@PhelypeOleinik you have an email :D

@JosephWright @DavidCarlisle Any possibility to http://www.latex-project.org/lppl.txt=>https://www.latex-project.org/lppl.txt in .ins|dtx (git LaTeX repo)?

@PabloGonzálezL Huh?

@PabloGonzálezL The legal text is 'fixed', I don't thing we want to touch that file at all

@JosephWright I think he means to switch to "https"...

@AlanMunn Yes, that looks like it would work. Thanks.

Back to palindrome path (modulo commas)

@Rmano yaaaay

@Rmano ooh httpSQUIRRELS

Squirrels are good at protocols :)

^^^ Australian "squirrel" advocating for https everywhere. ;-) @PauloCereda

@JosephWright I think the xetex/dvipdfmx backend uses the wrong literal command for opacity. And the docu of l3opacity will have to adapt as they no longer use stacks. Also I wonder if they (and also dvips) should set the opacity to 1 in the reset. That is not perfect but probably better than doing nothing.

\DocumentMetadata{uncompress}
\documentclass{article}
\usepackage{transparent}

\ExplSyntaxOn
\cs_set_protected:Npn \__opacity_backend_select_aux:n #1
  {
    \tl_set:Nn \l__opacity_backend_fill_tl {#1}
    \tl_set:Nn \l__opacity_backend_stroke_tl {#1}
    \pdfmanagement_add:nnn { Page / Resources / ExtGState }
      { opacity #1 }
      { << /ca ~ #1 /CA ~ #1 >> }
    %\__kernel_backend_literal:n
    \__kernel_backend_literal_pdf:n %<--- new
      { /opacity #1 ~ gs }
    \group_insert_after:N \__opacity_backend_reset:

@Rmano Ah, right

@UlrikeFischer Could well be: when I backed out the stacks, I might have made an error

@Rmano oh no :)

@JosephWright so... a fix is in progress? :)

@PabloGonzálezL only occurs in base/doc/* I think, not in base/*.dtx or base/*.ins (@JosephWright)

@DavidCarlisle https:// quack

@DavidCarlisle Sounds about right: I remember fixing some of these

@PauloCereda spdy://breakfast

@DavidCarlisle scp://oh_no

@JosephWright should I open an issue?

@UlrikeFischer Please

@JosephWright I went with \protected, seemed more in the spirit of the PR:-) Le's see what the test suite does now, last night their were failures due to \protect going (understandable) but also several \special{} appearing in the ltmarks and footmisc tests that I didn't trace

lots of helicopters overhead today....

@DavidCarlisle I'd seen - still failing but the plan is sound :)

@DavidCarlisle We are going to see more and more robust commands move to \protected: I think it's one we really should talk about seriously

@JosephWright yes I didn't update the \protect tests in case we decided to go back to using it, more interested to see what happens to the diffs with \special

@JosephWright @DavidCarlisle the new code is doing something curious to the bookmarks:

@UlrikeFischer Example?

@JosephWright like this

\documentclass{article}
\usepackage{hyperref}
\begin{document}

\section{\MakeUppercase{abc grüße $a=b$}}

\end{document}

hyperref maps MakeUppercase to the expl3 function in pdfstringdef.

@UlrikeFischer I'm not seeing that with:

\documentclass{article}
\usepackage{hyperref}

\ExplSyntaxOn
\cs_gset:Npx \MakeUppercase #1 { \exp_not:c { MakeUppercase ~ } {#1} }
\cs_gset_protected:cpn { MakeUppercase ~ } #1
  { \text_uppercase:n{#1} }
\ExplSyntaxOff

\begin{document}


\section{\MakeUppercase{abc grüße $a=b$}}

\end{document}

@UlrikeFischer Ah, do you mean the UTF-8 business?

@JosephWright yes, the small ü and ß. Actually all non-ascii are not uppercased anymore.

@UlrikeFischer Hmm, but they are in the main text, so this must be a question of what is being passed to expl3: I'll see if I can work it out

@JosephWright perhaps it is a timing problem in \pdfstringdef.

