« first day (3775 days earlier)      last day (1142 days later) » 

8:29 AM
@PauloCereda (sorry, that I'm evil)
 
@Skillmon <3 we will apply your suggestion for version 8 :)
 
@PauloCereda why 8? Thought the next major release (that can include breaking changes, since it's a major release) would be 7.
 
@Skillmon Scheduling. :) We will see what can be done. :)
 
@PauloCereda You're the maintainers :) (ooh, and original authors)
 
@Skillmon ooh
> Linux Mint Developers Will Force Updates on Users Like Microsoft Does with Windows 10
@Plergux quack
 
8:43 AM
@PauloCereda Baaa!
 
@Plergux ooh hi sheep person
 
@PauloCereda Oooh! hi duck person? How are things in "Brazzaland"? :p
 
:57212206
 
@Plergux oh same old same old, blood, devastation, death, war and horror. :) What's next, a man who does gardening? :)
@Plergux and how are things in the land of ice? :)
 
@PauloCereda lol
 
8:48 AM
@UlrikeFischer For xetex.def, I think we need \def\define@color@named#1#2{\expandafter\edef\csname col@#1\endcsname{#2}}
 
@JosephWright sorry I had no time at all yesterday, need me to do any color checks? (I didn't really look at the issue at all) is it just in the def files or do you need to adjust color.sty as well?
 
@DavidCarlisle I think it's just the .def
@DavidCarlisle Currently, named colours are saved like dvips but that doesn't wok if one is using pdfTeX-style specials
 
@PauloCereda shaky. Literally :p Since we had the big one last wednesday (around 5.4) there have been literally hundreds of quakes per day, many over 5, in the same area. So quite possibly an eruption could happen.
 
@Plergux oh no
 
@JosephWright so xetex and dvipdfm presumably
 
8:54 AM
@Plergux I am no seismologist, but I suspect the recent changes in l3color might have something do to with these earthquakes...
 
@PauloCereda Lol. So all the l3color people jumping up and down in frustration are causing the plates to shift? :p
 
@Plergux ooh more theories
 
@DavidCarlisle Yes
@PauloCereda My TUG talk intro :)
 
@JosephWright ooh :)
@JosephWright I can prepare a better image :)
 
@PauloCereda Probably use the hummingbird :)
 
9:06 AM
@JosephWright ooh
Wants a photo of Fubá as member team? :)
 
@PauloCereda Sounds good
 
@JosephWright while looking at this: there is something wrong when using the dvipdfmx backend with the pdfmanagement:
\RequirePackage{pdfmanagement-testphase}
\DeclareDocumentMetadata{uncompress,backend=dvipdfmx}

\documentclass[dvipdfmx]{article}
\usepackage{graphicx}
\usepackage{color}
\begin{document}
\textcolor{blue}{NavyBlue}
\end{document}
gives a warning:
test-utf8.dvi -> test-utf8.pdf
[1
dvipdfmx:warning: Stadk ID=2 already initialized?
dvipdfmx:warning: Interpreting special command pdfcolorstackinit (compat) failed.
dvipdfmx:warning: >> at page="1" position="(72, 720)" (in PDF)
dvipdfmx:warning: >> xxx "pdfcolorstackinit 2 page direct (/opacity1 gs)"
dvipdfmx:warning: >> Reading special command stopped around >>page direct (/opacity1 gs)<<
]
3564 bytes written
 
@UlrikeFischer Hmm, looks like they are using the same stack
@UlrikeFischer I've applied your source3 change: better but not 100% yet
 
@JosephWright You could also blame it on the butterfly effect. :p
 
9:23 AM
@JosephWright yes, it was a bit late yesterday to really think it through. Probably it would be better to allow maketitle to create chapters in the implementation part too, perhaps even with the same numbers as in the main part, but with some prefix.
 
