@DavidCarlisle Oh, that is your way out.

@DavidCarlisle I am sorry, but I don't even understand what that means.

(I must admit that I do not even understand what OP looks for in the question. Why change the look of the question mark?)

Karl seems happy he already has an article for 2023

@mickep well for whatever reason the OP wanted it, it is not an unreaonable question to ask why \mathbf{a} is a bold a but \mathbf{?} is not bold. To make ? pick up the current \fam I changed its class from 5 (close) to 7 (variable) but class 7 is spaced like class 0 \mathord I wondered if in the new engine you could have a variable font family without changing the math spacing.

@JosephWright but doesn't math-active mean that your code would then not accept key=val input in math mode? Seems like a bad idea (think tagging for instance)

@Skillmon math active (ie \mathcode"8000) only kicks in if the = would be typeset

@DavidCarlisle ah, ok. Wasn't fully aware of the implications of math-active, not something I've played much with, yet.

@Skillmon I was looking for what is out there already, not suggesting we make , math active :)

@Skillmon My point was I can't find a case that makes a comma generally active: I suspect it must be a historical thing

@JosephWright and I was just confusing when math-active kicks in, see @DavidCarlisle's message above. Even if the two classes turn , math-active it's fine.

@JosephWright I could imagine some people turning it active to have some "intelligent" spacing to commas... But honestly, the overhead of active commas isn't that much, basically it's just one loop-iteration and one test for a stop condition more (as long as it doesn't show up) compared to if we dropped support for it.

(and the loop is quite fast at that spot, see how fast expkv is for an empty argument to get an idea of the overhead, hint: it beats keyval)

@Skillmon Sure, but the wider context here is trying to get a proper grip on what is active and what is not: I can't stop the odd user doing strange things, but I can look at the package landscape and try to see what makes sense there

In other news: Expansion control in keys is falling in shape. There is one syntax decision I'm not yet sure about, and whether I should provide something like array's \newcolumntype, like \ekvnewexpansionrule.

But I have a working prototype of the parsing and the base-expansion types.

@English language: How would you call this part of this machine? "worktop"? "work-plate"?

@cis Looks like a hotplate to me

@JosephWright Oha

@cis as Joseph said...

And I learned that supporting 3 category codes of : as a separator (which might turn up all at the same time in theory) in an expansion-only context is a pain!

I have one constant time implementation of the splitting that takes 110 lines of code, and one that is faster if only a single category code is used in the argument, but slower for most mixed-cases taking 226 lines of code...

@DavidCarlisle If there were a bold ? in the math font, I agree that \mathbf{?} makes sense. Regarding the font families, I don't know how they are supposed to work (I do not think context uses them).

@mickep and it turned out 0:0, got not a single piece of that melon :(

@mickep ah that's the issue: the Unicode math ranges are somewhat limited so here you really need to switch fonts. Same issue just got asked again this morning by someone else, for digits asking why the command for bold italic a doesn't give a bold italic 1 I doubt "because the Unicode committee decreed it" is the answer they want.

@DavidCarlisle :)

@DavidCarlisle Well, for that one {\bi 0} gives the bold 0 in context.

@mickep \bi for bold-italic?

@DavidCarlisle I asked Hans, and he suggested to perhaps use maxfam as currentfam, instead of using class 7.

@Skillmon Indeed.

@DavidCarlisle But maybe that is the answer they should get. Then they can complain to the Unicode people. (But I do not read the question like that. I think OP there wants just bold and not bold italic from symbfit in the case of numbers)

@mickep ah thanks, I'll look...

@mickep yes here they want bold upright, but it's hard to give a consistent set of commands here as the characters in the math alphabet ranges are so arbitrary, digits, greek, lowercase, all come and go with no documentable logic

@DavidCarlisle but I wonder if unicode-math is doing the wrong thing here. From the code I would have expected it to use the bold digits with \symbfit, but it doesn't. At least that here looks like this:

\usv_set:nnn {bf}     {num}   {"1D7CE}
\usv_set:nnn {bfup}   {num}   {"1D7CE}
\usv_set:nnn {bfit}   {num}   {"1D7CE}

@DavidCarlisle Yes, the Unicode situation is not good. I don't know really know how these things work, but I think there is some clever tagging on the alphabets somehow.