@UlrikeFischer it's inputenc, you should use enctex

@DavidCarlisle boooo ;-)

@DavidCarlisle ooh boo

@UlrikeFischer I see what's happening. In the running text, the actives have the standard definitions, so they are OK. In the hyperref part, they are defined to expand to their byte definition, so \text_expand:n is converting the input to gr\80\374\80\337e $a=b$, and the case changer never sees the right thing

@JosephWright so the support for 8bit broke hyperref?

@UlrikeFischer Expanding active chars did, yes, seems so

@JosephWright so this is already broken in PL3?

@JosephWright we should drop 8bit ;-(

@DavidCarlisle No, because it relates to the expl3 change not the LaTeX2e one

@UlrikeFischer With the new \protected definitions of actives, couldn't hyperref just leave UTF-8 alone?

@JosephWright probably...

@JosephWright where? only for makeuppercase or in general?

@UlrikeFischer I was thinking in general, but I guess that doesn't work as this is how it deals with making escaped-byte output?

@UlrikeFischer, @DavidCarlisle Key is I think this is an issue at the hyperref end - we haven't 'messed up' in expl3 or in the proposed kernel change. Of course, we still have to fix it, but we can do that in a separate step

@JosephWright the expansion you are doing for the 8bit, can one somehow disable it locally (in pdfstringdef)? After all it worked before.

@UlrikeFischer If we know what form the hyperref redefinition takes, I could set up to avoid expanding them - I guess I'll need to dig through

If I redefine the command we added to expand the utf8 with \def\HyPsd@expand@utfvii{} it gets a bit better, one gets uppercase, only wrong chars .... I should check what happens in ansi.

\documentclass{article}
\usepackage{hyperref}
\ExplSyntaxOn
\cs_gset:Npx \MakeUppercase #1 { \exp_not:c { MakeUppercase ~ } {#1} }
\cs_gset_protected:cpn { MakeUppercase ~ } #1
  { \text_uppercase:n {#1} }
\makeatletter
\cs_gset:Npn \__text_expand_explicit:N #1
  {
    \token_if_cs:NTF #1
      { \__text_expand_exclude:N #1 }
      {
        \bool_lazy_and:nnTF
          { \token_if_active_p:N #1 }
          {
            ! \bool_lazy_any_p:n
              {
                { \token_if_protected_macro_p:N #1 }

@UlrikeFischer, @DavidCarlisle ^^^??? (with a bit of refinement)

\documentclass{article}
\usepackage{hyperref}
\ExplSyntaxOn
\cs_gset:Npx \MakeUppercase #1 { \exp_not:c { MakeUppercase ~ } {#1} }
\cs_gset_protected:cpn { MakeUppercase ~ } #1
  { \text_uppercase:n {#1} }
\makeatletter
\cs_generate_variant:Nn \tl_if_head_eq_meaning_p:nN { o }
\cs_gset:Npn \__text_expand_explicit:N #1
  {
    \token_if_cs:NTF #1
      { \__text_expand_exclude:N #1 }
      {
        \bool_lazy_and:nnTF
          { \token_if_active_p:N #1 }
          {
            ! \bool_lazy_any_p:n

Better ...

! Missing number, treated as zero.
<to be read again>
                   \__bool_@_0:
l.38 \section{\MakeUppercase{zzüße $a=b$}}

?

(I removed the makeuppercase def, and used a format built from the gh881 branch)

@DavidCarlisle I think you also removed the \makeatletter

@UlrikeFischer What you said ;)

@UlrikeFischer b@@gger

@JosephWright works

@DavidCarlisle Cool - it's a bit slower but I think should allow us to transition, and at some future date we could revisit

@JosephWright meanwhile I just ran l3build check in base locally, seemd to get to th end but ended

 Check failed with difference files
/usr/local/texlive/2022/bin/x86_64-cygwin/l3build:159: attempt to call a nil value (global 'ordered_filelist')

@DavidCarlisle Er, that's @MarcelKrüger's fault ;)

am i missing an l3build update here

@DavidCarlisle OK, le's see what @UlrikeFischer thinks

@JosephWright @UlrikeFischer a bunch of \protect have gone as expectd but any idea about this?

