Dammit, someone pointed out both were explained in the documentation, I just read the wrong parts

@yo' happy new year

@DavidCarlisle tl2020?

@UlrikeFischer seems so but still 2020-02-02 not totally surprisingly

@Canageek you clearly took earlier advice on board:

@schtandard looking good (just read the comments in the PR not the code) I actually started to look at highlight after it switched here but haven't any time so disn't get anywhere. I agree with the direction of going for a single language that picks up l3 and @ most of the time.

@JosephWright since @DavidCarlisle gave me the idea, I'd like \lastnamedcs backported into pdfTeX and XeTeX. Any chance this might happen? (Remember, blame @DavidCarlisle if something goes wrong with it)

@Skillmonlikestopanswers.xyz process error, you are supposed to blame @UlrikeFischer not me.

overleaf.com/read/chkkmrymnwsq

@yo' ^^^^

@DavidCarlisle I was expecting a document saying "You should always blame @UlrikeFischer".

Anyone want to guess which is the most viewed lesson at learnlatex.org according to google search console?

@DavidCarlisle I'm lost, is anything wrong? (Sorry I'm not at a PC)

@yo' No it works !!

@DavidCarlisle great :) Is it on TL2020?

@yo' I had a special configuration for marathi documents that uploaded marathi package into each project and I have now ripped all that out

@yo' yep

@DavidCarlisle shhhhh that's still a top secret :-) (the announcements shall go out today)

@yo' I didn't use any secret information just looked in the sidebar of my projects:-)

@DavidCarlisle I know... It's been deployed yestreday and we wait for a bit whether we don't have to roll back before we announce it.

@yo' That never happens with latex releases :-)

@DavidCarlisle isn't the rule "don't look in sidebar bad things can happen"?

@daleif is there a problem with memoir and texlive? It wasn't in todays update.

@yo' True (I have no idea how reading vs. writing translates into the needs of syntax highlighting)

@Skillmonlikestopanswers.xyz Certainly doable: @PhelypeOleinik and I can take a look I'm sure

@JosephWright thanks. Perhaps I should learn WEB, so that I can someday do things like this myself...

@Skillmonlikestopanswers.xyz LuaTeX is not in WEB, so we'll have to work it out a bit more

@JosephWright Yes, but pdfTeX is, so backporting would require me to "speak" WEB...

@UlrikeFischer I got the announcement email from the CTAN people last night. And the proper version is on CTAN. So it probably hasn't reached the mirrors yet (not sure how the updated are taken from CTAN into TL)

[ 8/22, 00:19/01:04] update: jadetex [80k] (54227 -> 56548) ... done

^^ really??

@samcarter_is_at_topanswers.xyz well, for me, the math highlighting in particular is a special thing useful to be sure the $s are correctly paired in a paragraph. That's useful for writing, but I'm not sure how much needed it is for Q&A.

@DavidCarlisle is it a real update or only some catalog/collection business?

@UlrikeFischer I didn't dare look

@DavidCarlisle ctan looks old.

@yo' I know editors which are unable to get $ within verbatim environments right so they think they switched to math mode and highlight the next line or even paragraphs until the next $. Sometimes a %$ at the end of the verbatim environment helps but non-broken math highlighting is non-trivial.

@TeXnician the one in vim is normally ok, although sometimes you have to break out some math block, yes (but it's difficult to trigger, and there is a magic comment provided to reset the state machine).

@Rmano \newcolumntype\math[1]{>{$}#1<{$}} triggers it, iirc. Something I've done in a few tables...

@Skillmonlikestopanswers.xyz Yes, and unbalanced delimiters also...

@JosephWright If that's to be added, then maybe \begincsname can go as well... I'll look into that, eventually

@PhelypeOleinik Probably yes: I guess the code goes together

@PhelypeOleinik You happy to look? If we do it, we need to do pTeX and upTeX as well as pdfTeX and XeTeX

@JosephWright If I recall correctly yes, it's the same procedure in LuaTeX, just with different flags

@JosephWright Yes, probably do in tex.ch (without affecting tex, of course) so all engines have it

@JosephWright Though I'm not sure how soon that will be ("soon" isn't the right word :-)

@DavidCarlisle Akira is fast: tug.org/pipermail/dvipdfmx/2020-October/000110.html

@PhelypeOleinik Sounds good

@PhelypeOleinik I guess if we can aim for TL21 it would be handy, so something like by the end of the yeaer

@JosephWright and hint and aleph and ... :-)

Do I need to show the .cls file with my MWE here?

@raf you may as well 126 lines == nothing

okay, I am adding it.

