, but you buy two or three a year, times decades and decades....

I had 3 bookshelves in my room, my Dad had 4 in the guest bedroom, each at least two books deep

@Canageek Sure, the great books are affordable! I just meant the space in the house

@StefanKottwitz See: Why we are selling our books now that I live in an apartment and he and my Mom are moving into a condo ;)

@Canageek Some of my shelve space is blocked by a thousand math books starting at 1770 :-)

Wait, why am I in the room "MiKeTeX needs improvement and I need to calm down"?

@StefanKottwitz I'm getting rid of a lot of my textbooks. Too heavy to ship, and I can just check them out of the library, plus it is hard to sell 10 year old intro chem books.

@Canageek hmm I thought condo = Americanism for apartment but you seem to be using them to mean different things:-)

@DavidCarlisle In Canadian English it means either an apartment or apartment sized house that you own outright instead of rent.

@DavidCarlisle I don't know about American English.

@Canageek I searched for an old photo of my books. Today it's too messy, now at least.

@DavidCarlisle You don't own it as totally as a house though: For example the one my Grandmother owns (that my parents will be moving in to) has a Strata Council that limits how you can change them, as they are all supposed to be identical (within reason). So no knocking down walls or painting the outside

That's some of my math books 17 years ago. Now I have proper furniture but need a wide angle lens.

@StefanKottwitz Here is my RPG collection when I started (or at least the items I wanted to sell) lh6.googleusercontent.com/-FHwfk2n17Po/WFgevOsnp0I/AAAAAAAACt4/…

@Canageek ah thanks:-). it's just not a word at all in english english:-)

@StefanKottwitz My book collection (that I'm selling: I'm keeping 4 boxes of them) lh3.googleusercontent.com/-sR6z8eV7HdQ/WG6z5_bt5aI/AAAAAAAADWI/…

@StefanKottwitz And my Dads: lh3.googleusercontent.com/…

@StefanKottwitz Do you want some linear algeabra ones I have for sale? ;)

@Canageek Only in leather and gold :-)

@Canageek My wife likes fantasy books and anime such as Battle Angel Alita. That's space too. It all comes to the question, is there room enough for the books. (and the kid's books now....)

@DavidCarlisle Yes, having amssymb on another page would make more sense to me, and it appears to be a reasonable request.

@Canageek I think it was Alan Dean Foster but not Arthur C. Clark - wrong memory just because of the name starting with A. Thinking forests, I remember.

Have to go through the ABC of your list. :-)

When a 4th dimension opens in my apartment, I'll buy all SF books.

@barbarabeeton Let me see if I understood correctly. amssymb is an add-on for amsfonts. Because of this, loading amssymb loads also its parent amsfonts. I found this answer, tex.stackexchange.com/a/32103/83045 . amsmath is shown in the same tier as amssymb. Loading amsmath loads amstext. Does this mean amsmath, despite being bigger, was conceived as an add-on for amstext?

Also, two of that answer's links are broken. amssymb's (URL not found), and amsgen's (not found either, but changing .sty to .pdf shows the documentation). What's the edit protocol in this scenario?

@Kurzd it's sort of interesting to look at the history of how these things developed but it's long and convoluted and I don't think should drive current documentation. what is true now is that ctan is trying to move to a catalog that distinguishes what it calls packages (sometimes called bundles) from files (what to latex are packages) and so it helps if they have different names, so latex-tools is array,longtable,bm, dcolumn etc and latex-graphics is graphics, graphicx, color, etc

@Kurzd the ctan amsfonts entry is very old so probably wouldn't hurt to do as we did for the latex graphics bundle, make the top level entry something like ams-amsfonts or amsfonts-collection and then you can have an entry for each latex package as wel including the amsfonts package itself (@barbarabeeton)

@Kurzd no harm in editing to fix the links, but as werner is very active here you could just leave a comment suggesting he do it