12:00 AM
@DavidCarlisle what is Murrays tr28?

@Ell tn28 sorry the plain text math encoding I linked to above

Oh I see yes
I guess the first step if I were to implement this then would be to write a tn28 parse to some kind of internal tree structure

12:27 AM
@Johannes_B Here five

Thanks for the input @DavidCarlisle, this info is very useful

@Alenanno 8, 2 before date change, 6 after

Sleep time, night all

12:45 AM
@Ell night!

8 hours later…
8:37 AM
@DavidCarlisle I was going to ask about JSbox and extended syntax for \<thing>code, but I see you've done it already :-)
@DavidCarlisle I think the point about \pdfstrcmp got missed: the comparison is very simple (probably the Lua version is the one I'd point to, as it's just a thin wrapper around C)

8:51 AM
@JosephWright I fear compatibility with other Unicode engines isn't a priority (or even on the list of things to be considered:-)

@DavidCarlisle Indeed
@DavidCarlisle I'm still not clear on how font loading and shaping works, either

@JosephWright also I don't know if he's said how it will be released, if he's been working on it for some years is it a commercial product?

@DavidCarlisle I foresee more of the 'read manual, give up, do tests, ask by email, get unclear answer, try more stuff, ...'
@DavidCarlisle Indeed: what's the 'client program' here?

@JosephWright surely that would never happen

@DavidCarlisle We are possibly safe at the moment as not only is Unicode cross-compatibly a priority but neither's working with LaTeX2e
@DavidCarlisle I don't get the file loading issue: even in plain you need more than one file for a serious document

8:57 AM
@JosephWright I think he means he's got file loading but no search path so everything needs to be in one directory, so he doesn't mind a few document files but doesn't want to flatten the whole of ctan into his working directory
@JosephWright we are safe, but if jsbox is open source and in texlive then Will is going to be expected to make fontspec and unicode-math work with it which will be a pain if there are no Uxxx commands.

@DavidCarlisle A bit like KerTeX then
@DavidCarlisle I think 'impossible' is the word your are looking for

@JosephWright never looked at that:-)

@JosephWright: I could not add you as a friend because I had to purchase something first.

9:15 AM
@PauloCereda :-(
@DavidCarlisle Perhaps one for the team meeting: probably others are not aware of what's going on

@JosephWright Now I bought it, but you don't appear as a valid name. :) Try adding me: paulocereda

@PauloCereda Will try later today

@JosephWright thanks. :)

@DavidCarlisle Reading the announcement, it seems pretty clear that there's no font shaper as it mentions .tfm files. So this looks a bit like upTeX: not really a Unicode engine.
@DavidCarlisle unicode-math needs expl3, which needs \(pdf)strcmp, which we are told isn't available

@JosephWright I know, I just want to get it on his radar
@JosephWright I know....

9:28 AM
@DavidCarlisle Fair enough
@DavidCarlisle As you said last night, the idea looks like one derived from an environment that ignores XeTeX and LuaTeX development

@JosephWright yes shame really, given the environment he's describing, I'm assuming it's C and he could have started from luatex. (although that may have slowed it down a little:-)

@DavidCarlisle Yes, although LuaTeX of course is deliberately not compatible with TeX90 and he might not like all of their decisions
@DavidCarlisle If the code where public somewhere ...
@DavidCarlisle Would it be worth sending him a direct mail just with the team on the CC list?

@JosephWright Did we fix a time for the meeting? 11 is probably OK for me, although as I say I may have to mute/drop out from time to time
@JosephWright possibly

@DavidCarlisle Seems to be 11, though I'll see if we can make it 12

@JosephWright that sounds like it may be getting late for Chris, and it isn't necessarily any easier for me, I'd just confirm 11

9:37 AM
@DavidCarlisle I've asked if we can do 12, but if not then 11 it is

@JosephWright oh yes got that now

10:27 AM
2

11:18 AM

@DavidCarlisle Coding in bigints.sty is “strange”? This should go in English dictionaries for explaining “understatement”.

11:36 AM
@egreg had you seen that package before? First time I looked...

@DavidCarlisle Yes, just to tell people “don't use it”. ;-)

11:51 AM
@egreg I have to confess that it was hard to hit +1 on widerequation

@yo' :)

@egreg just please promise you'll never package it.

