are we not meant to remove comments? somebody just told me it is annoying, but I always thought we were meant to if they'd served their purpose. I do try to leave a comment saying I've done this, though.

should I just leave them?

Does this look like a correct 3D representation of a hole-cylinder?

@cis OK, but can be improved where the grey meet the blue in the outer circles.

@mickep With a lower height this becomes nicer:

That is also a perspective

@cfr and @Skillmon I have my keyboard completely personalized, rlog.rgtti.com/2014/05/01/…, but it's a continuous fight with updates (no way to have a local configuration that really works always, even on login, and Wayland makes things more interesting).

git:

==> WARNING: PACKAGER should have the format 'Example Name <[email protected]>'
==> Making package: auracle-git r408.3889e6f-1 (Dydd Gwener 20 mis Medi 2024 07:53:20 BST)
==> Checking runtime dependencies...
==> Checking buildtime dependencies...
==> Retrieving sources...
==> ERROR: /home/software/sources/auracle is not a clone of github.com/falconindy/auracle
    Aborting...
1 ~/builds/git-cl

this is true, but damned if I know what I'm meant to do about it. suspect I should remove the sources.

@Rmano really interesting you prefer the us mapping. that's one reason I have a thinkpad. it's also why I don't have a pound sign, even though it's a uk laptop.

@Rmano what am I meant to use dead grave for in vim? apparently, I'm missing out ....

@cfr that's basically because I grow up with it... The same reason I configure my PC in English (Irish, really) to have an English interface but with Euro and sane dates 😉. But I surrendered with the keyboard, I have Spanish layout in all my PCs now...

@cfr move to mark, I think, row and column

@Rmano latexdef is unhappy:

latexdef \begin
Duplicate specification "Find|F!" for option "find"
Duplicate specification "no-list-def|no-L" for option "no-l"
Duplicate specification "no-list-def-all|no-LL" for option "no-ll"
Duplicate specification "environment|e=s" for option "environment"
Duplicate specification "pgf-Keys|K" for option "pgf-keys"

@Rmano marks are in my heard-of-but-no-idea-beyond-that category.

I keep triggering things by mistake. recording, especially, is a pain.

@Rmano I'm doing similar things. PC in English, keyboard layout is Neo2 (though I can also type on a German QWERTZ layout -- just not as fast, and symbols are a mess there)

@Rmano I grew up with a uk keyboard, but prefer us ;).

@cfr hmmm

% latexdef begin

\begin:
macro:->\protect \begin


\begin :
macro:#1->\UseHook {env/#1/before}\@ifundefined {#1}{\def \reserved@a {\@latex@error {Environment #1 undefined}\@eha }}{\def \reserved@a {\def \@currenvir {#1}\edef \@currenvline {\on@line }\@execute@begin@hook {#1}\csname #1\endcsname }}\@ignorefalse \begingroup \@endpefalse \reserved@a

The backslash is eaten by the shell anyway ;-)

@egreg regarding your answer to tex.stackexchange.com/a/726841/3929, I've been meaning to ask, why do you use \cs_new_protected:Nn here and not \cs_new:Nn? I've been converting out UNI letter class to expl3 and is primarily using \cs_new:Npn and that is problably wrong in many of my cases.

@daleif It contains non-expandable tokens, so is itself not expandable -> protected

@JosephWright So for the most part it is ` \cs_new_protected:Nn` one ought to be using not \cs_new:Np (the current code for my project is working just fine).

A collegue just told me that Foo\textcolor{red}{ Bar} produces "FooBar", but Foo\textbf{ Bar} produces "Foo Bar". No question, this is bad markup; but shouldn't both macros behave the same way with regard to trailing spaces?

@Rmano no necessarily, \b is the bell, iirc. Depends on your terminal what it does with it.

@Lupino well yes, open an issue.

@UlrikeFischer i'm not even sure which one should be the correct one: leaving the whitespace, or consuming it to teach users where to put braces?

