@DavidCarlisle ooh Pokémons

@Skillmonlikestopanswers.xyz Sleep is generally considered to be a good thing. Within reason.

@JosephWright I see you were busy already, and from the texlive list it looks as if we have another day for another one ;-)

@PauloCereda so I followed the flow chart and concluded that twitter is the wrong channel and such things should go to stackexchange chat

@DavidCarlisle exactly :)

/quacks in despair

@DavidCarlisle Here you have two sources directly form the Italian yellow area (@egreg)

@CarLaTeX which is why I'm supplying important advice such as to eat more pineapple pizza to keep your vitamin levels up.

@DavidCarlisle More pineapple (not on pizza) :P

@CarLaTeX you can't leave the pizza out - the tomato and the cheese are important for the health too, pineapples alone are not enough.

@DavidCarlisle it's actually a good advice. :)

> This is the complete DNA of the Coronavirus (SARS-CoV-2). We are being attacked by a 8 kilobytes virus. Remember this when you hate on computers security.

twitter.com/ortegaalfredo/status/1233360209181061120

@UlrikeFischer Surely we can eat them both (separately)!

ooh 8KB

@UlrikeFischer ooh

@PauloCereda he's VIM user: we're safe :)

@yo' ooh :)

@UlrikeFischer miktex xelatex ocg question...

@DavidCarlisle texlive too for me.

@UlrikeFischer hmm could it be a pdf viewer thing? let me try a different viewer

@UlrikeFischer the pdf I get works in xpdf and acrobat pro 9

@DavidCarlisle Can you send me your pdf? I will look more closely in a few minutes - I have to sort out some chess business first - the opponent for the match on sunday just cancelled, they are coming from Heinsberg - the center of the virus panic here.

@UlrikeFischer OK

@UlrikeFischer yours is blue and made with a newer xdvipdfmx it seems

$ pdfinfo.exe test-utf8.pdf
Creator:        LaTeX with hyperref
Producer:       xdvipdfmx (20191114)
CreationDate:   Fri Feb 28 13:59:50 2020 GMT
Tagged:         no
UserProperties: no
Suspects:       no
Form:           none
JavaScript:     no
Pages:          1
Encrypted:      no
Page size:      612 x 792 pts (letter)
Page rot:       0
File size:      3996 bytes
Optimized:      no
PDF version:    1.5




$ pdfinfo cc661.pdf
Creator:        LaTeX with hyperref
Producer:       xdvipdfmx (20190503)

@DavidCarlisle yes, that is the problem. If I create the xdv with tl18 and then use the tl19 xdvipdfmx it gets blue. the tl18 xdvipdfmx is okay.

@egreg LOL

@UlrikeFischer cygwin texlive hasn't got a new xdvipdfmx yet as far as I can tell

@DavidCarlisle can you sent me an uncompressed pdf? Add

\usepackage{l3pdf}
\ExplSyntaxOn
\pdf_uncompress:
\ExplSyntaxOff

@UlrikeFischer done

@FaheemMitha deeply sorry: buzzfeednews.com/article/ryanhatesthis/…

@PauloCereda Not sure what you are referencing.

@DavidCarlisle @UlrikeFischer There is an issue with xdvipdfmx-2010-01-16. For the ocgcolorlinks-implementation (breakable ocg-colour links) in the ocgx2-package to work, It is necessary to place a msall rectangle

... as small rectangle at aan off-page location that is filled with the link colour. The current xdvipdfmx seems to clip it away from the PDF. Here is the relevant dvi-code:

@AlexG yes, it is obvious that the text rendering mode is set to fill when ocg-rectangles are drawn, but I don't quite understand why. Does the additional q/Q pair reset the text rendering mode? Is it not part of the graphic state?

