7:58 AM
quack

@PauloCereda A DUCK

@JosephWright ooh another duck

@PauloCereda Something like that

@JosephWright as long as pizza is safe, everything will be fine. :)

@PauloCereda Pizza should be fine, that part I'm not worried about!

8:12 AM
@JosephWright Meanwhile...

8:36 AM
@PauloCereda @JosephWright Three ducks in chat

@CarLaTeX ooh more ducks

@JosephWright when do you start?

8:52 AM
@CarLaTeX so I won't go hungry

@DavidCarlisle oh no

Oct 26 at 14:28, by Paulo Cereda
@DavidCarlisle you are mean

Nov 12 at 8:49, by Paulo Cereda
@CarLaTeX you are mean

@DavidCarlisle oh, no dungeon

@DavidCarlisle oh no!

8:59 AM
@UlrikeFischer to allow quick escape, should that prove necessary

@DavidCarlisle there are even windows!

9:14 AM
@DavidCarlisle ;-) wishing you a nice day doesn't sound appropriate - perhaps good luck?

9:34 AM
@UlrikeFischer there isn't even any pineapple on the pizza menu :(

@DavidCarlisle shocking

@DavidCarlisle I better don't ask about duck - it would shock @PauloCereda even more

@UlrikeFischer oh no
/quacks in despair

1 hour later…
10:41 AM
@JosephWright at the meeting?

@DavidCarlisle by the time, lunch I guess?
I am waiting for 14 GMT. :)

11:07 AM
Now Boris Johnson does not know the lyrics of The Wheels on the Bus.
@mickep ^^
This girl returned a rock to the park. :)

@PauloCereda rage!!

11:32 AM
@mickep woo

12:29 PM
@PauloCereda the teabag incident was way worse.

1 hour later…
1:29 PM
@PauloCereda -- The girl who drew the picture, or the person holding the picture -- who is a uniformed ranger? (The Smokies are beautiful, especially in the morning, when there are mists that outline the hills.)

1:46 PM
@barbarabeeton the person holding the picture. :)
10-minute countdown.

2 hours later…

3:50 PM

@CarLaTeX I think it becomes obvious if you follow the link …

@HaraldHanche-Olsen @yo' Oooh I didn't notice it is a link, sorry!

@CarLaTeX Yeah, links don't stand out very well here.

4:07 PM
Ah sorry for the confusion

@yo' I've got some vision problems :)

2 hours later…
6:02 PM
I many package documentations I've seen \begin{decl} being used to markup commands
Where is that environment from?

6:17 PM
@BrainStone Which package, for example?

graphicx and I think fontspec

@BrainStone not graphicx but: texmf/tex/latex/base/ltxguide.cls:\newenvironment{decl}[1][]%

I see
@DavidCarlisle Is that an extension of ltxdoc?

@BrainStone no a custom class just for clsguide

Hm ok

6:23 PM
@BrainStone also used for some latex3 xpackages xpackages/xlang/xlang.dtx for example

Any recommendation for documenting commands?

@BrainStone I like l3doc it is simplistic in its looks but really clear.

I'm currently using ltxdoc
Which one is the better?
And what's the markup for commands then?

@BrainStone the question "Which one is the better?" cannot really be answered as there are no well-defined criteria to measure....

@Skillmon Which do you consider to be better?
And can I just switch between them?

6:35 PM
@BrainStone \begin{function}{\foo}\begin{syntax}\cs{foo}\oarg{<opt>}\marg{<arg>}\end{syntax}<description>\end{function}.
@BrainStone I never used ltxdoc (I guess I'm too young, l3doc already existed when my first packages were created and back then I decided to use it instead of the other possibilities), so I can't really say

@BrainStone I never used l3doc, (I guess I'm too old, l3doc was written after I had written most of my packages) so I can't really say either, @WillRobertson

6:58 PM
@Skillmon I couldn't really find any (rendered/pdf) docs on the class. Do you know where I could find them?
@DavidCarlisle How would you document commands then?

