@UlrikeFischer yes

@UlrikeFischer Did you check in?

@JosephWright I thought I did, I will check in a few minutes if I forgot to push

quack

@JosephWright the commit is there github.com/latex3/latex3/commit/…, but it seem to have broken a test.

@PauloCereda XD love how the devil is poking a hole in his bug bag and you'd probably end up with two or three lurking around for the one he took off you. XD

@Plergux ooh

@barbarabeeton enjoy: vimeo.com/75532294

Snow -- Unfortunately it will be gone in a few hours

@UlrikeFischer awwwwww <3

@UlrikeFischer it's the closest I will ever be to snow. :(

@PauloCereda Someday you'll make it to Germany, and then we're going on top of some mountains so that you can be in the snow!

@Skillmon ooh

@PauloCereda and we this year ;-(. We would have been skiing now if it weren't for Corona.

@UlrikeFischer oh

/sad quack

@UlrikeFischer Same here /sad

@Rmano ooh

Remnants of the Filomena's storm (one week later!) around here

@UlrikeFischer in Bayreuth is enough snow so that we can go tobogganing

(and yes, I had to look up that word O_o)

ooh rabbit sports

ooh hi mr. velociraptor

@DavidCarlisle quack

@PauloCereda bad timimg, too late for breakfast, too early for lunch ;-P

@Rmano ooh I am saved

@PauloCereda brunch

@Rmano English has that covered ^^

@DavidCarlisle oops

@DavidCarlisle oh no

@Rmano you are mean

@PauloCereda my daughter has got the habit of singing "Daddy mean" (on the tune of Boney M. "daddy cool") later, so yes, I know ,-)

@Rmano If your daughter remembers Boney M you are older than I thought.

I remember watching Armstrong's moon landing (almost) live when I was 5 yo, on our first TV set...

@Rmano oh so a bit younger than me (I was 8) but quite old for a squirrel

@DavidCarlisle ;-) long-term squirrel

@Rmano ooh

@Rmano Me-me-me-mean! (to the tune of Ma Baker) :)

@DavidCarlisle ooh

@PauloCereda wasn't that for mean cats? ;-P

@Rmano ooh :)

I don't suppose anyone wants to mod ping Martin about

@UlrikeFischer (or any miktex user) vvv

@MartinScharrer ^^^

@DavidCarlisle Done

@JosephWright relying on the old or new package begin document hook to inject a missing {} into package code always seemed a bit fragile:-)

@DavidCarlisle That's the issue? I thought reading the question it was suggesting there is a hook missing

@JosephWright in tex.stackexchange.com/questions/141520/… egreg points out that there is a missing {} in storebox.sty but if you do \AtBeginDocument{{}} as a "temporary" fix the {} magically drops in to the right place, but with the new hook code it has started failing, but that temporary fix was in 2013, it might be time to actually fix the package.

@DavidCarlisle I meant tex.stackexchange.com/questions/579189/…

@DavidCarlisle I added a comment

@JosephWright I think that is just suggesting avoiding the issue by using a different code path, but the reason it changed this release is just that the {} doesn't get added as the immediately following tokens in the hook macro since the last release

@DavidCarlisle I'm not sure. The wrong atbegindocument is naturally something he should fix. But I'm not sure about the hook part.

@UlrikeFischer you got me a tick, thanks:-)

@DavidCarlisle a "uff it works now" tick ?

@JosephWright well, there is in fact a hook missing. AFAIK using only the current hook mechanisms the thing OP wants to achieve isn't possible, as the shipout hooks can't alter the header/footer in the page box. Using everypage-1x gives the correct behaviour. So the point at which the available shipout hooks are executed miss the one point that the original everypage-1x package was using.

@UlrikeFischer a happy user tick

@Skillmon There are differences in the hook placement, but I'm not convince yet, that a new hook is really needed: the user said that xsavebox work, and imho there is not difference in the core code: it also needs the xform stored before the reference.

@UlrikeFischer hm, didn't think of that. But I see no reason why the code of storebox should fail here, other than either being too late, or limited by a scope not there when using everypage-1x. (that's just my guts, I haven't really investigated, yet)

@UlrikeFischer and since xsavebox works (something I have ignored up to now), my guess would then be the scope thingy, there might be a local assignment in storebox that's done globally in xsavebox. Afterall the \mybox is an undefined macro if one uses the storebox code. And if one initializes it with \newstorebox\mybox, the code then throws an error because of a missing number (because the definition of \mybox isn't correct)

@UlrikeFischer I have an idea what to test next.

