6:37 AM
@DavidCarlisle wait, does that mean that mylatexformat.ltx got broken as well?

6:56 AM
@UlrikeFischer stop outsmarting me :)

@UlrikeFischer I'm glad that your solution goes more or less in the same direction (but requires less manual work, that's for sure)

@Skillmon well as long as there is no expandable file_input in expl3, using the primitive is the only choice. But they had imho simply luck that it worked before. \input was never advertised as expandable.

@UlrikeFischer I know (hence my comment that I always would've used the primitive even before, and actually did a few times). An expandable \file_input would be interesting, but I'm not sure how feasible in a robust way. I guess @JosephWright will figure :)

7:17 AM
@JosephWright I don't remember what four-fingers do in Mac. :)

@PauloCereda They type "asdf".

@Skillmon ooh

7:46 AM
@Skillmon The issue is you can't have before/after file hooks: you can arrange to have a hookless/no-name-recording version using \pdffilesize for the searching
@PauloCereda Shows all the open windows

@JosephWright The Super key (the one typically with the Windows icon)?

@PauloCereda Ah, yes

8:02 AM
@JosephWright Why can't I have hooks? I certainly can (checking for a defined command is expandable, expanding a command is). It just can't work with LaTeX's current hook management system :)

@Skillmon Hooks can contain arbitrary code ...
@Skillmon More significant is tracking files: we can't add to the file list by expansion, etc.

@JosephWright Correct and not my fault if the user defines an unexpandable hook :)

@Skillmon ooh rabbits go fishing :)

@JosephWright that one is the real culprit.
@JosephWright more seriously though, I think an expandable file-input would be useful (sometimes) even if it didn't play nice on LaTeX's file tracking and hooks.

@Skillmon Yes, it's been mentioned! The reason for going to \pdffilesize for doing the searching is precisely as it allows expandability at the lower level: the older code uses \openin, which could never work
@Skillmon I've wondered if we should simply give up tracking files at the start of the document, and accept that \input is done by expansion but therefore not ttracked

8:14 AM
@JosephWright doesn't the current \input support hooks? I think I'd stay with this, but provide \ExpandableInput as an alternative (and the CamelCases already give a hint that this is intended for people who know what they do).
It's official: We now have to wear masks even if we keep the distance as soon as more than one person is in the room.

@Skillmon Lucky you
@Skillmon Even though the effectiveness is marginal at best ...

@JosephWright especially if you're wearing the mask for longer time periods (which is exactly what will happen with these new rules). The rules even apply for staying outside. Masks are mandatory as soon as people are "closer" or "meet for a longer time period" (no definition for "closer" or "longer time period" given).
Funny thing is that this does apply for the university. I think the city has to make masks mandatory only on certain squares, everywhere else people will most likely ignore this...

8:48 AM
@Skillmon To be fair, we have much the same except wet labs are currently exempt; luckily, that's the main reason to be on campus

@Skillmon possibly, need to fix the original and best version first:-)

@DavidCarlisle I never used your version (look at me, I'm young!)

@UlrikeFischer @PhelypeOleinik it's clear that I need to take the \endinput out of \document in new releases but not so clear where to put it back? This seems to work but may be abusing the kernel hook?
@Skillmon neither have I
\makeatletter
\let\MYLATEXdocument\document
\let\MYLATEXopenout\openout
% new for 2010-10-01
\ifx\@kernel@before@begindocument\@undefined
\let\MYLATEX@old@endinput\endinput
\else
\let\MYLATEX@old@endinput\relax
\fi
\def\document{\MYLATEX@old@endinput
% end of change
{\setbox\z@\hbox{\normalfont% normal
{\ifx\large\@undefined\else\large\fi
\ifx\footnotesize\@undefined\else\footnotesize\fi}%
{\bfseries\itshape}% bold and bold italic
{\itshape}\ttfamily\sffamily}}%
er replace endinput by endgroup in the above (clearly I didn't test it on an old format:-)
\makeatletter
\let\MYLATEXdocument\document
\let\MYLATEXopenout\openout
% new for 2010-10-01
\ifx\@kernel@before@begindocument\@undefined
\let\MYLATEX@old@endgroup\endgroup
\else
\let\MYLATEX@old@endgroup\relax
\fi
\def\document{\MYLATEX@old@endgroup
% end of change
{\setbox\z@\hbox{\normalfont% normal
{\ifx\large\@undefined\else\large\fi
\ifx\footnotesize\@undefined\else\footnotesize\fi}%
{\bfseries\itshape}% bold and bold italic
{\itshape}\ttfamily\sffamily}}%

@DavidCarlisle You're living in the past!