@UlrikeFischer Something like that, yes, but I'm happy to fix in increments :)
 
@Plergux ooh
 
@JosephWright even more if you have to do a full compilation circle for 1000 pages to check what happened ;-)
 
@UlrikeFischer Nott that bad here on my MacBook Pro - one reason for having a fast machine
 
@JosephWright my desktop pc is quite fast, but the laptop takes its time. I should try to get a new one in the next time, I had to delete the doc folder in texlive 2020 to cram 2021 in ;-)
 
9:30 AM
@UlrikeFischer That's why i got a new one!
 
@JosephWright we are grand, aren't we? Off to play the grand piano, aren't we? No more buttery scones? :)
 
@UlrikeFischer it probably takes you less time than making source2e.tex on a sun3...
 
@DavidCarlisle :)
 
9:51 AM
@UlrikeFischer I just bought more external hard drives. My laptop now has two 1TB ones and one 65G for smaller files (and since I have Windows 10 I can install apps on the external drives!).
 
/dev/sda2       465G   79G  385G  17% /
/hides in shame
 
@PauloCereda oooh! 80GB of space! Quick! Download some games! :p
 
@Plergux actually, 385G :)
 
@Plergux or run a beamer document with \tracingall
4
 
@DavidCarlisle ooh
@DavidCarlisle beamer + TikZ + \tracingall :)
 
10:00 AM
@PauloCereda ah! I got it the wrong way around! How do you only use 17% of your space? I'm lucky if I have that left on any of my hard drives XD
@DavidCarlisle oooh... tempted...
 
@Plergux ooh :)
 
@JosephWright it improves the situation as one can get the color, but there seems still to be a step missing. But as a side remark: with the patch I insert in color with the pdfmanagement, the l3color is setup correctly and works out of the box:
\RequirePackage{pdfmanagement-testphase}
\DeclareDocumentMetadata{uncompress}

\documentclass{article}
\usepackage{graphicx}
\usepackage[dvipsnames,usenames]{color}
\ExplSyntaxOn
\color_show:n{NavyBlue}
\ExplSyntaxOff
\begin{document}
\ExplSyntaxOn
{\color_select:n{NavyBlue} Navyblue}
\ExplSyntaxOff
\textcolor{NavyBlue}{NavyBlue}
\end{document}
 
@UlrikeFischer I'll keep working on it
@UlrikeFischer I've tracked down the stack number issue: fix at both l3 and graphics-def ends needed
 
10:19 AM
@JosephWright I know that in math mode, there are reasons why one might prefer \begingroup...\endgroup to \bgroup...\egroup, but for general notions of scope setting, is there a preference?
 
@StevenB.Segletes No
 
@JosephWright Thank you. Helpful.
 
ooh FMiTeX
 
@UlrikeFischer Fixed here: release upcoming
@UlrikeFischer,@DavidCarlisle Fixes for l3backend and graphics-def off to CTAN
 
10:45 AM
@JosephWright did you update the dates? I still see 2021-02-18 in l3backend-xetex.def
 
@UlrikeFischer Hadn't pushed :0
@UlrikeFischer Try now
 
@JosephWright better, but I can't install the new xetex.def, if I call l3build install it does nothing ...
@JosephWright I have now the new xetex.def and l3backend, but I still don't get navyblue. Do I need something else?
 
11:06 AM
@UlrikeFischer No?
@UlrikeFischer No, should work with
\documentclass{article}
\usepackage[dvipsnames,usenames]{color}
\begin{document}
\textcolor{NavyBlue}{NavyBlue}
\end{document}
@UlrikeFischer Only needs the corrected xetex.def: the other fixes are about the excess/incorrect colour stacks
 
@JosephWright permission to come up with a title artwork for your talk?
 