@raf better still of course would be to check if you can show the problem with article (apart from requiring less code to post, it helps narrow down the issue if you kno wit isn't in that class

@DavidCarlisle :)

@DavidCarlisle HINT is actively developed, so we are in the clear; for Aleph, I think you know what I think!

@UlrikeFischer do you know what changed in gs ?

@DavidCarlisle no, they seem to change something constantly ;-(. Perhaps Akira's check-in gives some clue, but I have to do some shopping now so can't check.

@JosephWright To keep the "one engine enhancement per year" streak? ;-)

@UlrikeFischer thank you so much. It's working perfectly now!

@StefanKottwitz ooh German snakes go hißßßßßßß

@UlrikeFischer Is the Greek issue fixed more-or-less now? Or is there more I should look at?

@JosephWright apart from the expanding business in hyperref it works fine.

@UlrikeFischer OK: that one is really at the hyperref end (unless we fundamentally re-think using \exp_not:n for 'return' of expandable tokens)

@JosephWright well the main question is if one can use \expanded there or if it will explode with other code. I will probably have to set up some tests.

@UlrikeFischer That comes down to if the content is 'text'. Probably for stuff that might contain LaTeX2e commands you have to use \protected@edef, so do the expansion twice; like I say, perhaps we should rethink using \exp_not:n (though it is handy)

@JosephWright could the \exp_not:n be added conditionally, as in a "texorpdfstring" ?

@UlrikeFischer Yes, if we had a flag set (so a bit like \protected@edef): one for the team list I suspect (would also be of interest for \prop_item:Nn, etc.)

Question: on a non updated memoir, aren't this suppose to clear the package/before/caption hook? It clearly isn't

@UlrikeFischer yep, the official announcements shall go out today.

@daleif Hm, it should... Let me check

@daleif You are at the top level, so the 'guess the label' stuff doesn't work, you need \RemoveFromHook{package/before/caption}[memoir] (@PhelypeOleinik)

@daleif No, it shouldn't. Inside memoir, the default label is memoir, so you need to provide the optional argument:

\documentclass{memoir}
%\DebugHooksOn
\RemoveFromHook{package/before/caption}[memoir]
\AddToHook{package/before/caption}[memoir]{\typeout{before caption}}
\ShowHook{package/before/caption}
%\usepackage{caption}
\begin{document}
test
\end{document}

@JosephWright ahh, it also work with [*] as there are no other caption hooks

@PhelypeOleinik What I said ;)

@daleif (what Joseph said :-)

Yeah

I'm slow :-)

@daleif Indeed, though risky of course

@PhelypeOleinik So the point is that the hook came from memoir then that is what one should need. Hmm, that ought to be explained a bit better in the manual.

@JosephWright Just wanted my doc to compile (though in my real case I just copied in the updated memoir)

@daleif Yes, the idea is that you don't "accidentally" remove other packages' code from the hook without explicitly saying what label you want to remove. Any suggestion where/what I can improve?

@daleif This should be documented in the long-ish section 2.1.3 Hook names and default labels

@daleif In your example code, the label is top-level (the default in the document body), so the \RemoveFromHook removes* the \AddToHook you do on the next line and memoir's code remains

*Actually since it comes before, it prevents that the code from being added

@PhelypeOleinik It might be an idea to mention along side \RemoveFromHook that one can use \ShowHook to figure out the label if one does not quite understand which label is needed. It took me a while to decipher how to read that information. Might be worth listing that \ShowHook{hook}` lists the labels associated with the hook code and that is the label one needs to remove that exact piece of code.

@daleif Got it. Indeed, the documentation of \ShowHook is lacking some detail. Thanks for the feedback!

@PhelypeOleinik Not quite understood, if I ion my preamble use \RemoveFromHook{hook} and then \AddToHook{hook}{code} than code is also gone because this was in the preamble.

@daleif Yes, that's right, because we want to allow packages to be loaded in arbitrary order. Suppose a package A.sty that does \AddToHook{hook}{code} and package B.sty does \RemoveFromHook{hook}[A]. It goes well if A is loaded before B. To allow the case where B is loaded before, I implemented \RemoveFromHook to remove code chunks not yet added

@daleif So (assuming the same label) \RemoveFromHook{hook} and then \AddToHook{hook}{code} has the same effect as \AddToHook{hook}{code} and then \RemoveFromHook{hook}

@PhelypeOleinik That ought to be listed somewhere.

@PhelypeOleinik That one caught me out :)

@daleif Hm, it is, but under the expl3 interface \hook_gremove_code:nn. I'll adjust this one too

@PhelypeOleinik Is it only in the preamble the order does not matter?