@yo' I solemny declare

@egreg s/mn/mnl/
:)

@yo' You're the expert in consonant clusters

11:54 AM
@egreg :)

Guys, do you think is there a way for us to create a TeX document to make our new series of ads? It had to be dimensions measured in pixels...

@PauloCereda you can create it at the correct size and then do convert -density 600 bla.pdf -resize 300 bla.png

@yo' Yep, that's the plan. :) Is there a way to create a document with this measurements?

@PauloCereda \usepackage[paperwidth=300px,paperheight=220px,margin=0pt]{geometry}
Or:
\documentclass[tikz,margin=0pt]{standalone}
\begin{document}
\begin{tikzpicture}
\useasboundingbox (-150pt,-110pt) rectangle (150pt,110pt);
...

@yo' OH MY!

12:01 PM
@JosephWright a little challenge for l3build log normalisation functions: mathemaesthetics.com/JSBox/triplog.pdf

@DavidCarlisle Looks like it won't work properly as an 8-bit engine

@DavidCarlisle I doubt he's worrying about using inputenc tricks

\documentclass[12pt]{article}
\usepackage[T1]{fontenc}
\usepackage[utf8]{inputenc}
\usepackage[
paperwidth=220px,
paperheight=300px,
left=0.5cm,
right=0.5cm,
top=1cm,
bottom=1.2cm]{geometry}
\usepackage{tgpagella}
\usepackage[demo]{graphicx}

\pagestyle{empty}

\begin{document}

\centering\large Do you understand cricket?\\ Neither do we.

\vspace{2em}

\includegraphics[width=3cm,height=3cm]{blokes}

\vspace{2em}

But we know \TeX.

\vfill

\end{document}
How do you guys like this code?
Also, any suggestions on how to include a border? :)

12:15 PM
@PauloCereda What kind of border?

@DavidCarlisle Uh-oh. :)
@Alenanno I thought of just a frame, a line surrounding the whole area. :)

Allan Robert Border AO (born 27 July 1955) is an Australian former cricketer. A batsman, Border was for many years the captain of the Australian team. His playing nickname was "A.B.". He played 156 Test matches in his career, a record until it was passed by fellow Australian Steve Waugh. Border still retains the world record for the number of consecutive Test appearances of 153 and the number of Tests as captain. He was primarily a left hand batsman but also achieved sporadic success as a part-time left arm orthodox spinner. Border amassed 11,174 Test runs (a world record until it was passed by...
@PauloCereda this one ^^^

@DavidCarlisle ooh

Like that?

12:19 PM
Add \usepackage{tikz} (you might have to remove the demo option for graphicx), and then add:
\begin{tikzpicture}[overlay,remember picture]
\draw[line width=5mm] (current page.north west) -| (current page.south east) -| cycle;
\end{tikzpicture}
to your document, of course. :P

@Alenanno (current page.south east) rectangle (current page.north west)? (Not that it matters much.)

@TorbjørnT. True, didn't think of that. :D Less code FTW!
@TorbjørnT. But I'd do (current page.south west) rectangle (current page.north east). I always draw rectangles bottom-left -> top-right. I thought everybody did that. Interesting.

@Alenanno It doesn't matter which two corners you use of course, or which direction you put them in. There was no particular reason why I did it in that order.

@TorbjørnT. Yeah I know. I was talking about the mind process. :D
Like when someone writes the number 3 starting from below vs starting from the top.

@Alenanno 'Convoluted' would be one to describe my mind process.

12:27 PM
@TorbjørnT. You never show it in your code, though. :P

\documentclass[12pt]{article}
\usepackage[T1]{fontenc}
\usepackage[utf8]{inputenc}
\usepackage[
paperwidth=220px,
paperheight=300px,
left=0.5cm,
right=0.5cm,
top=0.5cm,
bottom=1.2cm]{geometry}
\usepackage{tgpagella}
\usepackage{graphicx}
\usepackage{tikz}

\newcommand{\drawborder}{%
\begin{tikzpicture}[overlay,remember picture]
\draw[line width=5mm, color=cyan] (current page.north west) -| (current page.south east) -| cycle;,
\end{tikzpicture}}

\usetikzlibrary{fit,calc,shadows,backgrounds,positioning}
Courtesy of Paul Gaborit's code. :)
How do you guys like it?

