@JosephWright I got with David's longtable branch a latex 20201001, an the updates format failed. Can the content of the latex be older than the current latex-dev?

@DavidCarlisle oh no

@barbarabeeton ooh :)

@UlrikeFischer Depends on where he branched from

@UlrikeFischer Looking at it, would be best to merge or rebase the branch

@UlrikeFischer @DavidCarlisle arara 5.0.2 hit pretest. :)

@JosephWright branched off develop (and claims to be 2010-10-01-pl1

@JosephWright @UlrikeFischer must admit I haven't used it as a real format other than temporary ones for l3build check. I could merge in develop again to bring the format up to date...

@UlrikeFischer try now should be same as develop apart from longtable and a few test files

ooh

@PauloCereda yep got it

@DavidCarlisle longtable

@mickep LOL

@JosephWright could I have rebased without sending everyone the commit mails again

@DavidCarlisle No, that would need a merge commit instead

@DavidCarlisle Anything I plan to rebase I do in a fork to avoid the mails (Git isn't really set up around the whole mail-every-commit idea)

@JosephWright Oh no

@JosephWright yes I may merge next time rather than rebase:-) it'll be a throw away branch after it's finally merged back so it doesn't really matter if the history in this branch is a bit convoluted

It's open now, guys: github.com/cereda/rhabarberbarbara

@Skillmonlikestopanswers.xyz fixed the rhab...bier thingy. :)

Of course it was not Rhabarberbarbarabarbarbarenbartbarbierbier, but Rhabarberbarbarabarbarbarenbartbarbier. Easy to spot. :D

@DavidCarlisle ^^

Look who's here tex.stackexchange.com/users/208556/donald-arseneau

@DavidCarlisle trying to steal your tick? ;-)

@DavidCarlisle looks like a complete (La)TeX rookie.

@PauloCereda yes, it actually was...

@Skillmonlikestopanswers.xyz Donald? yes probably never used tex at all:-)

@DavidCarlisle Sure, not like me, a well known TeXie contributing many packages.

Hello everyone.

@Broxigar hi

Would you mind taking a look at a post to help me out with an Issue that I am trying to resolve?

@Skillmonlikestopanswers.xyz at least he's learned the basic rep harvesting mechanism, start with inserting a change that breaks your package then harvest rep fixing it.

@Broxigar that's how the site works:-)

I know :p but I knew that you would be here so I entered the chat xD

@DavidCarlisle Now the only thing that's missing is a user base of my packages :)

because Its been Idle for a couple of hours and I couldn't solve it by myself

here's the link

@DavidCarlisle you can upload a new longtable version.

@UlrikeFischer Oh gods, no, don't give him ideas!!!

@Skillmonlikestopanswers.xyz he already has ideas ...

@Skillmonlikestopanswers.xyz She was being serious:

@DavidCarlisle the code in the issue looks quite good.

@DavidCarlisle I would, but I fear that this will break my master's thesis severely, so next month?

@UlrikeFischer yes I'll pull it back to the dtx and see what breaks (apart from @Skillmonlikestopanswers.xyz's thesis)

@UlrikeFischer I know, we chatted about it a bit like 2 years ago. He has ideas for a very long time

@Skillmonlikestopanswers.xyz oh a test file. If you install with l3build it goes in the latex-dev part, so you can try it with pdflatex-dev ;-)

@Skillmonlikestopanswers.xyz but these are more focused implement by next latex release version of the ideas. 4.14 not 5.0

@DavidCarlisle what did you implement?

@DavidCarlisle is there finally some firstfoot built into the package? :)

New variant of the miktex update mismatch: format creation failed because of Package expl3 Error: Mismatched expl3 files detected.

@Skillmonlikestopanswers.xyz fixes for "float on same page as start of table", "shrink glue anywheer near start of table" increase chunksize t0 200 in case your machine has more than 500K of memory and a few other bits and pieces

@Skillmonlikestopanswers.xyz I might do that, not sure yet

@DavidCarlisle how do you check for memory from within TeX?

I got texlive 2019 pdftex to segfault, tl2020 gives ! TeX capacity exceeded, sorry [expansion depth=10000].

@Skillmonlikestopanswers.xyz wait 30 years then assume it's safe to increase the limit a bit.

@DavidCarlisle really segfaulting? Or just the capacity exceeded?

@DavidCarlisle but what if I want to run TeX on my father's old Win95 machine?

@Skillmonlikestopanswers.xyz well the terminal output is Segmentation fault (core dumped) so that looks like a real segfault to me

@DavidCarlisle Windows can throw segmentation fault? Or are you in Cygwin?