../build/test-config-OR/footmisc-002.pdftex.diff
::::::::::::::
*** ../../build/test-config-OR/footmisc-002.tlg Thu Jun 30 22:44:17 2022
--- ../../build/test-config-OR/footmisc-002.pdftex.log  Thu Jun 30 22:44:27 2022
***************
*** 79,89 ****
--- 79,91 ----
  ....\write1{\pp@pagectr{footnote}{1}{\theabspage }{\thepage }}
  ....\penalty 10000
  ....\penalty 10000
+ ....\special{}
  ....\hbox(8.14003+0.0)x4.48613
  .....\mathon
  .....\hbox(4.51111+0.0)x4.48613, shifted -3.62892
  ......\OT1/cmr/m/n/7 1

@DavidCarlisle I'm going to check main

@JosephWright oh is this git again, where is that file...

@JosephWright it's required/latex-lab, shouldn't that have been cleared from build if I run check just in base?

@JosephWright looks ok. Even \MakeUppercase{zzüße $a=b$ \texorpdfstring{xx}{🦆}} works in the bookmarks.

@DavidCarlisle that's me. This test loads the pdfmanagement and tagpdf-base. And some of the dummy tagging insert empty whatsits now. You remember the discussion on the team list?

@UlrikeFischer Great: I'll adjust

@DavidCarlisle Did you l3build clean?

@JosephWright boo

@UlrikeFischer, @DavidCarlisle We almost good to go?

@JosephWright did you move the new expl3 code to latex2e for the tests?

@UlrikeFischer Not yet - give me 10 minutes for the test suite to pass at the l3 end

@UlrikeFischer Done

@JosephWright OMG EXACTLY TEN MINUTES

@PauloCereda :)

@PauloCereda Unique time lord skills!

@samcarter ooh

@Skillmon ooh twitter.com/WholesomeMeme/status/1542857801181143041

@JosephWright I was working through test failures, but needed to stop ad do some Java:-)

@JosephWright us? tex.stackexchange.com/q/649555/1090

@UlrikeFischer so it already failed at PL3 or it changed since then?

@JosephWright of course not :-)

@DavidCarlisle well it couldn't have failed then. But I uploaded a pdfmanagement mid june, so it could fit.

@JosephWright I commited the ones I had done but probably can't do more until this evening, work calls

@JosephWright feel free to update if you want to merge before then. I looked over the diffs before

@DavidCarlisle I wondered

@DavidCarlisle :)

@DavidCarlisle OK

@DavidCarlisle I think I'll wait: lots to do here

@JosephWright I suspect an engine difference. The format with miktex starts with This is e-pTeX, Version 3.141592653-p4.0.0-220214-2.6 (utf8.euc), texlive has (utf8.sjis). But I don't know where one can set that.

@UlrikeFischer Ah, that could do it, yes

@PauloCereda -- I'm very fond of bunnies; there are several in our backyard wilderness. But we don't have a vegetable garden -- that would strike terror in the heart of a serious gardener.

@barbarabeeton ooh a bunny sanctum :)

@barbarabeeton in a pie?

@DavidCarlisle -- No pie, pizza or otherwise. We also get occasional possum visitors, and our former next-door neighbors once photographed a coyote hanging out there.

@DavidCarlisle oh no

@PauloCereda our cats also like them although they skip the pie making stage.

@JosephWright It seems to be this mysterious unicodedata failure again: github.com/MiKTeX/miktex/issues/1121

@JosephWright @UlrikeFischer ^ Green

@UlrikeFischer Oh goody

@DavidCarlisle Yay! expl3 release, merge, release?

@JosephWright that's the plan I think, yes.

@JosephWright there's a disproportionate number of \MakeUppercase in the test suite:-)

@DavidCarlisle I'll see if I can also sort @UlrikeFischer's issue before I do - this evening defo

@JosephWright where did we get to with \stock... that for -dev and next time?

@DavidCarlisle Yes: not a PL as it's not urgent (we generally don't add anything in a PL unless it's required to fix something)

@JosephWright ooh more stuff to break

Can I help break stuff?

@barbarabeeton :) yes, a very accurate representation of academic German.

@samcarter -- I particularly liked the picture of the little trailer tagalong. Almost an afterthought.

@barbarabeeton Aren't rabbits very destructive?