@Skillmon well one simple difference is that the storebox doesn't store the reference globally:

\documentclass[a4paper]{article}
\usepackage{storebox,xsavebox}

\begin{document}
{\storebox{\mybox}{some page-specific stuff repeated multiple times on that page}}
\usestorebox\mybox %error

{\xsbox{blubbox}{blblblb}}
\xusebox{blubbox} %work

Page 1
\end{document}

@UlrikeFischer what I've said guessed :)

@Skillmon well I saw that already in the morning. But storing the command globally doesn't resolve the error in the question. There is a still another timing problem.

@UlrikeFischer yep, doing \global\let\mybox\mybox does fix it in your MWE, but not in the OP.

@UlrikeFischer I think the issue is that the first execution of shipout/background is too late, if one adds an identical \AtBeginDocument hook the code works (with \global\let\mybox\mybox added)

@Skillmon yes, background is certainly too late, the question is if shipout/before can work.

@UlrikeFischer no, not according to my tests.

we are missing a box operation on coffins bbc.co.uk/news/av/entertainment-arts-55683508

@UlrikeFischer also, the contents of the box can't be updated :(

@Skillmon well no, that is the point of an xform, and the reason why I don't really understand the hook setup of the OP.

@UlrikeFischer yes, but using everypage-1x the box is correctly updated.

@Skillmon the everypage hooks were earlier, but this basically mean that they can't really react to page content, so why isn't the box simply stored earlier, e.g. at the begin of the document?

@UlrikeFischer even if I store it once \AtBeginDocument, I can't alter it in shipout/after or shipout/before.

@Skillmon no you can't. You need a new box if you want something different.

@UlrikeFischer just take a look and play with the following:

\documentclass[a4paper]{article}

\usepackage[pagestyles,extramarks]{titlesec}

\newif\ifstorebox
\storeboxtrue

\newif\ifoldhook
\oldhooktrue

\ifstorebox % use storebox or xsavebox
  \usepackage{storebox}
  \AtBeginDocument[storebox]{{}} % due to tex.stackexchange.com/a/141540
  \newcommand*\usemybox{\usestorebox\mybox}
  \newcommand*\updatebox
    {%
      \storebox\mybox{This is page \thepage.}%
      \global\let\mybox\mybox
    }
\else
  \usepackage{xsavebox}
  \newcommand*\updatebox{\xsbox{mybox}{This is page \thepage.}}

@UlrikeFischer it uses storebox and everypage-1x with the booleans set as they are in the code, and it updates the box contents as wanted for every page.

@UlrikeFischer and if you set \storeboxfalse you also get the correct results using xsavebox. But as soon as you use \oldhookfalse you don't get the wanted output in either case of \ifstorebox.

@UlrikeFischer but the contents are altered with \storeboxfalse and \oldhookfalse, but not for the second page (i.e. too late).

@Skillmon sorry I have to make lunch first, I will look later.

@UlrikeFischer no problem. Go have lunch, that's more important :) Guten Appetit.

@Skillmon \global\let\mybox\mybox looks a bit dubious in any case:-)

@Skillmon don't you want \global\setbox not a local \setbox followed by \global\let ?

as in1-2 is not the same as 3-4 vvvv

\newbox\zzz

\setbox\zzz\hbox{aaa}

{
\setbox\zzz\hbox{bbb}
\global\let\zzz\zzz
1 \copy\zzz
}
2 \copy\zzz


\setbox\zzz\hbox{ccc}

{
\global\setbox\zzz\hbox{ddd}
3 \copy\zzz
}
4 \copy\zzz

\bye

@DavidCarlisle I don't know how \storebox works internally and whether \global would do the right thing :)

@DavidCarlisle yep, \storebox expands to \begingroup... so \global doesn't work. The "box" in \mybox is just a number which is \chardefed, so \global\let\mybox\mybox does no harm here.

@DavidCarlisle the \mybox is not a box but simply a command (defined with \mathchardef)

@DavidCarlisle but the \global\let\mybox\mybox is necessary if \storebox is used in the new hooks, as they are used in a scope.

@Skillmon it makes mybox gloally point to the local value of the box register, it does no harm but doesn't do anything useful

@Skillmon surely you want the box assignment to be global, the chardef of \mybox is done when the box is allocated at the top level so it is already global and globally redeclaring it to its current definition does nothing

@Skillmon see the 1-2 example above: you get exactly the same without the \global\let\zzz\zzz

@Skillmon I wasn't suggesting you put \global in front of \storebox but rather define a \gstorebox that makes a global box assignment

@DavidCarlisle there is no (tex) box. It not even need to get allocated first, you write directly to the pdf, and then store the object/xform number.