@PauloCereda Ack
@UlrikeFischer To fix the colour stack numbering error one also needs updated l3kernel
 
11:32 AM
@UlrikeFischer I think we are getting close to having this working
@DavidCarlisle, @UlrikeFischer Does a talk about l3color sound OK for TUG? I was thinking covering transparency, colour spaces, etc., on top of the basics
 
@JosephWright sounds like a topic
 
@JosephWright I don't get the color, and if I uncompress I see 0 G 0 g 0 G convert@color@ {NavyBlue} BT in the stream. xetex.def 2021/03/02 v5.0j Graphics/color driver for xetex
@JosephWright sounds good.
 
@PauloCereda and bye bye, Linux Mint, you were an often used beginners distro.
 
@UlrikeFischer Oh, I forgot to push some of the changes - try re-downloading now
 
@JosephWright success!
 
12:08 PM
@UlrikeFischer I think we are basically sorted for dvipdfmx, and using multiple stacks is a lot easier
@UlrikeFischer, @DavidCarlisle I might raise the footnote-colour business: we could update the kernel to do that for pdfTeX/LuaTeX/dvipdfmx, meaning only dvips and dvisvgm would have an issue
What I need is a PostScript wizard to work out how to implement a colour stack manually ...
 
@PauloCereda How about TMITex? :p
 
12:41 PM
@JosephWright or extend the stack code in dvips, or do you think it needs to be in PS in the end?
 
@DavidCarlisle In dvips would be much easier!
 
@JosephWright how's your C ....
 
@DavidCarlisle Part of the issue is the interface: I guess we need to make an entirely new one, as in dvipdfmx
@DavidCarlisle \special{colorstack <n> <value>} perhaps?
 
@JosephWright Just love the motivation for the design of the color commands in dvips:
 *  The motivating idea:  we want to be able to process 5,000 page
 *  documents using lots of colors on each page on a 640K IBM PC
 *  relatively quickly.
 
@DavidCarlisle Ah, that again
@DavidCarlisle I wonder if Tom R. might be convinced to look at it? Would be better than me trying
 
12:44 PM
That 640K memory segment limit is such a pain to work round.
 
@DavidCarlisle Expanded memory ;)
 
@JosephWright He made some updates for DEK last year didn't he?
 
@DavidCarlisle Yes, exactly
 
color.c is really surprisingly small (although you may need changes elewhere)
 
@DavidCarlisle I'll take a look then
 
12:49 PM
Is there written down a spec of what's missing?
}
/*
 *   This routine is a bit magic.  It looks up the page in the current
 *   hash table.  If the page is already entered in the hash table, then
 *   it restores the color to what that page had, and sets the last
 *   color.  This will occur if this is not the first time that this
 *   page has been encountered.
 *
 *   If, on the other hand, this is the first time that the page has
 *   been encountered, then it will create a new hash entry and copy the
 *   current color information into it.  Since we can only encounter a
 
@DavidCarlisle Ideally we'd be able to handle multiple stacks: one is needed for transparency, one for footnotes. Transparency in PostScript is of course 'interesting', so that makes it harder (and less motivated): in the PDF routes, we already have the ability to deal with 'graphic state' more generally using a 'colour' stack
 
@JosephWright imho it would be worth a try to get someone in. There is also the xform question, if @AlexG is right and one need different variants depending on if it is used as appearance or in the stream and additionally need to detect an unknown landscape setting it would be a rather a pain to add sensible support.
 
1:05 PM
@UlrikeFischer Isn't that at the PDF end though? So has to be pdfmark?
 
@JosephWright I have no idea what is where needed ;-). I only see that Alex is passing an argument to tell its code if the xobject should be used in the stream or as appearance, and this looks quite wrong as interface (and clashes with all the other backends).
 
2:02 PM
@Plergux :)
 
2:25 PM
@UlrikeFischer I'll take a look at that code later and see whether a change in dvips could help
 
3:03 PM
@JosephWright artwork is done
@JosephWright promise me you will use it?
 