@BrainStone texdoc interface3, for example

@PhelypeOleinik I'm talking about markup in a tex document

@BrainStone The entire LaTeX3 kernel uses l3doc

That's fair
I like the function and syntax environments form that class
I think I'll stick with that
Though changing brought this funny issue onto me: https://gist.github.com/BrainStone/e56904cea2ee611a9ec5034e7706f0fb
In short it's now complaining about the unicode characters. Seems like it ignores the options I pass to it.

@BrainStone l3doc loads hyperref, and you're using something not expandable in a section header. Please show what's on line 74

7:12 PM
\newpage
Line 73 is \tableofcontents
And I'm really not using anything in the headers

@BrainStone Multiple files, perhaps? The message just says the line number, not the file name.

That's the only file I've touched that even has line 74

@BrainStone Odd... Can you make a compilable example?

@PhelypeOleinik Doesn't happen with the ltxdoc class
I can give you the whole repo I'm working on
It's open source

@BrainStone ltxdoc doesn't load hyperref

7:15 PM
But I do

@BrainStone Serves the purpose

The branch l3doc-test has the change
Run make in the main dir. And be sure to have pandoc installed

@BrainStone ! Package fontspec Error: The font "LexendDeca" cannot be found.

Oh. Clone the repo recursively
After the fact you can do it like this: (one sec)
@PhelypeOleinik git submodule update --init --recursive

7:36 PM
@BrainStone as I say most of my packages use ltxdoc

I haven't found a decent way to markup commands with that class.
I don't really care about using either

@BrainStone The problem is the \url in the \author:
\author{%
Yannick Schinko\\
\texorpdfstring{\url{github.com/BrainStone/latex-lexend}}{}%
}

@PhelypeOleinik I'll try that in a sec
And a honest question, when I add the command docs, is this good to publish?

@BrainStone it's good enough to document all the packages in the base tools and graphics distributions.

@DavidCarlisle I'm not saying it's bad, just that I don't know much about it and that I don't understand how they markup commands

7:47 PM
@BrainStone I don't think your licence statement clear. pick LPPL or GPL, or give the users a choice of either, but what does it mean to say the package is LPPL but the source is GPL?

That the final package (aka the zip file or the one you find on CTAN (soon)) is LPPL, and all the code used to create that package is GPL

@BrainStone you can look at any of the sourcre filesa as examples,

@DavidCarlisle I have and I don't understand. The dtx syntax is throwing me way off

@BrainStone I can't guess what that is supposed to mean, if the sources are GPL then (more or less) anything you make from the sources is GPL too

@DavidCarlisle It doesn't have to be. I mean I can (as I did) say it doesn't

7:51 PM
@BrainStone if you mean the tooling to build the package is GPL but none of that actually ends up in the final distribution then OK, but i wouldn't have guessed that was what was meant from the licence statement

@DavidCarlisle That's exactly how I mean it
How would you word that?

@BrainStone don't refer to your build scripts as "source" as in a tex context that usually means the package itself, take for example the graphics package it's source is graphics.dtx which contains within it the full text of graphics.sty, it would be odd to make graphics.dtx GPL but graphics.sty LPPL. the actual ctan distribution is made by the l3build tool, that could easily be under a separate licence which happens to be LPPL as well in this case, but I could have used gmake (which is GPL)

Ok
I've called it "source repository" though
In my docs I put it like this:
The final package, the one you find on \href{ctan.org}{CTAN}, is licensed under
like you would use most other packages on \href{ctan.org}{CTAN}.

The source repository, the one you find under \url{github.com/BrainStone/latex-lexend},
License v3.0}.
Is that clearer?

@BrainStone yes:-)

Ok. Will copy that section into the README then too

7:58 PM
@BrainStone meanwhile what's in the fonts/vf directory, it looks like another ttf font, not a .vf ?

Yeah. It's just ttf's for now
Currently it's designed to require fontspec
I'll get to that eventually
Though would that be an issue for publishing it on CTAN?

@BrainStone oh it's planned as opentype variable not a tex v(irtual)font :-)

Yes

@BrainStone requiring fontspec is more or less a given for otf/ttf fonts. for the variable font you may want to change the file path as vf is taken: /usr/local/texlive/2019/texmf-dist/fonts/vf are tex vf fonts

Inside my package I've put them under <package>/font/xxx.ttf
Is that not ok?

