@UlrikeFischer proud owner of a new git repo

@DavidCarlisle Which one?

@PhelypeOleinik @UlrikeFischer this one github.com/davidcarlisle/graphics-pln

@UlrikeFischer most of the code in it is less than new:-)

@DavidCarlisle 20 commits already. You are fast! :-)

@DavidCarlisle oh you splitted. Is the issue new?

@yo' I realised that the drawing (not the transparent background) of the attached image don't have the same color when I drag it to the document. See:

Do you know how to change it?

@UlrikeFischer define "new" I think the code that is wrong was added in 1996, which doesn't seem that long ago.

@manooooh just edit it in a bitmap editor eg put a transparent panel in front not a lot of point doing that in tex

Oh actually with decodearray={0.2 0.3 1 0 1 0.8} I could change the draw coloring to a light blue. But I don't understand how 0.2 0.3 1 0 1 0.8 works

@DavidCarlisle thank you David! Unfortunately, I am not a designer, so I don't know how to complete your steps

@AlanMunn -- Reminds me of my all-time favorite cartoon: A robed individual carrying a sign saying "It's just going to continue and continue". (The usual sign would say "The end is nigh.")

TeX Live 2020 pre-testing has started

@JosephWright oh no

@JosephWright but not quite yet for cygwin

@DavidCarlisle ooh

@PauloCereda I've got the sources and have already compiled my own, but I still feel I'm missing out as the cygwin binaries aren't in the pretest yet:-)

@DavidCarlisle :)

@PauloCereda although, to be different perhaps I may install the linux pretest under wsl instead

@DavidCarlisle hmmm

@PauloCereda wsl is the future I suspect

@DavidCarlisle ooh welcome to the world of tomorrow

@PauloCereda Where's Maggie Philbin when you need her.

@DavidCarlisle ooh

@DavidCarlisle or the Futurama guy

@UlrikeFischer The v3 code works entirely differently: my plan is to reset the series, etc., to the default unless specifically asked not to. But that means the options don't really map between v2 and v3. I'm also seeing an oddity in the above, which I need to track down

@UlrikeFischer I know I need to finish v3 ASAP

@JosephWright ooh versions

@JosephWright yes I know that it is different and I didn't try to suggest some concrete code. I just wanted to mentioned the new "force a fontseries" command.

\sffamily abc \fontseriesforce{m}\selectfont abc
\qty[mode = text,
  reset-text-series = false,
  reset-text-family = false]{1}{\km} abc

@UlrikeFischer ^^ In v3, gives the result I'd expect

@JosephWright what would it give if the user uses only \fontseries{m}? still m?

@UlrikeFischer Yes, because I've set rest-font-series to false

@UlrikeFischer Possibly I'll need to look at Frank's new axis, of course

\bool_if:NT \l_@@_text_family_bool
  {
    \fontfamily { \familydefault }
    \selectfont
  }
\bool_if:NT \l_@@_text_series_bool
  {
    \fontseries { \seriesdefault }
    \selectfont
  }
\bool_if:NT \l_@@_text_shape_bool
  {
    \fontshape { \shapedefault }
    \selectfont
  }

@UlrikeFischer ^^^ The key code

@JosephWright couldn't you use something like \l_siunitx_seriesdefault? Which could then be set by the user if wanted?

@UlrikeFischer Well, I could, but I'm still exploring the font stuff; also, there is a reasonable starting point that \seriesdefault has to be the correct position as it's, like, the global default

@JosephWright it is certainly a reasonable default. But I believe that there is always a user who has special wishes and asks e.g. how to set the default to bold zapf chancery ...

In other news:

This is pdfTeX, Version 3.14159265-2.6-1.40.21 (TeX Live 2020/W32TeX)

@UlrikeFischer yay! I am finding the instruction on installing pretest a little confusing --- Am I right that if I download the install-tl as indicated is then a matter of doing the same as a "portable" normal installation?

@UlrikeFischer Like I said, a work still in progress, and certainly adjustable even if I don't do it now

@JosephWright yes, should be easy.

@DavidCarlisle TeXLIve on WSL works much faster than under "cygwin" or the windows binaries themselves and with the official WSL2 output this will be much better :)

@PauloCereda Including "arara" under WSL :)

@JosephWright Hey, have you had time to take a look at l3build?

@PabloGonzálezL sure but we have a massive cygwin infrastructure here (including X) so moving over to wsl will take time.

@DavidCarlisle Yes, X is all one of them (for now)... of course there are some advances but not to the extent :(

@DavidCarlisle If you have some time left you can take a look at (tex.stackexchange.com/q/530890/7832) it's a bit silly to doubt, but I can't find the answer :(

@PabloGonzálezL I'll email

@PabloGonzálezL Give me an hour or so

@JosephWright Thanks, time is not a problem :) ... I have one more morning to have fun installing TeXLive 2020 :-)

@UlrikeFischer Here's a much better one

> pdftex --version
pdfTeX 3.14159265-2.6-1.40.21 (TeX Live 2020)
kpathsea version 6.3.2
Copyright 2020 Han The Thanh (pdfTeX) et al.
There is NO warranty.  Redistribution of this software is
covered by the terms of both the pdfTeX copyright and
the Lesser GNU General Public License.
For more information about these matters, see the file
named COPYING and the pdfTeX source.
Primary author of pdfTeX: Han The Thanh (pdfTeX) et al.
Compiled with libpng 1.6.37; using libpng 1.6.37
Compiled with zlib 1.2.11; using zlib 1.2.11

@egreg finally you catch up:

@DavidCarlisle I'm even ahead of you: pdfTeX 3.14159265-2.6-1.40.21

@egreg I have that as well (just installed the pretest but...)