1111: xxx 'pdf:code /OC/rm@oc0 BDC'
1136: xxx 'pdf:put @resources <</Properties <</rm@oc0 @pbs@obj1>>>>'
1194: xxx 'pdf:code -88888 -88888 99999 99999 re f'
1235: xxx 'pdf:code EMC'
1249: xxx 'pdf:code /OC/rm@oc1 BDC'
1274: xxx 'pdf:put @resources <</Properties <</rm@oc1 @pbs@obj0>>>>'
1332: xxx 'color push rgb 0 0 1'
1354: xxx 'pdf:code -88888 -88888 99999 99999 re f'
1395: xxx 'color pop'
1406: xxx 'pdf:code EMC'
1420: xxx 'pdf:code 0 Tr Q'
1437: xxx 'pdf:eann'

@AlexG why are you saying that it is a small rectangle? I didn't checked but simply assumed that -88888 -88888 99999 99999 is quite large, isn't it?

@UlrikeFischer Yes you are right. It is a large rectangle. I am about to compare the PDF page content streams. There seems to be an additional q/Q pair in the new xdvipdfmx-version, before BT and after ET:

NEW xdvipdfmx:
 q 1 0 0 1 72 720 cm 0 G 0 g 0 G 0 g q 7 Tr q BT /F1 9.9626 Tf 76.712 -62.765 Td[(link)]TJ /F2 .01 Tf 16.325 0 Td[( )]TJ ET Q /OC/rm@oc0 BDC -88888 -88888 99999 99999 re f EMC /OC/rm@oc1 BDC 0 0 1 RG 0 0 1 rg -88888 -88888 99999 99999 re f 0 G 0 g EMC 0 Tr Q 0 G 0 g q BT /F1 9.9626 Tf 231.133 -630.635 Td[(1)]TJ ET Q 0 G 0 g Q

OLD:
 q 1 0 0 1 72 720 cm q 7 Tr BT /F1 9.9626 Tf 76.712 -62.765 Td[(link)]TJ /F2 .01 Tf 16.325 0 Td[( )]TJ ET /OC/rm@oc0 BDC -88888 -88888 99999 99999 re f EMC /OC/rm@oc1 BDC 0 0 1 RG 0 0 1 rg -88888 -88888 99999 99999 re f 0 G 0 g EMC 0 Tr Q BT /F1 9.…

@AlexG @UlrikeFischer one of you raise with the xdvipdfmx tracker?

@AlexG yes, I have seen this. But I don't understand why it matters.

@UlrikeFischer @DavidCarlisle I am going to investigate this.

@DavidCarlisle sure.

@UlrikeFischer If I remove the q/Q pair in question, the result is ok again.

@AlexG I would say the problem is that the q Q pair "hides" the text. I can force something similar with pdflatex. As soon as the BT/ET pair is empty the page gets blue:

\documentclass{article}
\usepackage{l3pdf}
\ExplSyntaxOn
\pdf_uncompress:
\ExplSyntaxOff
\usepackage[colorlinks=true,urlcolor=blue]{hyperref}
\usepackage[ocgcolorlinks]{ocgx2}
\begin{document}\pagestyle{empty}\showoutput
abc
\pdfliteral direct{
q 7 Tr
%q
%BT
%/F8 9.9626 Tf 148.712 657.235 Td [(a)]TJ/F25 0.01 Tf 4.979 0 Td [(abc)]TJ
%ET
%Q
BT
/F8 9.9626 Tf 148.712 657.235 Td [()]TJ %with x is works
ET
/OC/rm@oc0 BDC
-88888 -88888 99999 99999 re f
EMC
/OC/rm@oc1 BDC
0 0 1 rg 0 0 1 RG
-88888 -88888 99999 99999 re f

@UlrikeFischer Thank you a lot! Thus, currently, glyph outlines (7 Tr) cannot be filled afterwards, and the colour spills the page :-(

@AlexG 7 Tr is clipping, so basically there is no clipping path right?

@UlrikeFischer Everything is that long ago. I need to reead the refs again.

@AlexG ;-) well I just read them, and 7 is clipping.

@UlrikeFischer Yes, clipping... I am going to make a minimal test file for a bug report.

@UlrikeFischer Minimal test file:

\special{pdf:code q 7 Tr} % add text path to clipping path
Text in rendering mode `for clipping'.
\special{pdf:code 0 0 1 RG 0 0 1 rg -88888 -88888 99999 99999 re f Q} % fill text clipping path

Normal Text (render mode `fill').