@PauloCereda Nice! :D By the way, no need for color= in this case, just cyan will do. And replace the line \draw (after the square brackets) with this, less code: (current page.south west) rectangle (current page.north east)

@Alenanno Do you mind editing my code? I have no TikZ fu, and I am sure I wll make huge mistakes... :)

\draw[cyan, line width=5mm] (current page.south west) rectangle (current page.north east);
:P
Or whatever line width you wanted. :D
How did you write the code above without TikZ-fu come on.
:D

@Alenanno CTRL+C fu. :)

12:35 PM
@PauloCereda Ahah well, that's how you learn. :P
You could add a pale background. like cyan!50 or less.
I'm not a color-theme master though

\begin{document}
\pagecolor{cyan!30}

You could modify the drawborder command, like
\newcommand\drawborder[2]{%
\begin{tikzpicture}[overlay,remember picture]
\filldraw[line width=5mm, draw=#1, fill=#2] (current page.north west) -| (current page.south east) -| cycle;,
\end{tikzpicture}}
And then you can write, \drawborder{cyan}{cyan!50}
or if you plan the colors to be simply different gradiations then even one argument is enough
and you write
fill=#1!50
and you only write one argument

@Alenanno Fantastic!

\newcommand\drawborder[1]{%
\begin{tikzpicture}[overlay,remember picture]
\filldraw[line width=5mm, draw=#1, fill=#1!50] (current page.north west) -| (current page.south east) -| cycle;,
\end{tikzpicture}}
so it's \drawborder{cyan}

@Alenanno I like the 2 arguments option. :)

12:41 PM
By the way, if you plan the whole document to be inside a tikzpicture, you can (should) remove the tikzpicture environment from the new command definitions.

@Alenanno Enhance our code, please. :) I don't understand half of the hints. :)
The other half is delicious. :)

You can write the code as
\newcommand\drawborder[2]{%
\tikz[overlay,remember picture]{
\filldraw[line width=5mm, draw=#1, fill=#2] (current page.north west) -| (current page.south east) -| cycle;,
}}
Without the environment tags
@Alenanno Which part of automatically do you not understand? I already wrote this in the question, but I’ll repeat it for you: if I spot these cases in the document, I can take manual action, but I want it to at least not spread content into the page margin automatically. — mirabilos 3 mins ago
Uuuh... ok.

@Alenanno done. :)

Basically instead of the tags, it's \tikz[overlay, remember picture]{ ...code... }
It's good when you have a text document and you want to insert a single tikz command without having to set the whole environment.

12:56 PM
4
3

:D

@PauloCereda Bahahah

This will be the best series ever.

47 mins ago, by David Carlisle

1:02 PM
@DavidCarlisle Oh no!

@PauloCereda doesn't time fly when you are procrastinating?

:D

@PauloCereda You naughty, naughty duck :D

@DavidCarlisle Yep. :)
@Alenanno Pretty much. :) Him and @barbarabeeton. :)

@PauloCereda speaking of which you could do an advert for people with missing shift keys

1:05 PM
@Alenanno I am the guy who cheers @PauloCereda up when @DavidCarlisle was angry with him ;-)

@DavidCarlisle ooh
3
@DavidCarlisle: ^^ :)

lualatex KERNEL.tex
......
Overfull \hbox (4.7766pt too wide) detected at line 1942
\T1/cmr/m/n/10 2.1940
) [34] (/usr/local/texlive/2015/texmf-dist/tex/latex/base/t1cmtt.fd) [35]
[36] [37]Segmentation fault

I wonder if it my large latex file. sometimes I get segfault from lualatex on very large latex.... trying to find out why...

@PauloCereda What about a: Quack! :-)

@RomainPicot coin coin!

@PauloCereda Quack Quack ! :-)

1:17 PM
@PauloCereda double rasterization, ouch! Hint: You can generate the gravatars and almost any size, by changing the parameter in tez linkz

3
@egreg: ^^
@yo' :(
@yo' oooh

@Paulo gravatar.com/avatar/… could be better

@yo' Not that better, sadly. :(

@Nasser gdb --args lualatex KERNEL.tex
Are there really no guillemets on OT1?

@PauloCereda indeed, I know. It should be possible to convert to curves, it's just squares and triangles.
Maybe the algorithm used by gravatar is known ... ?