8:04 PM
@BrainStone it the github you have one under vf and this is wrong.

Oh
I see
That's just a submodule

@BrainStone ah but is that submodule just an import of the upstream font development so the file structure given?

@DavidCarlisle In the final .zip it's different
For creating the zip I create a helper dir
And put the files in the correct places there
As I need to create actual files from the templates anyways

@BrainStone OK you realise I'm just flicking over the file layout from a tex viewpoint not actually reading the documents or code, so just initial impressions not all of which may make sense:-)

@DavidCarlisle All good
That's why I'm explaining it ;)
I can give you a preliminary version of the zip I intend on releasing to CTAN

8:07 PM
@BrainStone not sure I'll have a lot of time next few days

I'll be done this evening, so no worries
All that's still missing are the command docs
And it's like 8
And 7 of them are the exact same

@DavidCarlisle I can't quite follow
Do you mean it can easily take 20 years for a simple bit of docs to be added?

@BrainStone no look at the date on that document, it is 20 years old today (and has been my primary programming language since before that first version was released)

Ah I see
Nice :)
Can I use the automatic labels hyperref creates?
Like ref a section without having to create a label myself?

8:25 PM
@BrainStone possibly but it seems like a bad idea

@BrainStone they are only destinations not labels.

Fair
I only need it once anyways
No harm in creating the target label

@BrainStone people are always asking for auto-labels but it always ends up with effectively generating a label like section.2 and then to reference it you need \ref{section.2} but that loses most of the original point of the system that you can add/remove sections and the cross references still work, if you reference an auto-generated label you need to edit any cross references after a document edit

@DavidCarlisle That's a fair point. I'd be using section names anyways
Which is not perfect but better

8:57 PM
Do you know how I can get current math font (which is set for any  instance) and change current text (non-math) font to the math font? I need to be able to access \bullet in the math font through \XeTeXglyphbounds, and I can't (doing \the\XeTeXcharglyph\bullet returns zero), so I was thinking changing the font to math font would be something worth a try.

@bp2017 if you are not using unicode-math but legacy math it imho can't work.

@UlrikeFischer, oh, I'm starting to understand now. I thought unicode-math conflicts with amsmath. But now it looks to me like unicode-math is akin to XeTeX for math.

@bp2017 there isn't (in the pdftex setup) a single math font, there are 16 (times three sizes) so you would have to look at the \meaning of \bullet and work out which \mathfam it is and then look up in the nfss tables which font is assigned to that \mathfam (which is basically hoe bm works)

By the way, I found this tex.stackexchange.com/questions/332203/check-math-font (would be of help to me).

@bp2017 ?

9:05 PM
@DavidCarlisle, well, in plain TeX we couldn't use unicode, in XeTeX we can. I read it that without unicode-math we can't use unicode in math so unicode-math is in a sense similar to XeTeX (although it may not be a good comparison).

@bp2017 but since you are using xetex presumaby you are using TU encoding so why use $\bullet$ rather than \textbullet which is character hex 2022 ie •
@bp2017 saying unicode-math is similar to xetex is like saying an apple is similar to a London Bus. I suppose they have some things in common, but it is hard to guess what similarity you are seeing

@DavidCarlisle, because I need two types of circle bullets: filled ($\bullet$) and hollow ($\circ$), \textbullet is a filled circle.
So I decided to pursue the math character route instead.
@DavidCarlisle, the similarity I described: both support unicode (please correct me if I'm wrong).

Then use U+25CB ○ or U+25EF ◯
@bp2017 unicode-math is a tex package, it supports unicode if the tex system it is running on supports unicode (and actually stops if not) so it works with xetex and luatex currently

@bp2017 there are circles in unicode:
\documentclass{article}
\usepackage{fontspec}
\setmainfont{DejaVu Sans}

\begin{document}
^^^^25cb \textbullet
\end{document}

@UlrikeFischer so slow

9:13 PM
@DavidCarlisle I needed to find a font ;-)
@DavidCarlisle what is with xcolor for the robustness? It would adjustment too right?