@UlrikeFischer ah I thought there was a box as well as the xform, only skimmed the code though. OK the macro shouldn't be called foobox then:-)

@DavidCarlisle well there is a box, but only a tempory while writing the form, it is not needed to reference the xform later.

@DavidCarlisle what @UlrikeFischer says :) Not my fault that the names are misleading :)

@UlrikeFischer I have a few more tidy-ups to do on that int issue :)

@JosephWright ok, I will not interfere.

@UlrikeFischer :)

@UlrikeFischer Next is the colour stack: am also doing some real work today so trying to fit it in

@JosephWright good. You saw my comment about the management code?

@UlrikeFischer No?

@JosephWright this one chat.stackexchange.com/transcript/message/56770432#56770432.

@UlrikeFischer Didn't I do that?

Posting for visibility. twitter.com/latex_ninja/status/1350820270009962496

@JosephWright Not fully. You are testing if the \pdfmanagment_add exists. But the /color.opa 1 ~ gs are added to the page outside of this test. The other question is if your should test there that the management code is actually active (it is possible to load it and then to deactivate it, I use that for testing).

@UlrikeFischer Well yes: I create a stack, but unless someone uses it, all is well. REally, this is experimental stuff with no interfaces ATM, and even once there are, they are going to say 'you have to have the mangaement code loaded'

@JosephWright but you write into the stream:

\RequirePackage{l3pdf,l3color}
\ExplSyntaxOn
\pdf_uncompress:
\ExplSyntaxOff
\documentclass{book}
\begin{document}
abc
\end{document}

gives :

>>
stream
/color.opa1 gs
/color.opa1 gs
/color.opa1 gs
BT

but /color.opa1 is not defined as you didn't add the definition to the resources.

@UlrikeFischer Oh: that I wouldn't expect (the initialise value should logically only appear if you 'bottom out' the stack)

@UlrikeFischer Feels like a bug in dvipdfmx to me ...

@JosephWright that was with pdflatex.

@UlrikeFischer Ah

@UlrikeFischer That's not what I expected

@UlrikeFischer I'll adjust, then

@UlrikeFischer I've moved: I think it's reasonable ATM to assume if PDF management is loaded it's active. We can revisit that once we have it more integrated

@JosephWright ok I will keep an eye on it. if needed it shouldn't be difficult to add the additional boolean.

@JosephWright around?

@PauloCereda Yes

@JosephWright ooh wait a minute then

@JosephWright mail

@PauloCereda Coool

@PauloCereda -- Wow! That's an old one! Thanks. Notice the French cuffs on the operator, and the lab coat on the tech in the darkroom. Very period.

@barbarabeeton A very nice finding!

@UlrikeFischer Exactly: I hope once we agree on a final home for the PDF management code it will be 'always on'

@UlrikeFischer Hopefully now fixed .. back to colour stacks

When I start a new document in overleaf, by default, I have;

\documentclass{article}
\usepackage[utf8]{inputenc}

\title{White spaces}
\author{Du Du Da Da}
\date{January BC205}

\begin{document}

\maketitle

\section{Introduction}

\end{document}

I'd like for the date to be in the upper right corner, preferably by default also. How could one implement that?

@schn not clear what you mean, you get the title formatting from article class, you could add definitions to change that but better to pick a different class, but the default for a new document is pretty arbitrary (I almost always delete that) as you normally have a specific class in mind.

@DavidCarlisle Say for an essay. How would you indicate the date? In Word there is the header option to put for instance the date and then the line currently occupied by date can be used for say a course code/school.

@schn well you could put it in the page head using fancyhdr package but do you want that as a default?

@schn vvv

\documentclass{article}
\usepackage[utf8]{inputenc}

\title{White spaces}
\author{Du Du Da Da}
\date{}
\usepackage{fancyhdr}
\fancyhead[R]{\today}

\begin{document}

\maketitle
\thispagestyle{fancy}
\section{Introduction}

\end{document}

@DavidCarlisle I guess not. But the article class is something I frequently use. Looks nice! If one now puts something in the \date which isn't actually related to the date, is this abuse of the command \date?

@schn yes but as you would be the 1000001'th person to do that, you shouldn't feel too bad.

@DavidCarlisle Feels better :)

@schn that said, not feel bad \neq the right thing to do :)

@schn basically if you are using a journal class to submit a journal article, best to stick to the intended use for commands. But article is latex's general basic anything document class, not necessarily just academic articles so it is often "abused" until a more specific class for the particular use case is written.