1:20 PM
@yo' :)
Or I could add them by hand in Gimp.

now I gotta go visit my mom. See you later guys
@PauloCereda the rotation would still cause issues.

@yo' Indeed. :(
@yo' All the best, pal.

@PauloCereda thanks!

@PauloCereda :D

@Nasser I think a segfault (for any input) is a bug in the system, so you should report it on luatex list or their bug tracker (luatex list first is usually best then let them open a tracker bug if needed)
@HenriMenke what? you aren't writing in English? :-) people used to use $\langle$ and some suitable negative spacing

1:31 PM
@DavidCarlisle So there is no shortcut to guillemets in OT1. I don't use these very often. First, I use T1 fonts or OpenType fonts and second, I use \enquote from csquotes. It just occurred to me when answering tex.stackexchange.com/questions/287254/…
I saw that there are new Lucida families available from TUG tug.org/store/lucida/order.html
I own the OpenType version and I was wondering if one could buy an upgrade.

@HenriMenke there are only 127 characters and have to get a-zA-Z0-9 and (of course:-) upper case Greek plus all the accents and Scandinavian letters doesn't really leave room for variant quotes.

@HenriMenke I was actually hoping she tried making a matrix, but it's ok lol :D

@HenriMenke: our new series is almost ready! (scroll up to see some samples)

@DavidCarlisle So according to your previous message, writing in languages other than English is unconceivable, but having Scandinavian letters but no French quotes is perfectly resonable? ;)

@HenriMenke Englishmen don't like French :'(

1:45 PM
@HenriMenke Knuth doesn't sound like a French name to me.....

@DavidCarlisle But neither does it sound Scandinavian

@DavidCarlisle the problem is that it does not happen all the time. so can't make a MWE still. never seen segfault from lualatex before.

@Nasser Sounds like a Heisenbug.

@HenriMenke so they got lucky:-) anyway if you haven't got the letters you need things are worse than if you just have to go to a different font to pick up some punctuation characters:-)

@DavidCarlisle This question just came to my mind. In plain.tex fonts are preloaded and hence compiled into the format. Is this also possible with OpenType fonts in LuaTeX?

1:50 PM
@HenriMenke no

@PauloCereda This is gold!
@DavidCarlisle Can luaotfload.sty be compiled into the format?

@HenriMenke the OTF are handled by lua callbacks with the font info in lua tables and no lua state can be stored in the format
@HenriMenke see above :-)
@HenriMenke however see recent thread on luatex list

@DavidCarlisle This makes things really awkward. For instance, when I first tried xetex I just wrote äöü\bye and was puzzeled why the document was empty.

English people: is it chat room or chatroom?

@PauloCereda the former (usually, except when it isn't)

1:54 PM
@DavidCarlisle I am stuck. :)

@HenriMenke yes, we know, sorry the latex user experience is terrible there. It's easier to see that it's bad than to see how to change it.
@JosephWright ^^^

@DavidCarlisle “\everyjob{\RequirePackage{luaotfload}}”. That's what I suspected.

\everybreathyoutake{\iwillbewatchingyou}
@egreg ooh

@egreg Thanks for making me loose my faith in English grammar googlefight.com/comprised+of-vs-comprises.php

1:57 PM
@PauloCereda This is English, we don't really understand compound words

@DavidCarlisle :)
3
@DavidCarlisle: ^^ SCIENCE

2 hours ago, by David Carlisle
2 hours...

@DavidCarlisle ouch
@HenriMenke yaaaaaaaaaaaaaaay

1:59 PM
@PauloCereda remember that usage consists mostly of people asking how to get out of vim.
3

@DavidCarlisle @PauloCereda or how to close it without using kill command

@DavidCarlisle wow

@PauloCereda I win

@DavidCarlisle This is ridiculous. When you Ctrl+C in vim it tells you how to exit.

2:03 PM
@HenriMenke but not if you made the mistake of accidentally getting into insert mode first

@DavidCarlisle smash ... oh great, now it's recording!

@DavidCarlisle -- when tex was first created, \outer was very useful for avoiding nested definitions and overrunning the hash table. maybe doesn't make so much sense now, but it really did save my neck then.