@UlrikeFischer I just had faith that you would have one installed locally and css would find it (fallback font handling is useful at times:-)
@UlrikeFischer yes I left a note in that issue somewhere plan would be to to color then ping Uwe for xcolor I checked that if we do color first it wouldn't break xcolor it would just over-write them to the current definitions

@DavidCarlisle well if we ping Uwe we should also ask to remove this
\DeclareOption{fixpdftex}{\XC@append\XC@@pkg{\RequirePackage{pdfcolmk}}}

@UlrikeFischer I'll do color then assign the issue to you:-)

The only other package using it is changes and there I already ping the author.
@DavidCarlisle who did last time reach him?

@UlrikeFischer me, I contacted him a year or two back to give him advance warning about something (I forget what exactly:-) He updated xcolor to avoid any public clash at the time
@UlrikeFischer in case you were wondering tug.org/pipermail/uktug-announce/2019q4/000251.html

9:20 PM
@DavidCarlisle no, only does nothing and good candidate for complete removal.

@UlrikeFischer ah I could have disabled it this time...

@DavidCarlisle I saw it already 4 hours ago. I keep track of such things ;-)

Would someone mind to check if this package I created is ready to be uploaded to CTAN?
https://gitlab.project-creative.net/BrainStone/latex-lexend/-/jobs/9410/artifacts/browse
The zip file is just the directory there zipped

@DavidCarlisle well the changes package loads it (for no reason at all), so a bit for warning is needed.

@UlrikeFischer yes but it could be a do-nothing package oh I see it's already a do-nothing package unless you have a pre-historic pdftex
changes is dated this october so actually has a developer who can recall what the code is doing?

9:27 PM
@DavidCarlisle yes, I thought it could move from do-nothing to don't-exist. (The code should be archive somewhere, perhaps it can be used for other stuff involving marks.)
@DavidCarlisle it add commands for editing commands (deleted, changed, added), uses soul .
@DavidCarlisle Bruno is working on spaces ...

@UlrikeFischer I'll put the dtx in the old-sources repo and set up a stub pdfcolmk that really does nothing, won't push it to ctan until next time though

@DavidCarlisle ok, and if the two packages remove it we can remove it too.

9:46 PM
@DavidCarlisle I have nice variant of the miktex admin/user problem: someone managed to install miktex twice in two different modes. And now he has an outdated hyperref in the admin tree which his single user miktex can't update but finds nevertheless.

@UlrikeFischer enable wsl and install the linux texlive in there?

@DavidCarlisle a windows texlive would work too ;-).

@UlrikeFischer what's windows?

@DavidCarlisle didn't you have some today?

@UlrikeFischer we did, also we had a lift which was needed, the dungeon is a distinctly non-accessible location so that was a good thing all round.

10:06 PM
@DavidCarlisle more people than usual?
@DavidCarlisle pdfx has lots of settings like \pdfglyphtounicode{integraldisplay}{222B FE02} with variant selector. Do you think that the variant selectors make sense here?

@UlrikeFischer a few 12, I think including a couple who had not been before, so they got a strange intro, with no tex talks at all just agm discussion

@DavidCarlisle, @UlrikeFischer, but if I use a unicode value directly, the currently selected font must be able to support the character, which is a (big) problem. (I wouldn't want to load an entire fontspec package just for the bullets). If I were to use $\circ$ and $\bullet$, however, no additional packages would be required and no font selection either. I just need to be able to use \XeTeXglyphbounds with those characters, that's all.

@UlrikeFischer well, what is VS-3 here? I only know about the VS-1 variants in the math block? but if that was a defined variant I suppose it would make some sort of sense but does the code really allow multiple uniocde numbers there, I'd have to check
@bp2017 huh? \XeTeXglyphbounds is a query to the font for the character size so isn't going to do anything useful in the case the font doesn't have the glyph

10:23 PM
@DavidCarlisle, I am confused here. Let me try to use \XeTeXglyphbounds on a character not supported by the current font to see what happens.

@bp2017 also I don't think you can use it on a classic tfm font such as the default math mode as tex has no access to the character shape for tfm fonts

@DavidCarlisle, actually, I just realized I wouldn't be able to, since \XeTeXcharglyph works only on characters that are supported by the current font (or am I wrong?). (\XeTeXcharglyph\circ doesn't work.)
@DavidCarlisle, good point, I forgot about tfm constraint.