@Skillmon ah yes --- my problem is that I am too bash/zsh oriented. To have a real \b you need to \\b (or quote it)

...or something similar

@Lupino that depends how you define "should"

@Lupino Foo\textbf{ Bar} is Foo{\bfseries\space Bar} so you get a space. Foo\textcolor{red}{ Bar} is Foo{\color{red}\space Bar} so you don't get a space as \color does \ignorespaces so that \color{red} Bar is like \bfseries bar and has no space....

i opened an issue: github.com/latex3/latex2e/issues/1474 Let the community decide…

@Lupino what a shocking suggestion, I thought this was a dictatorship, not a democracy

@daleif The function performs assignments, notably \NewDocumentEnvironment

@DavidCarlisle either way, i'm but a humble servant

@Lupino can wolves really be tamed/trusted?

@DavidCarlisle uh, a fellow cat-person :)

@egreg So more or less the protected one is the constructor one ought to be using by default, and only use \cs_new:Nn if the contents is expandable?

@daleif Yes - with lots of hindsight, we likely should have had this the other way around, i.e. some \cs_new_expandable:Npn and \cs_new:Npn making protected functions

@JosephWright noted.

@JosephWright why? I need \cs_new:Npn considerably more often than \cs_new_protected:Npn.

@Skillmon Look at the balance of the kernel or siunitx or ... expandable operations are the minority - once you do any typesetting you are always non-expandable

@daleif Yes

I have a couple of long documents where I use the outlines package. I use the default enumeration, which is a black dot for the first level and then a long minus sign for the second level (I don't use more than two levels). I am a bit lost now since I want to change this enumeration style in all of my documents to a number in the first level and a letter in the second level, however, I want the enumeration of the number to continue throughout the document. Here's an MWE:

\documentclass{article}

\usepackage{outlines}

\usepackage[shortlabels]{enumitem}
\setlist[enumerate,1]{label=\arabic*)}
\setlist[enumerate,2]{label=\alph*.}

\begin{document}
\begin{outline}[enumerate]
    \1 The first bullet point which is numbered.
    \2 The second one which has a letter.
\end{outline}
\begin{outline}[enumerate]
    \1 Continue enumeration of the number.
    \2 Start over enumeration of the letter.
\end{outline}
\end{document}

Preferably, I also don't want to add [enumerate] every time I type \begin{outline}.

@JosephWright who wants to typeset? It's about expandable sorting stuff! No one uses TeX for typesetting

@Skillmon Typesetting would only slow things down. That really can't be the way to go :)

@Skillmon The nearest anyone gets to typesetting is upper/lower case and clearly that is expandable

@Rmano yes, but you use ubuntu, right? so your perl will be older. you have the messages to look forward to .... (at least, I think this is why. I came here to ask @Skillmon, who I think uses arch.)

[2024-09-20T00:41:53+0100] [ALPM] upgraded perl (5.38.2-2 -> 5.40.0-1)

@psie \NewDocumentEnvironment{myoutline}{O{}}{\begin{outline}[enumerate,#1]}{\end{outline}}? (untested)

@cis What is missing are the vertical lines that represent the sides of the cylinder, or not?

@cfr Yep, I am on Ubuntu LTS at work (compromised from the first position of "Windows or Mac"). But I have a Manjaro at home, although I think both are on perl 5.38.2)

@cfr This is an old problem that has been fixed a couple of releases ago, latexdef should update a line when loading Getopt ...at the moment it only returns a warning...in the future it will be an error.

@PabloGonzálezL which probably means patching texdef every time I update TeX Live ... again ;). (because I doubt texdef will get updated anytime soon.)

@cfr yep, I'm on Arch :)

(but haven't done any system updates in a while so might be on an older version than you are, mine still says This is perl 5, version 38, subversion 2 (v5.38.2) built for x86_64-linux-thread-multi)

@cfr no, it will not annoy me if you do remove obsolete comments below my answers. I think that's absolutely your decision to make, as the side effects for me are beneficial: I got an improved version of my answer and everyone will think it was my brilliance, not yours :P

@Skillmon ah, yes. well, when you do update, I'd be interested to know if you see the same thing. last time this happened it eventually became an error and stayed an error for rather a long time. I'm hoping it won't be the same with 5.40.

@Skillmon thanks ;)

@PabloGonzálezL do you have a ref for that or should I not be so lazy and look it up?

@DavidCarlisle so, are you responsible for the decision to add welsh words to the oed? my mother is outraged: apparently, it's a slippery slope to the inclusion of gaelic. who knows what that will lead to? they might start adding bits of latin.