2:17 PM
@barbarabeeton but it doesn't really work even for that, it doesn't make people avoid nested definitions it just makes them convolute their code to avoid the error from \outer. See the etex.sty from yesterday I had \let\globcount\newcount which is readable and having found that people use it with plain I changed it to \expandafter\let\csname globcount\expandafter\endcsname\csname newcount\endcsname which is the same except that it is unreadable mess that does not generate an error from \outer

@PauloCereda None for me? I am curious to see how you stereotype me xD

@Alenanno <3

@yo' -- why not bp instead of pt?

2
@Alenanno: ^^

@PauloCereda :D ahah I'll make one too.

2:31 PM
@PauloCereda May I have a stereotype too? :3 Quack !

@DavidCarlisle -- maybe "people" do, but i learned very early that it was helpful. (as i said, running out of hash table space was endemic for math.)

@RomainPicot ooh sure!

@barbarabeeton yes I know: we had fun getting 2e+amsmath to fit in emtex's hash table. I struggle to see how \outer would have helped, but (of course) I believe you if you say it did!

@Alenanno -- a bit heavy. to me, it looks rather like a funeral card. don't think that's quite the effect we're looking for. but i suppose that's only because it's black/white -- better in color.

@barbarabeeton It was just to show the border @PauloCereda was asking for, the color came later. :P

2:37 PM
3
@RomainPicot: ^^ :)

@PauloCereda Coin Coin ! <3

@RomainPicot <3
If I ever manage to go to France, I will wear a coin coin T-shirt!

@PauloCereda -- \uppercasenonmath

@JosephWright on jsbox and logging tug.org/TUGboat/tb35-2/tb110mckenna.pdf

@barbarabeeton oh I will fix it. :)

2:49 PM
@DavidCarlisle Yes, well I wasn't involved in the decision to stick with OT1 in 1992 so ...
@DavidCarlisle I've probably read it once but will do again ...
@DavidCarlisle More seriously, we do have to consider how to deal with this longer-term (unless of course I can force people to get L3 finished)

@HenriMenke -- s/loose/lose ... but don't underestimate archaic (human) editors; i'd never let you get away with that, even if googlefight shows that that usage is taking over.

@JosephWright well only if it's open source if it is commercial they can make it work with latex:-) (@barbarabeeton do you know the plans for jsbox?)
I took this course with Nick:

@DavidCarlisle You work with Nick Higham?

@RomainPicot we were same undergraduate year
@RomainPicot took almost no courses in common by the end:-) But we both took the course described there.

3:04 PM
@DavidCarlisle Oh nice ! I have one of his book on my desk.
@DavidCarlisle My PhD subject fit in some part in his research field :-)

@RomainPicot Given that I avoided all the Numerical and Applied courses, odd that I should end up working here (NAG=Numerical Algorithms Group:-)

@DavidCarlisle I never had any Numerical or Applied courses too you know ;-)
@DavidCarlisle But I read and learn fast

@DavidCarlisle -- don't know jsbox plans, but intend to inquire. after all, he owes me a tugboat writeup. at the tug meeting where he made the presentation, he was most adamant about maintaining trip test compatibility, modulo the enhancements he wanted -- interactive graphics capability and (much!) better and more complete error reporting. his demonstrations were staggeringly impressive.

3:33 PM
@DavidCarlisle I like them all. :-)

@JosephWright slides look more promising (MATH table is mentioned and "Many sub-problems left to solve; XETEX primitives to incorporate; font feature support; etc.")

@barbarabeeton does it matter? I'm for "create a large bitmap and scale down" approach anyway... But you're right that with bp you get the font sizes roughly correct...

@yo' -- probably doesn't matter, but i'm lazy and try to avoid having to make small adjustments when a reasonably precise match of units is already available.

@barbarabeeton so you would then simply export with 72dpi and "be sure the result is exactly 300x220px"?
@JosephWright Do you think the SO event is valuable at TeX.SX? I was just told that it's you who can remove it...

3:49 PM
@yo' -- i'm making the rash assumption that the bp measure is closer to the physical characteristics of monitors, whereas using 72.27pt = 1in is inevitably going to result in rounding. (i could be wrong.)

@barbarabeeton well, the point is, convert (ImageMagick) takes the vector PDF file and creates a raster PNG; there is no rounding going on, basically. Unless you include a raster in the PDF file and want to preserve its quality on output; but then, much more need to be done.