@schn ^^

@PauloCereda True. Where in the code to your answer is the date inserted?

@DavidCarlisle Out of curiosity, is it possible to change something to default in Overleaf? Not expecting you to know, but too lazy to google.

@schn er, it's in the screenshot. :) Date is inserted after the author name, as expected in the typical \maketitle from the article class. I simply added a new element with \postauthor which adds arbitrary things after the author info.

@schn your text is far longer than what you would actually type in a Google query box, you know. :)

@schn I only use overleaf occasionally to test things. I prefer to use emacs (on @PauloCereda's recommendation)

@DavidCarlisle oh no

@DavidCarlisle s/use/break/g

@PauloCereda I see. Is \date compensated for your new element \postauthor? With a default article document in Overleaf, \date is explicitly written.

@schn the new element is simply added after the author data. If you really want to get rid of \date being printed in \maketitle you can either patch it or redefine it.

@PauloCereda I see, it follows automatically from \maketitle

marmotghost submitted the tikz-bbox package.

ooh ghosts

Vaccines were finally approved in Brazil!

@PauloCereda oh good luck!

@yo' Thanks <3

To us all!

This naughty virus prevented me from visiting my friends in Europe. Soon I will Europe the hell of this virus. :)

@PauloCereda when you come here, block enough time. (I'll write you a PM)

:(

@yo' Ojh heck!!!

@JosephWright well, I needed a package manual, so I was fine with texdoc, but...

@JosephWright I have a timing problem with the pdfmanagement loading order.

@UlrikeFischer ??

@JosephWright well you are using \pdfmanagement_add in the color backend. This means I must load the code for it before, but this requires that I load the backend and then I'm going in circles.

@UlrikeFischer Ah, right: I still feel the backend should not be fixed until \begin{document} ;)

@UlrikeFischer I'll sort that

@UlrikeFischer I guess I should have \AtBeginDvi, most likely

@JosephWright Currently I'm doing \RequirePackage{l3pdf} which loads the normal backend, and then \file_input:n {l3\g__sys_backend_tl-pdf.def} to load the extra backend commands. But you need stuff from the later. Perhaps you can delay the actual use of \pdfmanagement? It doesn't matter much if it is done only at the begin of the document.

@UlrikeFischer I will do: this is of course still the issue that we really need to finalise the management so the load order is predicatable

@yo' Seems to be back

@UlrikeFischer Try now ;)

@UlrikeFischer Do your destination changes need a check?

@JosephWright well the order is quite predicatable, as I'm forcing everything to load at the start, it only wasn't the order you expected ;-)

@JosephWright \AtBeginDvi doesn't work ;-( But it is not your fault, it is the problem for which I needed a new shipout hook. Leave it there for now, it is good test-- when it works I will know that I have solved the core problem.

@JosephWright well it doesn't hurt if you look ;-) perhaps you have a better idea to do this than with two temporary boxes. But from the interface I think it is fine, better than the box content version.

neural network library for latex?

@Minsky And?

@Minsky I guess you meant to ask "Does anybody know a library to draw neural network topologies for LaTeX?"

@Skillmon And?

@Diaa place the contents of $? inside the file latexalpha2_check.tmp (overwriting the file). The special shell variable $? stores the return value of the last command. So you're saving the return value of the last run command inside a file. The return value should be 0 if the command was successful, else it'll be non-zero.

@Skillmon ;-)

@StefanKottwitz my answer would be: Search the site, there are plenty of examples.

@Diaa I have no idea what command wolframscript does, though :) Anyway its return value is stored.

@Skillmon Thanks! Do you know if the special shell variable $? can be used as it is in Windows?

@Diaa I don't have much experience with scripting in Windows batch/powershell, so no, I'd have to search that as well :)

@Skillmon I think it asks to open wolfram. Just a guess :)

@Diaa seems pretty far fetched :)

@Skillmon What keyword should I search for or start my search journey with?

@Diaa that's quite dated, but was the first hit from my search (should work, but I can't test)

@Diaa so, the variable is called %ERRORLEVEL% in Windows it seems.

@Diaa and that was my search: "windows store return value of last command in file"

@Diaa I wonder why the term "exit code" didn't pop up in my mind, but instead "return value" was returned by my brain query.

@Diaa that looks wrong. At least in *nix && means "and", so the second half will only be executed if the first half was successful (which is the one case you don't need to store the exit code...)

@Skillmon You are right. It should be &

@Diaa I think ; should work in Windows as well (again untested)

@Diaa that looks wrong as well, from a *nix perspective... :)

@Diaa a quick query showed that & is correct in Windows :)