@JosephWright Why so?

@daleif Anywhere, as long as the label is the same

@PhelypeOleinik Most users probably stop reading when they see the first _ in a macro name

@daleif :-)

@PhelypeOleinik Does the delete happen at the first/next use of the hook?

@daleif It works like this: \RemoveFromHook{<hook>}[<label>] checks if <label> exists in <hook>; if it does, then it is removed normally, and if does not, then (<hook>,<label>) is added to a “removal queue”

@daleif \AddToHook{<hook>}[<label>]{<code>} then checks if (<hook>,<label>) exists in that removal queue, and if so it doesn't add, otherwise it adds normally

@daleif At the point the hook is used, no addition/removal operation takes place: only relations added with \DeclareHookRule are evaluated

TeX Live 2020 Now Available in Overleaf!

^^ @DavidCarlisle @UlrikeFischer it's official now!

@yo' Just wondering, what are your update policies on the images you use?

@daleif in very short: bugfixes only. In reality, some things are possible, but we're extremely reluctant.

@yo' so the 2020-10-01 format might not be available

@daleif that's a tough one, yes.

@yo' At least I'd wait a few months on that one (jf memoir for example)

@yo' Pre-loaded expl3 :)

Note that it took us several weeks to make sure it's fine.

@PhelypeOleinik I guess as I was thinking one would \@ifpackageloaded{foo}{\RemoveFromHook{begindocument}[foo]}{\AddToHook{begindocument}{\@ifpackageloaded{foo}{\ERROR}{}}} or similar, i.e. remove stuff if it's present, but not try to prevent it adding 'aafter the fact'otherwise

@JosephWright That can be done, yes, but the idea we discussed was that the effect of 'removing before adding' should be invisible, so package loading order loses some of its importance

@PhelypeOleinik Sure, I guess I'm just used to a particular approach that relies on interlocking tests

@JosephWright It has the advantage of much better control, indeed (and will probably be needed in some edge cases), but it soon becomes the \AtBeginDocument clutter we always had

@JosephWright But it's definitely not intuitive to 'remove before adding'

@PhelypeOleinik Certainly needs mentioning

@JosephWright I'll improve this part

@PhelypeOleinik No rush

@yo' Twitter announcement due soon?

@JosephWright yep, maybe an hour or so

@yo' postal news? :)

@PauloCereda I'm sooo busy, I didn't have time to get there yet :(

@yo' oh

@JosephWright Yep, pre-loaded expl3 :) overleaf.com/…

@JosephWright auto-pst-pdf just tried one windows to delete my whole test folder ...

@JosephWright It can fasten up checking whether a macro is defined quite considerably... :)

@UlrikeFischer Oops

@JosephWright it tries to run del "\@tempa" where \@tempa can be empty, and windows interprets this as del *, happily there is a security question first.

@UlrikeFischer Oops

@JosephWright what do you expect here:

\documentclass[]{article}
\begin{document}
\IfFileExists{}{true}{false}
\end{document}

@UlrikeFischer wtf, why would Windows think that del "" is the same as del *?

@UlrikeFischer false (though my name isn't Joseph).

@yo' less timeouts!

@Skillmonlikestopanswers.xyz Why do you use "why" and "windows" in the same sentence?

@samcarter_is_at_topanswers.xyz "Why would anyone use Windows?"

@Skillmonlikestopanswers.xyz hopefully! We do timing tests and comparison, and things might have gone slightly better on avarage.

Strange question: anyone have an example of a 4byte unicode char? I need to test mysql utf8 vs utf8mb4

@Skillmonlikestopanswers.xyz touché :)

@yo' expl3 is huge, putting it in the format will be considerably faster (which reminds me, I wanted to work on including mylatexformat into a latexmk rule for Overleaf sometime in the past... Never got to it). Same would be true for TikZ/pgf, KOMA, and others, all of which could be built into a project-based custom format (but there are things that don't work in formats and have to be executed later).

@UlrikeFischer Hm... \IfFileExists{} is \IfFileExists{.tex}, which is true

😃 seems usable ;-)

@PhelypeOleinik well it is different to tl19, and it caught auto-pst-pdf by surprise: it has a test only to delete only existing files, and not suddenly got an empty argument.

@PhelypeOleinik yes, but .tex is a file that only does echo ignored, so better ignore it :)

@Skillmonlikestopanswers.xyz well why does latex think that {} is the same as {.tex}?

@UlrikeFischer TL 19 tested csnameendcsname.tex, which would give true if it existed. Not good either

@UlrikeFischer because whenever you give LaTeX a file without extension it appends .tex to it.