@yo' -- i think you've just demonstrated more actual knowledge about the conversion than i have. so as long as you can ascertain that the horizontal pixel count is 300, the result should be acceptable. and, by the way, with 300x220px, wouldn't the image be wider than it is tall (i.e., landscape)? i was visualizing the new format to be more square than the old one.

@barbarabeeton I haven't checked for the height of the new ads...

@yo' -- i just looked back at the announcement, and it says 300x250px. (you should check; don't automatically assume i got it right.)

You're right, the new ad size is 300x250 px landscape

3:59 PM
@egreg Where in the TeXbook do I find information about the behaviour of the character | in horizontal and vertical mode?

@PauloCereda You should re-do all your new sexy ads @ 300x250 px

@yo' Oh my!

@HenriMenke -- i'm not @egreg (i'm purple, not green). in the texbook, p.52, it says "... you don't really want them when you're typesetting text. [...] vertical lines and relation signs are needed only in math mode."

@barbarabeeton Hm. {\tracingall| || |||}\bye outputs three very long rules of increasing length.
The log say {horizontal mode: the character |}.

Hi friends :)
My goodness @PauloCereda – you've been busy with those imgur links XD

4:08 PM
@SeanAllred <3

oh my god…
@PauloCereda <3

@barbarabeeton The rule resulting from | is also slightly shorter that the one resulting from --\relax--.
\setbox0=\hbox{|}\showthe\wd0
\setbox0=\hbox{--\relax--}\showthe\wd0
\bye

@HenriMenke -- hmmm. have to look further, i guess (the index has several more entries, but all the rest appear to deal with math); could this be one of dek's "white lies?

This reports: 10.00002pt and 10.00003pt.

@HenriMenke looks like 1sp, so it could be a rounding error... could you try \showthe\number\wd0 ? Oh I can do that right? :)

4:13 PM
By the way, this isn't exactly a TeX question, but it would probably use pdfbook: does anyone know how I would prepare treesmapsandtheorems.com/barriers/pdfs/TT&GL-1.01.pdf for distribution (other than printing and stapling the pages together at the corner)? I feel this has a lot of good things to say and I'd like to leave a few copies outside my office.

@SeanAllred I'm not sure it's possible, given there are pages that have text over the binding...

@yo' That was my original thought, too

@SeanAllred the only chance is an A4 landscape brochure, but that's not what you should want...

Besides, I don't have paper that big :(

@SeanAllred that's a technical problem. These do not count.

4:17 PM
@yo' ;)

@HenriMenke -- okay, more on the |. if you look at the code chart on texbook p.367, you will see that the | is in cell "7C'. in the chart on p.427, cmr10, that cell is occupied by the em dash. what you have in your code is two en dashes; if you actually look at the output, you will see that there's a hairline space between them. if you replace \hbox{--\relax--} by \hbox{---} you will get an identical value for the width. but since | isn't intended to be used in text, ...

@HenriMenke it's just a character, same as : or .

@HenriMenke Appendix F, Font Tables

@DavidCarlisle I wouldn't be that mean to @PauloCereda :)

4:32 PM
4 hours ago, by David Carlisle
@SeanAllred no! It's a kindness

@DavidCarlisle <3

@yo' So we can
@DavidCarlisle I meant the Unicode engine font business more generally: loading OT1 'for historical reasons' doesn't really make sense for them (even for pdfTeX it's at best a pain)
@barbarabeeton Shame that (at least from his TUGboat) the logging changes are not switchable: would be v. handy to have the classical trace for those of us used to it or using it for validation purposes.

4:48 PM
@JosephWright yes I think we should probably change for unicode engines

@JosephWright -- i agree that automated validation is a compelling reason for the "classical trace". make a suggestion, or if you'd like me to do that, let me know and i will. however, although the "new" logging changes are considerably different and more verbose, they might actually not be as inappropriate as you think; i'm interested in that too.

@barbarabeeton Didn't say it was 'inappropriate' :-) (Though I guess I'm quite happy 'thinking like TeX' when doing tracing)

@JosephWright -- maybe "inappropriate" isn't a good description. how about "disconcerting" or "distracting"? or maybe just that "they might not add as many discrepancies as you expect"?