@DavidCarlisle Babel shorthands :)

@JosephWright seems to be active character season

@JosephWright ...there are still things exploding in tikz with active babel shorthand... as circuitikz maintainer, I fear active things a lot ;-)

@Rmano yeah, we know: the plan is to try to get some order

@DavidCarlisle So close to Halloween, they get very active :)

@samcarter don't worry it'll soon be xii season

@DavidCarlisle Ohh, something to look forward to :)

@Rmano The idea is that if we can sort this out in the kernel, packages won't be faced with a variable situation for catcodes: that should work much better

@DavidCarlisle Is the unicode-math package still actively worked on?

@mickep We can make changes, there is discussion about high-level ideas - we need to eventually move something like it into the format

@JosephWright OK, but do you make changes? (It looks like the last one was 9 months ago, or do I misunderstand github.)

Moving the math into the format sounds like a good idea.

@mickep That sounds about right: it's not got lots of stuff going on, but we can make changes (also @DavidCarlisle is meant to be thinking about \math... vs \sym...)

@mickep At some stage, but it's tricky: math mode is very different using Unicode fonts, and you can't assume that 'Unicode engine = Unicode math'

@mickep it's easier if you only need to worry about one engine:-)

@JosephWright Oh, so it is David we should push :)

@DavidCarlisle True.

@DavidCarlisle Btw, somebody was bored today, and now there is:

@mickep we could just close all issues with fixed-in-context and you can support our users,

@DavidCarlisle Oh, good tag!

@JosephWright but what about the pgf-layer, isn't that kernel agnostic?

Question about enumitem: is there a simple way to tell it to wrap every item in a samepage environment?

@JosephWright @DavidCarlisle @UlrikeFischer twitter.com/SoVeryBritish/status/1585252889705861120 What's a pig in blanket?!

@PauloCereda sausage wrapped in bacon

@DavidCarlisle ooh

I wanna be English

@PauloCereda you know it's bad when even the British shy away from giving it a name that suggests it's food ;P (just trolling XD)

@Bordaigorl ooh

@DavidCarlisle ooh

@DavidCarlisle ooh

@Bordaigorl Maybe have a look at the getitems package?

@DavidCarlisle ooh

@PauloCereda can your allegiance be relied upon?

@DavidCarlisle definitely :)

@PauloCereda don't worry, we Germans also put bacon around sausages.

@PauloCereda OK you are in today's cabinet then

@Skillmon ooh das Bakon

@PauloCereda we call it Speck though...

@DavidCarlisle ooh British politics

@Skillmon German is so easy :)

@PauloCereda I think "bacon" originated in French.

@Skillmon ooh

@samcarter that's a very cool package! Thanks, I did not know about it (although it's probably overkill for what I wanted to do)

@PauloCereda counter-ooh

@Skillmon Yes, but the idea is that if you set in the kernel, packages simply inherit the catcodes that are present. So as long as pgf isn't forcing the catcode for chars, things will 'just work'

@DavidCarlisle Ohh

@JosephWright good plan, although premise is false for half of Heiko's packages for example...

\begingroup\catcode61\catcode48\catcode32=10\relax%
  \catcode13=5 % ^^M
  \endlinechar=13 %
  \catcode35=6 % #
  \catcode39=12 % '
  \catcode44=12 % ,
  \catcode45=12 % -
  \catcode46=12 % .
  \catcode58=12 % :
  \catcode64=11 % @
  \catcode123=1 % {
  \catcode125=2 % }
  \expandafter\let\expandafter\x\csname ver@hyphsubst.sty\endcsname

@DavidCarlisle I know

@DavidCarlisle I didn't say it will work ;)

@JosephWright I know you know...