@DavidCarlisle how?

@Skillmonlikestopanswers.xyz you wo't have an issue as the format will run out of memory loading expl3

@Skillmonlikestopanswers.xyz cygwin, I've never used a native windows tex:-)

@DavidCarlisle true that.

@DavidCarlisle I did a few weeks ago to try benchmarking on that buggy OS.

@UlrikeFischer adjusting bm for Frank's color example. the expansion loop was not totally unexpected but I'm not sure why it crashed rather than just fill a normal tex input buffer stack.

@Skillmonlikestopanswers.xyz you do realise that win95 is modern compared to the longtable design (I could probably already have increased the limits for that:-)

@DavidCarlisle yes, I do.

@DavidCarlisle ah, regarding loops: there is also latex/3887

@DavidCarlisle but that's the oldest PC I have access to, I guess.

@DavidCarlisle Is this the final nail on the coffin of c.t.t.? :)

@AlanMunn doubt that.

@UlrikeFischer yes although the discussion between Morten Chris and me in that one isn't that confident of finding any fix

@DavidCarlisle hm, it mentions a new primitive ;-)

Hello all!!

I am trying to use the Consolas with Sans Serif in listings

I tried with basicstyle=\ttfamily but it produces a Consolas with Serif

@manooooh you need to give at least some clues as to what your input looks like

@manooooh Are you doing this? tex.stackexchange.com/q/180331/2693

@DavidCarlisle ^^^^^^^^

@AlanMunn thanks for the link! I don't know the names, but please see the image. Is what I want teletype font? In that case, I am using pdfLaTeX :(

@manooooh and how is anyone supposed to debug an image?

@DavidCarlisle the new longtable seems to solve 4091.

@UlrikeFischer interesting audit trail in that one:-)

@DavidCarlisle the image of what I want is from the Visual environment; I can't do it

@manooooh this doesn't use consolas so it is quite unclear how it relates to your question.

@DavidCarlisle yes ;-)

@manooooh that isn't using consolas at all

@UlrikeFischer but how do I share the font of the Visual Studio program?

@manooooh sorry I have really no idea what you want.

@manooooh you said I tried with basicstyle=\ttfamily but it produces a Consolas with Serif so where do you have an example that is using consolas?

@manooooh The font of your editor has nothing to do with the font in the output.

@AlanMunn I thought yes. I opened a code with Visual Studio. How can I send you the font used of the Visual Studio program? I sent you the image

@manooooh How about this.

\documentclass{article}
\usepackage{listings}
\usepackage{inconsolata}
\begin{document}
\texttt{This is a test}
\begin{lstlisting}[basicstyle=\ttfamily]
This is a test
\end{lstlisting}
\end{document}

@DavidCarlisle you just have to use the source code and see the result. The font of what I want I don't know how to send it, I used an image for reference

@manooooh you only provided a test file when we asked for that, prior to that you said you had an example "that produces Consolas...." but it appears that you have no such example. If you do not provide code and give wildly inaccurate descriptions of your problem it is hard to understand what your issue is.

@DavidCarlisle I think 3619 was resolved earlier (perhaps with my patch).

@AlanMunn I think it is the solution! Why incosolata don't change normal text (inside document) but the text inside lstlisting does? That's very interesting

@manooooh Because it just sets the tt font, which isn't your main font. But \texttt worked in my example.

@UlrikeFischer \pageshrink=\z@ seems very suspect to me, even though it appears to work

@DavidCarlisle what could it have as bad sideeffect?

@UlrikeFischer opposite of the effect it's curing, the output routine being confused about the shrink on the page and making a page that assumes there is no shrink, when there is, so it ends short or more likely a possible page break is not taken or... I have no idea really, I'm not sure I have ever set it rather than read it before:-)

Does anyone here with an iPhone have any recommendations about software to transfer messages and other stuff from the phone to the Mac. E.g. PhoneView, TouchCopy, iMazing?

@DavidCarlisle this probably means that one will have to try it ;-) If you add the code, I will use it as often as possible.

@DavidCarlisle you could perhaps make it optional, \iflongtablesuppressshrink. Then one could more easily compare the output.

@UlrikeFischer not sure, there are lots of potentially page breaking changes here and I may end up with lots of options, there is always [=2020-01-01] )or maybe I'll allow [=v4] to use the current version

@UlrikeFischer but even if I don't make it an option I could pull it into a small internal macro so you wouldn't have to redefined the whole of LT@start

@DavidCarlisle yes, that was what I meant.