Friends, a more serious note: should we try this new ad series? :)

@barbarabeeton I think 'not expecting \tracingall with beamer' :-)
@barbarabeeton I have a feeling a lot of it is for low-level stuff (paragraph breaking/alignment) which is probably handy for writing a TeX engine but not so useful finding macro level issues

4:55 PM
@JosephWright I tried that first time I saw a beamer question here after I joined the site...

@DavidCarlisle Indeed
@DavidCarlisle I doubt you'll make that mistake twice

@JosephWright no I just do it with tikz and expl3 now

@JosephWright -- is \tracingall really required for all the latex tests?

@barbarabeeton no

@barbarabeeton Not really the point: the moment you change how everything gets logged we have a bit of an issue
@barbarabeeton Box output is important, as is \show (or for expl3 more \showtokens, which you can do some clever formatting with)

4:58 PM
@JosephWright -- indeed. obviously, if the logging is changed, then there's for sure a massive effort to establish a new baseline, but what if the results are improved? might it not be worth it? (nonetheless, the request for a switch is justifiable.)
@DavidCarlisle -- i always suspected you were daft. i think this squares it.

Something really weird is happening. My sections are printing "SECTION" instead of the number and "page" instead of the page number.
What the heck is going on?

@barbarabeeton no for us they unlikely to be ever worth it as it is extra work forever, every new test that is added will always need a separate jsbox baseline, and there is no way of checking (as has happened many times with luatex) that teh new engine is not doing the right thing
@barbarabeeton If I had any lawyers they would be writing to you
@Alenanno there is an error in your document

@DavidCarlisle Clearly, but I didn't touch anything.
._.

@Alenanno users always say that, they are always not telling the entire truth:-)

@DavidCarlisle I'm serious, I didn't touch the preamble between earlier and now

5:02 PM
@barbarabeeton Once there is something we can test I'm happy to see if we can find a middle ground where the changes can be normalised away so we can compare with pdfTeX. That's doable with (u)pTeX and XeTeX and at least workable with LuaTeX (which after all deliberately breaks various things and accidentally alters quite a few more)

@DavidCarlisle -- oh, i'm rather sick of lawyers right now. i'm glad you don't have any. (and i never said there was anything wrong with being somewhat daft; makes people more interesting, as long as it's not directed in a destructive manner.)
@JosephWright -- okay. so what you really need is a protocol that you can compare directly with the other existing engines, which will continue to exist in their present form. forget i made the suggestion; wasn't thinking clearly enough.

@DavidCarlisle Instead of \thepage and \thesection, I had page and section ._.
AH.
I understand what happened.

@Alenanno you changed something? :-)

@DavidCarlisle I was working on an example in another document, did "find and replace" for \the, but I realized it did that to all my opened documents. So I did "UNDO" on my keyboard, and I thought it was fixed, but it only fixed that one document, ruining my other ones. Now it seems OK.

10 mins ago, by David Carlisle
@Alenanno users always say that, they are always not telling the entire truth:-)
:-)

5:13 PM
@DavidCarlisle I was telling the truth, I simply forgot what had happened.

@Alenanno same thing:-)

@DavidCarlisle Your wording makes it sound like I'm hiding it intentionally. lol

5:47 PM
shit, the new sidebar is wide.

Ug, the new version of chemmacros is missing a feature the old one had.

6:21 PM
@Canageek Yes?

6:42 PM
@JosephWright \chemsetup[iupac]{coord-use-hyphen=true} (or false) now gives an error.

7:20 PM
Does anyone know why \pageref{} is printing the plural "pages" instead of page?

8:11 PM
@Alenanno by default it will just print a number so ...

@DavidCarlisle ... I feel stupid.

@Alenanno well at least you don't have eminent members of the TeX community calling you "daft" :(

@DavidCarlisle Someone called you that?

@Alenanno search for that word a few lines up:-)
@JosephWright could be an interesting meeting tomorrow

@DavidCarlisle Indeed

8:16 PM
@DavidCarlisle Oh

1 hour later…
9:25 PM
What do you use when you add letters? Like in “ommitted”, for instance. ;-)`egreg 9 hours ago
@egreg you are so snarky :)

@barbarabeeton see what I mean ^^^ :-)

@SeanAllred :)
@SeanAllred The serious comment just after mine tells something about the site. ;-)