\bye

@AlexG texlive list or dvipdfmx list. Should both work, but on the first probably more people will see it and so at least know about the problem.

@JosephWright is \l_text_expand_exclude_tl meant to be extended by the user? I mean should I be able to add more commands here?

@UlrikeFischer Yes

@JosephWright then it doesn't work ;-(

@UlrikeFischer Example?

@JosephWright I'm just checking if I had some old code interfering.

@JosephWright here:

\documentclass{article}
\ExplSyntaxOn
\cs_new:Npn\mylabel #1 {\label{x#1}}
\tl_put_right:Nn \l_text_expand_exclude_tl {\mylabel}
\cs_set_eq:NN \textuppercase\text_uppercase:n
\ExplSyntaxOff
\begin{document}

\textuppercase{abc\label{blub}\mylabel{blub}}

\ref{blub}, \ref{xblub}

\end{document}

I get in the aux:

\newlabel{blub}{{}{1}}
\newlabel{xBLUB}{{}{1}}

@UlrikeFischer But you are also changing case: you need to provide an equivalent there too ...

@JosephWright ? what does that mean?

@UlrikeFischer \tl_put_right:Nn \l_text_case_exclude_arg_tl { \mylabel}

@JosephWright that's not in the documentation ;-). Do one need both? (only the second seems to work fine).

\documentclass{article}
\ExplSyntaxOn
\cs_new:Npn\fooa #1 {\label{x#1}}
\tl_put_right:Nn \l_text_case_exclude_arg_tl {\fooa}
%
\cs_new:Npn\foob #1 {\label{y#1}}
\tl_put_right:Nn \l_text_case_exclude_arg_tl {\foob}
\tl_put_right:Nn \l_text_expand_exclude_tl {\foob}
\cs_set_eq:NN \textuppercase\text_uppercase:n
\ExplSyntaxOff
\begin{document}

\edef\x{\textuppercase{abc\fooa{blub}\foob{blub}}}
\show\x

\end{document}

@UlrikeFischer ^^^ \l_text_expand_exclude_tl makes the macro itself not be expanded, and \l_text_case_exclude_arg_tl makes its argument not be case-changed

@PhelypeOleinik ah, ok. But it can handle only one argument right? That's why here only \food works as it expands first to \label which has one argument:

\documentclass{article}
\ExplSyntaxOn
\cs_set_eq:NN \textuppercase\text_uppercase:n
\cs_new:Npn\fooc #1#2{\label{#1x#2}}
\tl_put_right:Nn \l_text_case_exclude_arg_tl {\fooc}
\tl_put_right:Nn \l_text_expand_exclude_tl {\fooc}
\cs_new:Npn\food #1 #2 {\label{#1z#2}}
\tl_put_right:Nn \l_text_case_exclude_arg_tl {\food}

\ExplSyntaxOff
\begin{document}

\edef\x{\textuppercase{abc\fooc{blub}{bla}\food{blub}{bla}}}
\show\x

\end{document}

@UlrikeFischer That's right

@UlrikeFischer I guess it could be expanded to add \l_text_case_exclude_two_args_tl...

Though most document commands like that have only one argument, I think...

@PhelypeOleinik well one can hide more complicated stuff, but it is interesting that it still expands inside the argument:

\documentclass{article}
\ExplSyntaxOn
\cs_set_eq:NN \textuppercase\text_uppercase:n
\cs_new:Npn\fooc #1#2{\label{#1x#2}}
\cs_new_protected:Npn \hide #1 {#1}
\tl_put_right:Nn \l_text_case_exclude_arg_tl {\hide}

\ExplSyntaxOff
\begin{document}

\edef\x{\textuppercase{abc\hide{\fooc{blub}{bla}}}}
\show\x

\end{document}

@UlrikeFischer I think it's because that's only for pretecting against case-changing, not expansion

@UlrikeFischer Yes, \l_text_case_exclude_two_args_tl only protects against case-changing, not against expansion