9:03 AM
oh 2020 then, best to forget the year 2020 I think.

@PauloCereda better than me at soccer for sure... ;-)

9:52 AM
Hi!
I am trying to create a table, and I intend to align the text in a certain column in the center, vertically as well as horizontally.
Now to center it horizontally, I am supposed to use, c in the \begin{tabular}'s options.
And to center it vertically, I need to use m.
But I cannot see how I can do both of them at once. Is there any way?
Using \centering in a m aligned cell doesn't work (it does not produce any error, but neither does it change the output)
Oh, I got it. I just created a new column type: \newcolumntype{C}{>{\centering\arraybackslash}m}. Works perfectly fine as of now.

hello!

3
ooh chicken peace

10:07 AM
@DavidCarlisle well kernel hooks are for the kernel, so .... But imho the first question is why there is the need to put an \endgroup back. I was too tired yesterday to look longer, but I had the impression, that there is a \begingroup too much, as \begin{document} is executed twice. Could that be?

Sep 10 at 23:23, by Skillmon likes topanswers.xyz
@cis There you go, two variants to vertically centre the cells of a column, one based on \vcenter, one based on box dimensions.
@FakeMod ^^^^

@Rmano Thanks! But, uhh..., that code there is out of my scope as of now, so I will just stick with the solution which I can understand :-)

@DavidCarlisle more people following your documentation suggestion... vvvv
This solution is a neat but I still need to consult the manual to understand these apparently predefined names before I can modify the diagram at will. Thanks you! Very nice! — Duns 16 mins ago

10:46 AM
@UlrikeFischer Yes I think the real question is how much of the new \begin{document} processing should happen in the format, and how much in the document. previously as \begin did essentially nothing the format version just preloaded some fonts then dumped the format and a more or less normal \begin{document} was executed in the document. Now it's harder to split. I decided if I was doing this at format creation time then it is essentially a (modified) kernel so I could borrow that hook.

@DavidCarlisle What happens if one use the % mylatex command and stop the format before \begin{document}? Does this work, if you only remove the \endgroup?

@UlrikeFischer I suppose an alternative would be to run \document not \begin{document} in the document pass, that would solve the grouping but means the after begindocument hooks never run at all unless I run them by hand, which seems to require knowing more internals than just adding to the hook
@UlrikeFischer yes I'd guess so, which means that my suggested change probably doesn't work in that case as then the hooks are only run once :(
Oh I suppose in the format run instead of patching \document I could patch \begin to dump before doing any of its normal code, so that all the hooks get run just once, in the main document run....

% new for 2010-10-01
\ifx\@execute@begin@hook\@undefined
\else
\let\MYLATEX@execute@begin@hook\@execute@begin@hook
\def\@execute@begin@hook#1{}%
\fi

% The version of \document to use on the initex run.
% Just preloads some fonts, puts back \document and \openout,
% sets up the banner to display the file list of files preloaded,
% then sets up some special catcodes so the preamble will be
% skipped on normal runs with the new format.
\def\document{\endgroup
{\setbox\z@\hbox{%

@DavidCarlisle I can't test now, only in the afternoon, but it sounds as if you should ask also @PhelypeOleinik and Frank.

@DavidCarlisle Actually you don't even need the \ifx\@execute@begin@hook\@undefined test
@DavidCarlisle This basically kills the new behaviour of \begin so that it behaves essentially like the old one, so you can do the \endgroup in \document. Then in the document run, \begin has the new hook-y behaviour

10:57 AM
@PhelypeOleinik Thanks, that sounds like a better plan. (could you have a look at the mylatexformat version is that broken in the same way?)

@DavidCarlisle Sure. Give me a minute
@DavidCarlisle It seems to work fine

@PhelypeOleinik so you only broke my version. Typical!!!