@PauloCereda Skype

Wanna hear a funny story from Overleaf? We accidentally broke projects that had \newcommand\beq{\begin{equation}} or alike :-) (Fix is on the way)

@yo' Ooops

@yo' ER, not such a bad thing

@JosephWright well, we also broke \newenvironment{foo}{\begin{bar}}{\end{bar}} and alike, which ... you know :)

@yo' how did you do it?

@yo' Ooopsy

@UlrikeFischer one of our parsers got into an infinite loop over any {\begin{foo}} or {\end{foo}}, causing the editor to eat itself out of memory.

@FaheemMitha -- Depends on where they are. You certainly wouldn't want them in your vegetable garden. But they do also eat "unwanted" vegetation including invasives, and provide food for animals higher on the food chain. Although these days it's not "politically correct", their skins are very warm, and were used for wraps for children by Native Americans. And these days, they're appreciated as pets.

@barbarabeeton Why isn't it "politically correct"?

@barbarabeeton our village allotments have literally hundreds of the things, definitely categorised under "pest" not "pet"

@FaheemMitha using real fur for clothing isn't acceptible in many contexts these days.

@DavidCarlisle Oh. Why?

@FaheemMitha -- Ever hear of PETA (People for Ethical Treatment of Animals)? They're opposed to the use of fur for any reason, and have all but ended the market for fur coats. They also oppose the eating of animal flesh, a practice that is also inimical to a number of religions.

@FaheemMitha google for animal rights :-)

@barbarabeeton I'm sure they make an exception for ducks

@DavidCarlisle -- I don't think so.

@DavidCarlisle Quack

@DavidCarlisle -- What's your opinion of Beatrix Potter?

@barbarabeeton she grew sheep to eat :-)

@barbarabeeton I've heard of them, yes.

There isn't much ethical treatment of people in this world, come to that.

@FaheemMitha -- And that's the real shame. I'm glad to have lived in more accepting and generous times. I am discouraged for the prospects of future generations.

@barbarabeeton that's a very filtered view of history:-) Things could always be better but may sections of society are far more accepted now than they were 20 or 30 years ago even.

@barbarabeeton Assuming there are future generations.

@DavidCarlisle Buildin l3kernel for release

@DavidCarlisle -- While I agree that's true, my childhood included the aftermath of WWII, when I was able to go wandering in the local parks and woods all by myself. Desegregation occurred while I was in public school, and in my school, went off without incident, and the "new" students were all welcome additions. I hit a snag in college, where the pure math faculty disliked women and applied mathematicians so much that I ended up in linguistics. So I've been fortunate, but not unique.

@FaheemMitha -- Many of them will have to learn to breathe underwater. Others will be dehydrated and roasted. I hope those in charge come to their senses and don't make things worse.

@barbarabeeton Sad to hear about the stupid math faculty. :(

@mickep -- Only pure math, and really, only one person. My applied math advisor was simply wonderful, as were the rest of the applied math and engineering faculties. It was really sad to leave them, but the faculties in German and linguistics were equally welcoming and inspiring. So I ended up with a wildly varied background, which proved unexpectedly useful in the long run.

@barbarabeeton To be clear, I think animals certainly need to be protected too. And their ecosystems as well.

Which, of course, are also our ecosystems.

@FaheemMitha -- Oh, I didn't think otherwise. If we don't protect everybody's ecosystems, the earth will go on without us, depleted in many ways, but it will go on. I think it's projected to go on for a bit over 4 billion years, until the sun expands to make the inner planets cinders

@DavidCarlisle :D

@JosephWright One more in doc \cs_argument_spec:N (Deprecated) in github.com/latex3/latex3/blob/…

@barbarabeeton It's nice to imagine the human race surviving to the far future, as the science fiction writers imagine. Sadly, the probability is heavily against it.

@FaheemMitha -- Even science fiction writers are less optimistic than they used to be.

@barbarabeeton I don't read much science fiction any more. I did once.

@PabloGonzálezL I'll have to do that for the next release: Ina has just installed today's update

@PabloGonzálezL Checked-in

@JosephWright following the texlive list? ;-)

@DavidCarlisle Oh yes :)

@JosephWright ctan.org/incoming

@DavidCarlisle That one was a bit late for Ina ...