@PauloCereda Of course
 
@JosephWright May I also offer a title suggestion? :)
 
@PauloCereda :)
 
@JosephWright You have no idea how serious this is. :)
 
yo'
3:25 PM
Tested COVID-neg today, quite relieved.
6
 
@yo' happy to hear it!
 
@UlrikeFischer I think I need to work on the PostScript, although I also need to read the PDF stuff - doubtless we need to manage this at the macro layer
 
@yo' that's really good news!
 
@yo' yay
 
3:51 PM
@yo' -- That's really good news! But it's possible to become seriously ill with something else, so take care of yourself. (Gordon spent a week in the hospital last month with asthma-induced pneumonia; they took good care of him, and he's all better now, thank goodness.)
 
 
3 hours later…
6:47 PM
user image
4
Fluffiness of the day :p
2
 
@Plergux ooh
 
7:01 PM
> Epic Games has bought Tonic Games Group, the company that owns Fall Guys: Ultimate Knockout studio Mediatonic.
Epic Games is the company which produces Fortnite...
 
@Plergux -- I love their little black socks above white (well, perhaps slightly muddy) "tennis shoes".
 
@PauloCereda lovely... :|
@barbarabeeton yeah, they're what we would call in Icelandic a "krúttbomba" (cutie-bomb :p) :p
 
7:21 PM
@Plergux wow, what a nice version of sheep!
 
@Rmano they look like a metal rock band :)
 
@Rmano It is! It's like somebody took a llama, a sheep, and a cow and smushed them together. :p
@PauloCereda They're baaaaad to the bone :p
 
@Plergux LOL
 
@Plergux ...and twisted their horns a bit, it seems
@Plergux we should ask @samcarter_waiting_for_summer to add a tikzling...
 
@Rmano :D
 
7:28 PM
What do I need to learn today?
 
@LaTeXereXeTaL It is a myth that a duck's quack won't echo. This has been conclusively disproved through different scientific acoustic tests, and was even featured as "busted" on an episode of the Discovery Channel show Mythbusters.
ooh a fact
@LaTeXereXeTaL and hi, we miss you. <3
 
I didn't know anyone thought duck quacks didn't echo, but still good to know. I've been quite busy teaching and working on my package update. I've learned a lot about using GitHub too.
 
@LaTeXereXeTaL that's good :)
 
@PauloCereda Trying to enter a new culture is sometimes difficult.
 
@PauloCereda hmph, how come I mention GH and you tell me off and say I should be using GL, but @LaTeXereXeTaL says he's using GH and you make encouraging quacks?
  github-0122 (24/444)
          --> failed

  github-0124 (25/444)
          --> failed

  github-0177 (26/444)
  github-0186 (27/444)
  github-0192 (28/444)
  github-0198 (29/444)
  github-0199 (30/444)
          --> failed
speaking of github :(
ah back to texlive 2020...
 
7:43 PM
@Ulrike, How can I add a single object reference of a Widget Annot to the Fields array in AcroForm, using the new l3pdfmanagement interface?
 
@AlexG I assume you mean this: \pdfmanagement_add:nnx{Catalog/AcroForm}{Fields}{<obj ref>}
 
yo'
@barbarabeeton yeah, all the other nasty germs obviously don't disapprear when there's a new one -- why would they? I'm glad to hear that you're OK, greetings to Gordon!
 
GL = GitLab?
 
@LaTeXereXeTaL yes
 
8:29 PM
@UlrikeFischer Yes, this one thank you. I wasn't quite sure whether <obj ref> is the object ref of the whole array or of a single element therein.
 
@AlexG it is additive, so if more than one package adds a field they are combined.
 
@UlrikeFischer This is the way it should be ;)
I am beginning to replace my old commands that do the same.
 
 
2 hours later…
10:17 PM
@DavidCarlisle oh
@DavidCarlisle /encouraging quack <3
And GL > GH :)
 
@PauloCereda dinner
 
@DavidCarlisle oh no
 

« first day (3775 days earlier)      last day (1142 days later) »