@DavidCarlisle :-)
@DavidCarlisle A nice sunday friday puzzle for you ;-)
@DavidCarlisle mylatexformat hooks onto \begin rather than \document to find out where the preamble ends, so it wasn't affected. Then in the document run it executes \begin{document} normally
@DavidCarlisle That's what you get for patching \document ;-)

@PhelypeOleinik it was safe enough in latex2.09 I could trust Leslie not to break my code

@DavidCarlisle Kids these days breaking code all around... U_U

11:12 AM
@DavidCarlisle :)

4 hours later…
2:59 PM
@FakeMod If you decide to use it sometime in the future, pick the one based on box dimensions.
@PhelypeOleinik and you're living in the past as well! :)

@PauloCereda The behind the scenes footage for your recent piano video: youtube.com/watch?v=-YriEyWpoiQ
3

@Skillmon Shush, I'm trying to introduce a bug in @DavidCarlisle's code!
5

3:16 PM
@UlrikeFischer In luaotfload I know added features lower and upper. They are not remotely on the level of @JosephWright's code yet, but it handles at least the not language or context-sensitive parts of Unicode's SpecialCasing. (aka it's enough to handle german)

@MarcelKrüger ooh Deutsch

@PauloCereda I think Portuguese works too :)

@MarcelKrüger Bah Portuguese is piece of cake. :)

@PauloCereda In this context that's actually true.

3:32 PM
@MarcelKrüger nice, I will try when I'm home again.

@UlrikeFischer ooh Haus
@MarcelKrüger Mein Deutsch ist sehr gut!
ooh ein Ente

@PauloCereda It's certainly much, much better than my Portuguese 🙈

3:48 PM
@MarcelKrüger ooh

4:00 PM
luatex list, hmmm
@PhelypeOleinik don't worry I'll make sure to blame you

@DavidCarlisle ooh every paragraph

@PauloCereda I thought the linked answer was abundantly clear, and bug free

@DavidCarlisle ooh bugs

4:16 PM
@PauloCereda heading for all the 8's chat.stackexchange.com/search?q=ooh&user=6485&room=41

@DavidCarlisle ooh

@DavidCarlisle hmmm

hmmmm

@UlrikeFischer Phil, who doesn't use latex or luatex seems ideally placed to answer.

@DavidCarlisle ;-)

4:42 PM
is it only me or is the percentage of uncomprehensible and incomplete question quite high currently?

@PauloCereda Ente is feminine, so "eine Ente" :)
@UlrikeFischer Is this the case because?

@Skillmon ;-) something like that yes.

5:03 PM
Anyone knows if there are plans for a new tikz version to be released anytime soon? These "someone ate my cookies bullet points" questions become boring ...

@samcarter_is_at_topanswers.xyz No idea, but I'd like a small change in documentation included.

@Skillmon what change are you waiting for?

@samcarter_is_at_topanswers.xyz Not waiting for, I have to make it myself :) In the documentation of \pgfparserreinsert it needs to state that the current contents of \pgfparserletter get reinserted (which don't necessarily have to be the last parsed token).

@Skillmon oh, killer rabbit power for tikz manual :)

5:10 PM

@samcarter_is_at_topanswers.xyz Will not be the first change I make to the manual :)

New repos on GitHub now use "main" instead of "master" for the main branch.

5:29 PM
@AlanMunn another reason to stop using GitHub... :)

@Skillmon why?

@Skillmon other hosts and git defaults will follow

6:19 PM
@MarcelKrüger german works fine, but greek as expected is missing rules.

6:29 PM
@Skillmon So it is your fault that the manual is so long :)

@AlanMunn I wonder how Metallica songs will sound now...
Main of puppets, I am pulling your strings...

@PauloCereda Having never listened to Metallica, this is not in my list of wonderings. :)

@AlanMunn oh no worries, nothing else matters. :)
2
/ba dum tss

@PauloCereda Shouldn't that be "nothing really matters"? :)

@PauloCereda pulling your sequences of characters

6:33 PM
@StefanKottwitz ooh
@AlanMunn That's Queen. :)

@PauloCereda dramatic lyrics

@PauloCereda Well you go with what you know. So I went.

@StefanKottwitz it is. :)

@PauloCereda awaiting your string btw. :-D

@AlanMunn ooh Joe Kaplow. :)
@StefanKottwitz oh sorry, I am in a hurry today...

6:36 PM
@PauloCereda I see, oh instead of ooh actually shows that

@StefanKottwitz ooh

@PauloCereda no woohrries

@StefanKottwitz ooh

@UlrikeFischer Thanks!
Let's try something: @Henri you'll get a box of chocolate if you release a new tikz version that brings back the bullet points :)

@samcarter_is_at_topanswers.xyz I WILL MAINTAIN TIKZ HENCEFORTH
2

6:50 PM
@PauloCereda :)

Hi!