@DavidCarlisle Excuses

@egreg and I got the tick for extracting numbers from \jobname

@DavidCarlisle we are driving through the Eifel ...

@UlrikeFischer it should be safe now. British teenagers have been repatriated.

@DavidCarlisle Leaving behind a defined macro? No way.

@egreg ? you think I should undefine, why?

@DavidCarlisle Because you used a “normal” name.

@egreg I could have used \zzzz you loaded xstring and that defined lots of commands with normal names but you don't undefine them all.

@PabloGonzálezL :)

running the 2e base test suite in the pretest.....

A palindrome ^^

@PhelypeOleinik wow!

@PhelypeOleinik :-)

@PhelypeOleinik ooh

@DavidCarlisle I'm tryinbg L3

  Check failed with difference files
  - ../build/test/tlb-callbacks-001.luatex.diff
  - ../build/test/tlb-extraprimitives-001.luatex.diff

@JosephWright the usual :-) ^^^^^

@PhelypeOleinik -- That one's special!

   dvifeedback
  dvivariable
  eTeXVersion
+ eTeXglueshrinkorder
+ eTeXgluestretchorder
  eTeXminorversion
  eTeXrevision
  eTeXversion

and

      *** ../build/test/tlb-callbacks-001.luatex.tlg    2020-03-03 15:32:49.229812600 +0000
--- ../build/test/tlb-callbacks-001.luatex.log  2020-03-03 15:32:51.335200600 +0000
***************
*** 1,5 ****
  This is a generated file for the LaTeX2e validation system.
  Don't change this file in any respect.
   LuaTeX Callback test
! + glyph_info
! + page_order_index
--- 1,6 ----
  This is a generated file for the LaTeX2e validation system.
  Don't change this file in any respect.
   LuaTeX Callback test

@DavidCarlisle Right, first order of business is to think about \input: we've got expl3 code that 'knows'about LuaTeX, but I should extend in some way for pdfTeX, etc., in TL 2020. I guess use \pdfmajorversion as a marker ...

@DavidCarlisle Ah yes, update ltluatex ...

@JosephWright and do we use that in 2e... one for team list I guess.

@DavidCarlisle Yup, I'll draft

@JosephWright do we know if MiKTeX got the input change already?

@UlrikeFischer we could ask @UlrikeFischer, try this:

\input{x}
\bye

in plain tex

@UlrikeFischer 'Yes'

@UlrikeFischer if it inputs texmf-dist/tex/latex/tools/x.tex you have the new \input

@JosephWright it would of course be possible to use a test as above and resurrect the old dirchk tests at format creation to test the input behaviour rather than rely on specific primitives being added at the same time, but perhaps not..

@DavidCarlisle Ulrike is sitting in a car and can't try now ;-)

@UlrikeFischer In a mail from Karl (about a month ago, I think) he said Christian had already added the \input{...} code to MiKTeX. I don't know if that means that users already have it

@UlrikeFischer no tex on your phone? disgraceful

@PhelypeOleinik @UlrikeFischer it is the headline news at github.com/MiKTeX/miktex/blob/next/CHANGELOG.md

@DavidCarlisle Oh :-)

@DavidCarlisle True, also for expl3, but I don't think we really want to go that way do we?

@PhelypeOleinik Wroking here

@DavidCarlisle, @UlrikeFischer Can you 'see' \current(x)spacingmode in upTeX? I can't ...

tug.org/texlive/pretest.html says they should be there

@JosephWright euptex has it

@DavidCarlisle Yes, tracked it down ...

@DavidCarlisle Had a chance to lopok at the keyval PR?

@JosephWright not really (not when I was awake)

@PhelypeOleinik Is the xparse stuff ready? I'm thinking it and the keyval material can go to CTAN now ...

@DavidCarlisle :)

@JosephWright Both seem good to go to me. I'll merge the xparse thing

@PhelypeOleinik OK, I think we'll merge the keyval one too, and I'll do a release (TL'20 pretest/MiKTeX-as-always)

@JosephWright yes seems good to me, I didn't see anything wrong with it just haven't had time to look as hard as I might have done

@JosephWright I spend more time trying to figure git out than doing actual changes :-)

@JosephWright Do I delete the cs-delimited-args branch or there's something else to do with merged branches?

@PhelypeOleinik I normally put a tag on them (like archived/cs-delimited-args) and then I delete the braches once they are merged, both locally and remotely. But I am not really a git expert, just trying to survive...