@JosephWright I wonder whether a big bang change would be best or start small eg just make " globally active, as the canonical babel shorthand most packages must have some defence for this already

With this plan babel shorthands would no longer be a problem: chat.stackexchange.com/transcript/message/19701043#19701043

@samcarter that is a good plan

@DavidCarlisle :)

@DavidCarlisle We need to some testing, I think: I suspect overall doing things once is likely easier

@DavidCarlisle breaking change, that should step the major version of the format. So either LaTeX2phi or LaTeX3 :P

@Skillmon \epsilon infinitesimally small so we can go to 2 epsilon+epsilon and still be at 2 epsilon

@DavidCarlisle \epsilon<0 !

:P

@Skillmon Provoking.

Today, I was paying the hospital a little visit.

@Skillmon you had a useful catch added to stop the ring sliding off?

@DavidCarlisle no, I got that one removed... :) Now typing takes a little longer than usual.

@Skillmon use a better editor

@DavidCarlisle you mean one that requires even more key-presses and finger movement? How should that help?

@Skillmon help like this:

@DavidCarlisle does that OS ship with a decent editor?

@Skillmon the best

@DavidCarlisle oh, it ships VIM?

@JosephWright any important expansion type missing from that list?

@JosephWright @UlrikeFischer we have a meeting?

@DavidCarlisle yes in ten minutes (at least my calendar claims that)

@DavidCarlisle Yes

@Skillmon x? As the handling of # is different from f?

@cis -- While I agree with @JosephWright regarding the object, I think you are looking for "work surface:.

@PauloCereda and @DavidCarlisle -- Hereabouts, a "pig in blanket" is a sausage (usually a hot dog) wrapped in biscuit dough. Sorry, David, no bacon.

@DavidCarlisle -- How about wanting to be a citizen of the civilized world? (But I despair.)

@barbarabeeton well one day we will get English spoken in your part of the world:-) A "sausage roll" is a sausage wrapped in pastry

@DavidCarlisle -- I think your hope for a common "English" is doomed. I'll accept "sausage roll" as long as the pastry isn't puff pastry; that's just wrong.

@barbarabeeton merriam-webster.com/dictionary/pig%20in%20a%20blanket :-)

@barbarabeeton The OED has its first recorded use of the term showing up in 1882 and referring to oysters wrapped in bacon

@PauloCereda all explained here: separatedbyacommonlanguage.blogspot.com/2020/02/…

@DavidCarlisle -- That's what I said. "Baked dough" is not the same as "pastry".

@barbarabeeton yep

@DavidCarlisle -- Oh. Oysters wrapped in bacon are "angels on horseback". (And I don't know how they got that name.)

@DavidCarlisle -- Thanks for that link. Makes sense to me. I grew up in oyster country, where they never would have been referred to as "pigs". (And where I grew up, a "pig" was also the name for a bar of just-smelted iron. Totally inedible.)

Here we wrap dates with bacon (and we add almonds or cheese, sometimes).

But we lack fantasy, so they are just "dates with bacon".

@JosephWright not possible in an expansion only context outside of LuaTeX

@Skillmon I never said it was ;)

@JosephWright but V is missing and could be implemented (obviously I can't include the expl3-equivalent of v, but that's just cv in this language).

@Rmano works with honey melon as well

@Skillmon Outch! A speedy recovery for you finger!

@samcarter thanks. Until now it doesn't really hurt to be honest. I just hope it doesn't get infected.

@Skillmon Keeping fingers crossed for no complications, we need your fingers to write all your lightning fast code! :)

@barbarabeeton Similar expression in German: Würstchen im Schlafrock

@samcarter :) Giving it a rest for now, good night!

@Skillmon Sleep well!

@samcarter -- The chef at one of our favorite restaurants specialized in pork, and made his own sausages. The restaurant opened a "Wurst window", where one could pick up an order without going inside. (This was great when restaurants had to close during the pandemic.) They advertised it as "You haven't had our best until you've tried our wurst." Sadly, they closed a few months ago, when he went off to teach charcuterie in Singapore. We miss them!

@barbarabeeton Oh, the people in Singapore are very lucky!