@FakeMod 'ello

I am using the menukeys package and I am trying to generate the key on a standard Windows keyboard which contains \  and | . It looks somewhat like this:
|---------|
| |~~~~|
| \~~~~|
|---------|

@FakeMod menukeys is currently broken, it doesn't work with a new latex.

(Sorry for the terrible ASCII drawing) How can I do this?
@PauloCereda Hello :)
@UlrikeFischer Didn't know that? It works for me...

(And that's probably because I haven't updated TexStudio and MikTex to the latest version)

@FakeMod texstudio is just an editor - the version of it does not matter

@samcarter_is_at_topanswers.xyz Oh sure. It was just that I hadn't updated both, MikTex and TexStudio.

7:43 PM
@samcarter_is_at_topanswers.xyz only if the parser documentation gets too long.
@DavidCarlisle time to build git from sources... :)

8:19 PM
@FakeMod If you're using the original menukeys, don't!

8:32 PM
(in case anyone is wondering, that^^^ is just because of catoptions)

@Skillmon ugh... cats... figures :p

@Plergux Well, the package isn't about the fur colours of different cat races....

@Skillmon no, but you said the problem was cat-options. :p

@Skillmon There's really no need to recompile git just for keeping "master", the plan for changing the default branch name is to first add a configuration option init.defaultBranch which might then get a new default value. If you don't like it, just change it in your config file.

2 hours later…
10:12 PM
%   \item[D] Stands for \textbf{Do not use}.  This special case is used
%     for \TeX{} primitives.  These functions don't adhere to the
%     \pkg{expl3} API, thus their usage is \emph{strongly discouraged}
%     outside of the kernel:  programmers should instead use the
%     interfaces documented in \texttt{interface3.pdf}%^^A
%     \footnote{If a primitive offers a functionality not yet in the
%       kernel, programmers and users are encouraged to write to the
%       \texttt{LaTeX-L} mailing list
@JosephWright @Skillmon What do you think? ^^

@MarcelKrüger Thanks, it was merely a joke.

@PhelypeOleinik replace 'their usage' with 'their use'. Are all the D commands truly TeX primitives? From other discussions I've seen the D designation included commands whose definition (also) might change, which doesn't seem like a TeX primitive.

@AlanMunn Thanks! Yes, all \tex_<name>:D are primitives (and in LuaTeX some are emulated primitives)

@PhelypeOleinik I would be more explicit, e.g. "These functions have no standarized syntax, they are engine dependent and their name can change without warning."

@UlrikeFischer Looks good! (gives a more concrete reason.) Thanks!
%   \item[D] Stands for \textbf{Do not use}.  This special case is used
%     for \TeX{} primitives.  These functions have no standarized
%     syntax, they are engine dependent and their name can change
%     without warning, thus, their use is \emph{strongly discouraged}
%     outside of the kernel:  programmers should instead use the
%     interfaces documented in \href{interface3.pdf}{interface3.pdf}%^^A

10:21 PM
@PhelypeOleinik I'm not convinced. There (at least at some point) was the general recommendation to \cs_new_eq:NN a private function to the happy-macros, and I'd really prefer if that wasn't completely discouraged (at least for more experienced users/coders) (I'm not sure what Frank would think of this). Also, what @UlrikeFischer says (evil engine dependent behaviour).

@Skillmon I thought about including the \cs_new_eq:NN stuff, but it would make the description for D longer than for the documented signatures (and I'm not trying that hard to make other people happy ;-). Also, (I just found) that explanation in l3styleguide.pdf; maybe a link to this file could be of use somewhere here.
@Skillmon They are not completely discouraged, only strongly :-) there are a few cases you can't really avoid, but who needs to use them most likely knows that (I think)

@PhelypeOleinik sure, but wording it "strongly discouraged" without showing or hinting at an approved way of access really leaves "completely discouraged" in ones mind, imho. (linking to the style guide, a document I had long since forgotten of, seems fine for me)

%   \item[D] Stands for \textbf{Do not use}.  This special case is used
%     for \TeX{} primitives.  These functions have no standarized
%     syntax, they are engine dependent and their name can change
%     without warning, thus, their use is \emph{strongly discouraged} in
%     package code (see the \href{l3styleguide.pdf}{l3styleguide.pdf}):
%     programmers should instead use the
%     interfaces documented in \href{interface3.pdf}{interface3.pdf}%^^A

DUMB question: Is there a way to define a command at the start of my document so my text editor knows it exists and stops highlighting it as wrong, but not actually change the definition the package has??