@UlrikeFischer yes but I think it would be an undocumented internal

@DavidCarlisle that would be okay. I can read code ;-) And if it looks as if users would need the ability to switch it off, you can add an interface and something to the manual later.

@UlrikeFischer I thought you read documentation:-)

@DavidCarlisle that too - I'm quite flexible ;-)

@UlrikeFischer can you type "the" ?

@DavidCarlisle the the the the?

@DavidCarlisle "the"

@AlanMunn thank you!

@DavidCarlisle but your question was wrong. It should have been "can you type 'teh''?"

Who was at that time?

@UlrikeFischer it took a couple of goes to get the version I posted:-)

@yo' there's an overleaf q at tex.stackexchange.com/questions/531738/… that you might want to look at

@DavidCarlisle thanks!

@yo' in view of your answer I deleted my final comment saying you should be able to tell minted where the files went:-)

@DavidCarlisle did you already manage to run a lualatex test in tl2020 somewhere? I'm just getting lost in the format settings.

@UlrikeFischer you can set checkengine to luahbtex then it actually makes a luahblatex format locally but it's Ok, I think, although in the end I build iftex on tl2019

@DavidCarlisle thanks

@UlrikeFischer although to be honest I'm not sure it was building the right thing, need to get back to check that, I pushed the pageshrink longtable as you may have seen

@DavidCarlisle I don't have an entry for luahblatex in fmtutil.cnf. Beside this I want to find out how it should be done properly ;-)

@UlrikeFischer yes it moans a bit, I think l3build needs extending but I need to check exactly what the current possibilities are for mapping engine names to latex format names, unless @JosephWright wants to suggest the intended form...

@DavidCarlisle it should work with the specialformats table. I only need to figure out the syntax. Just checking the documentation ;-)

@DavidCarlisle this seems to work. Now I need a way to check if I'm running under tl2020 or not:

checkengines = {"luahbtex","pdftex","xetex"}

specialformats = specialformats or {}
specialformats["latex"] = specialformats["latex"] or {
    luahbtex = {format = "lualatex"}
  }

@DavidCarlisle I can test for tl2020 with something like if string.find(kpse.find_file("article.cls"),"2020") then. But l3build doesn't compare with the luatex files, but with the pdftex files. Unless @JosephWright has a good idea it looks as if one has to make copies of all the xx.luatex.tpf to xx.luahbtex.tpf etc.

@UlrikeFischer can't be hard to make it ise .luatex. for the luahbtex tests, I was looking for something to test for 2020, the find-file works if you sue the standard form but wouldn't work eg for a linux packaged version installing in an undated texmf tree

@UlrikeFischer print(string.find(status.banner,"Version 1.12")) ?

@DavidCarlisle yes, will probably work. But currently I'm wondering if one can tell l3build to use the luahbtex engine even if one do "-eluatex" (and to use the .luatex extension).

@UlrikeFischer yes but you need to do it only for latex tests not plain

@UlrikeFischer a simpler plan would be to persuade the luatex team that luahbtex is a really good idea and just call it luatex :-)

@DavidCarlisle eek. You are right, that could be an additional complication.

@UlrikeFischer that's what killed me doing iftex under tl2020 it has latex and plain tests and getting both working seemed more work than I wanted just to delete that scrbase block, so i used tl2019 but planned to get back to it

@DavidCarlisle it works. I only had all the time engine=luahbtex instead of` binary=luahbtex:

specialformats["latex"] = specialformats["latex"] or {
    luatex = {binary="luahbtex",format = "lualatex"}
  }

@UlrikeFischer ah looking good (although once you have it working we should probably try to get the defaults to do that otherwise every build.lua is going to have to set up lualatex) @JosephWright

with this l3build -eluatex test uses luahbtex --fmt=lualatex

@UlrikeFischer and .luatex.tlg files?

@DavidCarlisle yes, the test passed, which means it used the right reference files (pdf tests).

if string.find(status.banner,"2019") then
  print("TL2019")
else
  -- tl2020
  print("TL2020 or later")

  specialformats = specialformats or {}
  specialformats["latex"] = specialformats["latex"] or {
    luatex = {binary="luahbtex",format = "lualatex"}
    }
end

@DavidCarlisle I was hoping the banner would show "luaHBTeX", but naturally l3build still uses luatex, not luahbtex. I switched the test, so that I don't have to restart in 2021. It looks now like this:

@UlrikeFischer yes I looked for 2019 originally but failed as my tl2019 tree has a self compiled luatex tl2020/dev in its bin directory:-) but that's probably better than testing the version string really I agree:-)

Oh one failure is, is that pdf 2.0 works:

@UlrikeFischer Looks right to me