@bp2017 you are asking about the inked size of a specific glyph in a specific font so naturally that is 0 if there is no such glyph.

So there's no way to remove the white space around the aforementioned bullets to prepare them for proper resizing/scaling.

@bp2017 that depends, it seems you are approaching this from a strange angle but I really don't know your use case .

10:32 PM
@DavidCarlisle, I am trying to use a character as a bullet now, without implicating additional packages or even drawing with picture environment (since circles are limited in size). This is the fastest way to display a bullet but, at the same time, until I figure out how to get hold of that math character and read the whitespace around it (to remove with negative kern in a box), there is no moving forward with this.

@bp2017 If you load pict2e (pretty stable, if that's your concern), then the size limit of the circle is no more

For reference:
%$\show\circ$ % \circ=\mathchar"220E
%\the\XeTeXcharglyph\mathchar"220E % ERROR
%\the\XeTeXcharglyph"220E % ZERO

@bp2017 The environment is LaTeX's picture environment. Only the internals are changed. And the time it takes is pretty much negligible, unless you are drawing hundreds of thousands of bullets, in which case you can save the drawn circle as a box and the time it takes reduces to basically the same as the character

@DavidCarlisle you can have multiple values in the argument, e.g. you could map a ffi-ligature glyph to f f i. But I'm not sure if the use is sensible here. He seems to want to distinguish size, e.g. a bigger brace \pdfglyphtounicode{braceleftbig}{007B FE01}. But in unicode.org/Public/UCD/latest/ucd/StandardizedVariants.txt such use is not mentioned.

@UlrikeFischer yes as far as I know only FE00 combinations are standardised in the math blocks

10:48 PM
@DavidCarlisle so it looks like some home made encoding ;-). My miktex problem reported back that Oberdiek was installed on 2/6/2019 and packaged on 11/10/2019. What a mess ;-(.

@UlrikeFischer my wsl suggestion looking more attractive?

@DavidCarlisle I don't think that I can tell him how to get wsl at this time of the night when I don't know how to get it myself ;-).

11:27 PM
Hello!
Do you recommend "\" for a space in math mode?

@manooooh What kind of space?

@manooooh unlikely (assuming you mean \  it's mostly a text mode thing

@egreg I am discussing with Carlos Ivorre (a logician) about the use of consecutive quantifiers in Logic. For example we have 4 options:
He told me he uses option (a), and I use option (d). So then I asked him if he uses a space to separate quantifiers, and sometimes he uses "\" (not \, nor \: nor \;), and between A and ( he does not always put a space in there
@DavidCarlisle yes I believe that, but a logician uses "\" for sometimes separate quantifiers :(
So do you recommend the use of "\ "?

@manooooh I'd use \, although actually I'd probably use . more like \exists e\in A . \forall x \in A . x*e = e*e = x

@DavidCarlisle yes I would use \,, too. May I ask why not "\ "?

11:39 PM
@manooooh but it's 30 years since I did logic for real so styles may have changed:-)

@DavidCarlisle WOW would you use a real dot (.) with no spaces?

@manooooh \  just isn't a math space, it's the word space from the surrounding text font
@manooooh a . I may make it have a bit more space, it depends

@DavidCarlisle but you don't have a criteria to determine "When place a space or when not", right?

@manooooh not after 30 years and 15minutes to midnight, no:-)

Hahahha
You're right, it is not time to ask things about like 30 years ago :)
@barbarabeeton what does AMS say about the above discussion? Thank you!
@DavidCarlisle 30 years and 10 minutes to midnight force you to transcribe e*e instead of e*x :P

11:51 PM
@manooooh -- I can't really answer this one; I'm not a mathematician myself, and I'm now retired from AMS, so don't have any ready contacts, especially not on Saturday night. The best I can do is recommend the AMS Style Guide, which may or may not mention it. Logic isn't the most frequently appearing area in AMS publications.

@barbarabeeton yes I checked the Style Guide and it seems it does not mention my question. Don't worry, I can survive without a convention, thank you for answering on a Saturday night!! Have a good night