@UlrikeFischer and certainly {} doesn't include a file extension... :)

@JosephWright the Twitter post is out! twitter.com/overleaf/status/1313468309212401666

@Skillmonlikestopanswers.xyz Neither \input{foo}, but it does read foo.tex

@PhelypeOleinik which is completely in line with what I wrote, isn't it?

@Skillmonlikestopanswers.xyz Ah, right. Didn't see the message above, sorry

@PhelypeOleinik no problem.

@Skillmonlikestopanswers.xyz Chats are confusing :-)

@Skillmonlikestopanswers.xyz KOMA is nothing like as big; pgf would benefit but largely as it has lots of files

@Skillmonlikestopanswers.xyz yes I know, but what I meant is: also tex tries to "translate" user input. It is not completly idiotic from windows to interprete del "" as del * .

hmm, as expected, web app did not like 😃 one bit. But that was the test, nothing else I did generated an error like that so my data seems ok.

@JosephWright well, considering that whenever I build a document using my custom class (which is based on KOMA) I can see a bunch of files in the log starting with scr and taking the majority of the preamble's log-space... It is nowhere near as big as pgf, that's right, but it isn't small either.

Breaking news! A mathematician has been awarded the Nobel prize for physics!

@egreg well, he has been working in a lot of fields

@UlrikeFischer There is a big difference from "Hey you omitted a file extension, let me try whether I can find that file with the default extension" to "Hey, you didn't say which file you want me to delete, so I delete every file, that should also include the file you meant"

@UlrikeFischer if Windows would instead try to delete a file with an empty string as its name, I'd totally understand (after all, I specified the file name to be an empty string, that's what "" usually means on the command line), but everything? Seems like a pretty drastic way to punish user input.

@Skillmonlikestopanswers.xyz It's what command.exe does (I just tested)

@UlrikeFischer I must say I find TeX's automatic .tex extension rather idiotic (maybe back then it was needed, for some systems, but nowadays it's a pain)

@PhelypeOleinik yes, you save yourself a bit typing at the cost of curious errors and lots of code gymnastic.

@PhelypeOleinik well, it at least makes some sense... But agreed, it gets in the way now and then (but not during code golf, those are valuable 4 bytes!)

@Skillmonlikestopanswers.xyz “now and then” because you didn't write the ltfilehook code :-)

@PhelypeOleinik and that's precisely why it only gets in the way now and then, because you cleared the way pretty good :)

@Skillmonlikestopanswers.xyz <3

@PhelypeOleinik oh, a carrot!

@Skillmonlikestopanswers.xyz ooh

U+1F955

ooh a uni-carrot

@Skillmonlikestopanswers.xyz :-)

@PhelypeOleinik it makes sense on the commandline, less so for internal input of files

ooh the duck

@matt ooh

@Skillmonlikestopanswers.xyz since you are our resident rabbit, wanna see an Easter egg? :)

@PauloCereda so many ooh's

@matt quite :)

@matt That's his speciality.

But maybe you haven't been here long.

o o h

@matt it has to happen naturally. :)

@PauloCereda always!

@Skillmonlikestopanswers.xyz ooh :)

@PauloCereda oh no!

(will remove the reference)

@Skillmonlikestopanswers.xyz oh no <3

@PauloCereda "Made with by the Island of TeX"

@Skillmonlikestopanswers.xyz Almost a carrot :)

@matt Nice effort. But next time, with more conviction.

@matt 8711 and counting

@HaraldHanche-Olsen ooh, someone who counts.

@DavidCarlisle thanks for the feedback! Any thoughts on highlighting braces and/or brackets?

how do you give transitions 4-6 in \tikzmarknode[show on=4] ?

I tried <4-6> but that does something terrible to latex

{4-6}is the answer

@Anush doesn't show on=4-6 work?

@Anush too slow... :)

@Skillmonlikestopanswers.xyz it turns out yes. Thank you

@DavidCarlisle someone uses your package: tex.stackexchange.com/questions/565684/…

@Skillmonlikestopanswers.xyz v5 :-)

@DavidCarlisle I fear the day it is released :)

@DavidCarlisle I guess it'd break a lot of my code (but making it easier to reimplement it)

@Skillmonlikestopanswers.xyz if I get back to that I'd probably start again, there is a less adventurous longtable4.14 in the main latex2e repo, I hope to actually ship that for the spring 2021 latex release, we'll see...

@DavidCarlisle my code is tested against 4.11 - 4.13, I hope it'll be ok with 4.14 as well.

@DavidCarlisle does 4.14 include \endfirstfoot?