Like this? ^^
@Canageek You got in the way of my comment :-)

10:28 PM
@Canageek \begingroup\newcommand\foo{}\endgroup.

@PhelypeOleinik Sorry!

@Canageek Just kidding. No problem :)

@PhelypeOleinik typo in "standardized"

(TeXStudio will load things from cwl files, but this package uses sty files :S

@Canageek \usepackage{color} \providecommand\color{zzz} assuming your editor recognises \providecommand

10:30 PM
@Skillmon I dont mkae tpyos! (thanks)

@Skillmon well I think it should be discouraged for "real" code. In experimental code, it can be needed for a short time, but it shouldn't stay. Btw: lualatex-math used them quite unnessarly, and it was broken for nearly a year because it didn't realize that the expl3 names had changed.

@DavidCarlisle Lets find out!
Nominating @DavidCarlisle for sainthood.....
2

@Canageek good plan

@UlrikeFischer then let's hope I'm a maintainer who notices that earlier... :) (I'm using a primitive \scantokens somewhere)
@Canageek but I was faster and gave a possibility using \newcommand (assuming your "smart" editor isn't that smart)

@Skillmon unsaintly suggestions, clearly.

10:33 PM
@Skillmon Fair, that is true. David's felt easier to understand though

@DavidCarlisle it is closed :(
@DavidCarlisle well, mine assumes the "smart" editor is dumber than it thinks, yours assumes it is smarter than I think :) Putting it that way, you're the saint indeed.

The biggest annoyance lately is 1) Discovering a bunch of my citation from my first paper are wrong as I somehow got title and booktitle backwars, which is just embarrassing (Still should be more then enough detail to find the work at least)
and discovering the font I had picked for greek letters looks terrible when used in context and can't find one I like
Tempted to go hunt down and try the lualatex font someone posted here earlier.
Its bad enough I'm tempted to go with italic greek letters even if they are a little wrong (I mean, something like 75% of chem papers get them wrong anyway)

Grr. why can't apple pdf viewer handle links with hashes correctly?

10:49 PM
@UlrikeFischer I'm very tempted to say things about apple, but instead I'll go to bed :) Good night everybody!

@Skillmon well I'm using them too, (well I'm now down to \tex_global:D \setattribute). But I'm not doing it as "experienced user", but because some interface is missing.

Suppose I have a horrendously named file such as "foo - bar - Oct 23, 2020.pdf". Is there any way around the prohibition on spaces in names that pdfpages so vehemently warns against? For various reasons I'd like to keep the names intact (they're generated by my LMS and linked in an HTML file etc., so changing them isn't ideal.)

@AlanMunn does it warn? I never tried.

@UlrikeFischer ⟨filename⟩– Filename of the PDF document. (The filename must not contain any blanks!)

11:04 PM
@AlanMunn Up-to-date TL 2020?

@PhelypeOleinik Within a day or so, yes.

@AlanMunn Strange. For me it errors with LaTeX Error: File foo - bar - Oct 23' not found (note the missing “, 2020”)...

@PhelypeOleinik My response to @UlrikeFischer was the documentation not the error.

@AlanMunn Ah, sorry

@AlanMunn I can look tomorrow.

11:09 PM
@PhelypeOleinik I get the following error:
Undefined control sequence.
<argument> LaTeX3 error:
Unbalanced quotes in file name '"Foo-Bar - Oct 23, 2020 1057 AM.pdf"}

@AlanMunn Oh, you quote the name in the input, Okay, let me check
@AlanMunn Ah, right, the problem with the quotes is that pdfpages takes "Foo-Bar - Oct 23, 2020 1057 AM.pdf" as two files separated by the comma, then quotes are unbalanced.
@AlanMunn The whole issue is indeed the comma and the fact that pdfpages thinks those are two files. If I remove the comma it works: the warning is probably from before LaTeX supported spaces in file names
@AlanMunn The dumbest solution is
\catcode\,=11
\includepdf{foo - bar - Oct 23, 2020.pdf}
:-)

@PhelypeOleinik That will work! This is a single use document. Thanks!

@AlanMunn You're welcome! I was about to say that you'll indeed have to use that trick. pdfpages uses the same routines for treating page lists and document lists, so if you disable one you disable the other (and it's deep down in the package code)