I have a mini-script to do that

@Rmano Well, I just spent 15 or so minutes figuring out how to merge a branch without messing the entire repo. I'm in the "trying to survive" boat as well ;-)

@Rmano Yes, the branch is already merged in master. So I should rename it and then delete?

@PhelypeOleinik This is what I do (found the info somewhere):

Substitute a branch with a tag, delete branch here an in origin

     git tag archive/<branchname> <branchname>
     git branch -d <branchname>
     git push --delete origin mybranch
     # In a command:
     B=fixdescription && git tag archive/"$B" "$B" && git branch -d "$B" && git push --delete origin "$B"

To restore the branch some time later:

     git checkout -b <branchname> archive/<branchname>

To list archive tags:

     git tag -l archive/\*

Supposing that the remote branch is the "origin" one (in my projects "origin" is my private fork; I branch there and then do a PR to the main repo).

@Rmano I'll try this. Thanks :D

@Rmano Hm... With git tag then I create an entry here, don't I?

@PhelypeOleinik I don't think --- the tag is local to my repo here, it is not even propagated on origin. At least I think...

@PhelypeOleinik there are two kinds of tags, one is signed and is an object of the repo, the other one is local only. The simple git tag is local only...

Again --- I hope. Please double check...

Ok, seems that you have to push the tag explicitly to have it in the repo: drupixels.com/blog/…

Or better git-scm.com/book/en/v2/Git-Basics-Tagging

@Rmano Hm... Tagging seems to be for marking a point in history (in latex3 we use for releases). It isn't of much use (in this case) if I only get to have a local version of it

@Rmano But thank you for the help!

@PhelypeOleinik stop breaking stuff. :)

@PauloCereda Oh no! :-)

@PhelypeOleinik oh no

@PauloCereda I just hope I didn't mess anything up in the implementation of \input{...}. The amount of angry users would be much higher

@PhelypeOleinik :)

@PhelypeOleinik Should be easy on GitHub

@PhelypeOleinik Please

@PhelypeOleinik Plese don't tag or at least please don't push a tag

@JosephWright Done already. After discussing with @Rmano I noticed that keeping a branch in our case is not really useful

@PhelypeOleinik Normally, for non-risky changes, I rebase so we get a linear history: it's no big issue, but we very rarely need to back-out merges

@JosephWright I imagined that was the case. After messing up here and there I learned to tidy up my local copy and only then push to remote, so I hope I won't make anything that requires that

@PhelypeOleinik a few million people sending you cross emails, it won't be so bad.

@DavidCarlisle ooh

@DavidCarlisle probably "tens of millions" actually.

@yo' but some of them might be too polite to hassle @PhelypeOleinik

@DavidCarlisle I don't believe so many of them come from the UK.

@UlrikeFischer should we risk a zref release?

@DavidCarlisle That's a relief ;-)

@PhelypeOleinik I first became aware of how many latex users there were when we put out an early release of 2e with \fbox broken, no one ever mails you to say things work....

@PhelypeOleinik I'm going to do a of release expl3 in a bit; have to build manually as the deployment didn't trigger ...

@DavidCarlisle :-)

@JosephWright Something broken?

@PhelypeOleinik No, but I want to get the keyval stuff and the xparse change into TL'20 pre-testing ASAP

@JosephWright I meant the deployment that didn't trigger

@PhelypeOleinik Oh, I think it's a Travis-CI issue: it did the build for the check-in but not the tag

@JosephWright did you merge without @DavidCarlisle's review? Also, yay, I now officially contributed to L3 :)

@Rmano correct, tags have to be pushed explicitly, using git push <remote> <tag-name>.

@Rmano but, iirc, there are two types of tags, one with and one without a tag-message, but that's the only difference there is, afaik.

@Skillmonlikestopanswers.xyz Er, no: git push --tags

@Skillmonlikestopanswers.xyz my in depth review was here, sorry:

@Skillmonlikestopanswers.xyz Yes, it's in and off to CTAN

@Skillmonlikestopanswers.xyz Congratulations, you are just in time to take the blame for anything that goes wrong in tl2020.

@JosephWright which still is explicitly pushing tags :)

@Skillmonlikestopanswers.xyz I meant that one doesn't normally push 'a tag', rather one pushes 'the tags as a set'

@DavidCarlisle I only take blame for \keyval_parse:NNn.

@Skillmonlikestopanswers.xyz that's not how it works, ask @UlrikeFischer

@JosephWright I tend to push "a tag".

@DavidCarlisle see, as long as we got @UlrikeFischer I don't have to take blame for anything, only for \keyval_parse:NNn. And since you gave an in depth review, I can blame you as well :)

@DavidCarlisle Yes, it looks ok.

@UlrikeFischer OK, wish me luck. If it's wrong we can blame @Skillmonlikestopanswers.xyz Hopefully I'll do better than the graphics-pln update this morning:-)

@DavidCarlisle What went wrong there?

@UlrikeFischer someone made a typo filling in the version date in the upload form

@DavidCarlisle But how could that happen? there are no teh in date.

@DavidCarlisle as texlive is frozen, it would only affect MiKTeX and we can always say that they didn't update in admin mode.

@UlrikeFischer merged to master and bumped date/version, running tests again...