@Skillmonlikestopanswers.xyz I can't remember:-)

@FaheemMitha ooh, a critic

@matt Nicely done.

@FaheemMitha thanks :)

@DavidCarlisle I guess we should disable kvoptions-patch?

@UlrikeFischer ?

@DavidCarlisle github.com/ho-tex/kvoptions/issues/5

@UlrikeFischer sorry got behind in mail:-) I think we looked at that before, is it ised much can we make it do nothing? Or give an error or ..?

@DavidCarlisle I can't remember to have seen it, but basically it now breaks \RequirePackageWithOptions{...}, and looking at the diff and all the patches I can't really see how to repair it.

\begin{filecontents}[overwrite]{foo-test.sty}
\RequirePackageWithOptions{bar-test}
\end{filecontents}
\begin{filecontents}[overwrite]{bar-test.sty}
\DeclareOption {quiet}     {}
\ProcessOptions*
\end{filecontents}
\documentclass{article}
\usepackage{kvoptions-patch}
\usepackage[quiet]{foo-test}
\begin{document}

Test
\end{document}

So I think we should make it do nothing in a new latex, perhaps along with a warning.

so which documents then break

@UlrikeFischer really need to get that raw option list code in next time, then it won't be needed at all and we can make kvoptions use the raw lists

@DavidCarlisle yes, but for now I guess we only have the options to either disable kvoptions-patch or to try to repair it as it breaks \RequirePackageWithOptions. We could ask @PhelypeOleinik he seems to know it: tex.stackexchange.com/a/444351/2388.

@UlrikeFischer yes I guess disabling is best if we don't see a quick fix and I'm not sure I want to look this evening

@UlrikeFischer Just yesterday I've removed kvoptions-patch from a class file I wrote :-)

@PhelypeOleinik so you agree with disabling?

@UlrikeFischer No objection from me

@PhelypeOleinik @DavidCarlisle ok. I will then answer the issue in this sense and prepare the upload.

@UlrikeFischer doesn't look too bad:

$ grep -rl  kvoptions-patch  /usr/local/texlive/2020/texmf-dist/tex/latex
/usr/local/texlive/2020/texmf-dist/tex/latex/confproc/confproc.cls
/usr/local/texlive/2020/texmf-dist/tex/latex/doctools/doctools.sty
/usr/local/texlive/2020/texmf-dist/tex/latex/kvoptions/kvoptions-patch.sty
/usr/local/texlive/2020/texmf-dist/tex/latex/kvoptions/kvoptions.sty
/usr/local/texlive/2020/texmf-dist/tex/latex/latexdemo/latexdemo.sty

@DavidCarlisle no, but kvoptions is used more, and the question is if documents load the patch (and rely on it) with the patch option, or locally. But I think there is not much one do apart from waiting what happens ... And time is better spent on improving the option system directly.

@UlrikeFischer yes OK

@UlrikeFischer we could also suggest this

\RequirePackage[2020-02-02]{latexrelease}
\documentclass{article}
\usepackage{kvoptions-patch}
\usepackage[quiet]{fontspec}
\begin{document}
Test
\end{document}

Looking in after multiple hours of Zoom classes today.

@LaTeXereXeTaL giving or taking

@DavidCarlisle Giving. I've been teaching from home since mid-March.

@LaTeXereXeTaL ah so it's harder to tune out and not follow then:-)

@DavidCarlisle I've been working expansion exercises from the TeX Book in between.

@LaTeXereXeTaL just skip straight to appendix D it's the best bit

I enjoy observing the developers create the software I use. The dedication is simply superhuman.

@DavidCarlisle Got quite a nice compliment from a current student today, who had a friend doing a Master's at a nearby university who said he was amazed that students were learning LaTeX in an introductory physics course at a community college. Made me feel good about the decision to use it.

@LaTeXereXeTaL I never used latex as a student:(

@DavidCarlisle I only learned of it my first year of grad school ('89) but didn't have a computer powerful enough to run it on. I started a few years later with emTeX in MSDOS and OS/2. Good times...

@LaTeXereXeTaL my thesis was submitted in 85 so in theory tex was available but wasn't well known in the Uk (nor easy to find a machine of any sort) so golfball typewriter it was (partly why I have little sympathy for students who claim their thesis requirement require some spacing specified to millimetre accuracy:-)

@DavidCarlisle Until I got my first PC in 1989 (December 15, as I recall) I used a Commodore SX-64. LOL

@DavidCarlisle I pushed the changes, but will now go to bed and do an upload tomorrow.

@UlrikeFischer thanks, goodnight