« first day (4788 days earlier)      last day (74 days later) » 

12:09 AM
@cfr I draw the line at !1.baseline :) I like my code to be understandable by me and others.
@DavidCarlisle still "impeccable pedigree" after I got my hands on it?
@UlrikeFischer I'm holding blame in reserve
@AlanMunn Me neither. I looked in the manual.
@AlanMunn That's why I didn't recommend it and said it was rather opaque. However, if you're writing long chains of these in package code, it can be more sensible to write !r1n or whatever than cluttering up the code with something more verbose with a lot of extra curly brackets to keep track of.
12:28 AM
@DavidCarlisle hyperref's color links broke with the new code:

gives in the pdf
/F38 14.3462 Tf 133.768 657.235 Td [(1)-1125(blub)]TJ/F28 9.9626 Tf 0 -21.821 Td [(See)]TJ
\xcolor@ {}{1 0 0 rg 1 0 0 RG}{rgb}{1,0,0}
@UlrikeFischer rember telling @JosephWright that \protected macros are hard to expand when you want to expand them......
@UlrikeFischer I guess that would have worked until I made \xcolor@ protected to gain xcolor compat on named model. Hmm back to \noexpand I guess. Not tonight...
1:07 AM
@AlanMunn Anyway, I'm sorry. I didn't mean the question to sound like that.
2 hours later…
2:51 AM
@cfr Oh it didn't. I took it purely as the helpful suggestion it was.
@DavidCarlisle -- (Okay, I know it's past your bedtime.) Regarding bm, why don't you just update the documentation to say that it's meant for use with "traditional" fonts, and can't handle Unicode fonts (yet)?
5 hours later…
7:55 AM
user image
1 hour later…
@barbarabeeton well no one reads doc:-) Actually just avoiding the tex error of not understanding \Umathchardef tokens is fairly easy to fix (I have a version in github somewhere) but harder is knowing what to do when the desired bod symbol may or may not be somewhere as \symbf or \symbfit or \symbfsf or ... in the same font or might (but probably won't) be available in the traditional way in a matching bold math font
@PauloCereda (a) breakfast (b) I think you posted that before (c) it's far too simple, eg milk is in pints if it's in a bottle, in litres or pints if it's in a carton depending on... things.
@PauloCereda knots and beaufort are missing at the speed branch (but those are of course not exclusively British, I admit)
@campa you deleted the pathname answer? (I voted to undelete)
9:48 AM
@UlrikeFischer I added a test for the unexpanding \xcolor@ (but the code doesn't pass it:-)
10:31 AM
@DavidCarlisle Well that looks very much like a tl (of course, I would like to simplify colour storage)
@JosephWright yes it's just a matter of where to adjust, I can get almost all of it working with just adjusting the def files but the named model (which xcolor mostly breaks anyway) goes through more expansion steps than the others which makes it a little tricky, so may have to adjust color and.or xcolor in the end
10:47 AM
@DavidCarlisle wouldn't it make sense to adjust xcolor anyway to expand?
@UlrikeFischer pragmatically yes although it depends on the "contract" with the driver files, currently it implicitly assumes the driver hands over just a fully expanded driver string. If it hands over an unexpanded xcolor internal, it could expand it away to add the xcolor wrapper or it could simply use the whole thing and not add the xcolor code again
11:14 AM
@DavidCarlisle hm, well in the current setup \current@color contains the \xcolor@ (which by the way probably means that the default 0 g 0 G should be changed for consistency), so I would say the driver hands over something that still must be expanded. But it should be expandable, so no \protected ;-)
@UlrikeFischer (@JosephWright) yep protected doesn't work, but as I say if the driver hands over \xcolor@ {}{0 0.5 0.7 0.1 k 0 0.5 0.7 0.1 K}{cmyk}{0,0.5,0.7,0.1} do we change xcolor to simply use that, or change it to expand it to 0 0.5 0.7 0.1 k 0 0.5 0.7 0.1 K so it can add \xcolor@ to get back to the same thing (which isn't always exactly the same due to numeric differences)
11:31 AM
I would say xcolor should "simply" use that. And seeing code like ` \edef\current@color{\XC@current@color}` I would say one should try to get rid of \XC@current@color and other parts that assume that there are two backend syntaxes.
4 hours later…
3:29 PM
@DavidCarlisle I hadn't noticed you already answered in comments
@campa yes sure but your answer was more complete and gives the OP somthing to accept
@DavidCarlisle Your wish is my command :-)
3:50 PM
! TeX capacity exceeded, sorry [input stack size=10000].
\let ->\let
l.737 % \paragraph
@DavidCarlisle ? how did you manage that? Or is the docu?
diff --git a/bookmark.dtx b/bookmark.dtx
index da2c324..014f326 100644
--- a/bookmark.dtx
+++ b/bookmark.dtx
@@ -727,12 +727,14 @@
 % \paragraph{Some helper commands}
 % Replaces some commands from ltxcmds:
+%    \begin{macrocode}
 \def\BKM@space{ }
+%    \end{macrocode}
 % \paragraph{Packages.}
I could push to the branch:-) l3build ctan
@DavidCarlisle ah, I forgot a macrocode.
@UlrikeFischer fixed
@DavidCarlisle I was just fixing it too ;-)
3:57 PM
@UlrikeFischer You have simply dropped the bookmark-frozen.dtx that Karl asked about?
@DavidCarlisle yes. Looking at it, it doesn't make much sense anymore. It was only meant as a temporary safety measure for the hooks.
@UlrikeFischer I removed the frozen.dtx from the documented source listings
@DavidCarlisle ah, I missed one. thanks
@UlrikeFischer what's stopping merging this with hyperref, are there extra incompatibilities?
@UlrikeFischer may look at color later, but need to drive home, been at mum's for a few days.
2 hours later…
6:30 PM
Hey all and @UlrikeFischer, replaced mathspec with unicode-math yesterday, thanks for the advice and patience
One curiosity, it seems that out of the box (xelatex), it changes the font of the mathcal/symcal family. And I don't really know what font is actually used originally and I don't really know how to attack that problem


% \usepackage{unicode-math} %Comment out to see diff

% $\symcal{Q}$ % comment out to see diff
compiled with xelatex
I have tried to inspect the document to guess the font used without unicode-math (I prefer the look of the calligraphic Q), but no real success
Seems like the change between both MWEs is the CMSY10 appears when unicode math is not used (so it contains the calligraphic Q I am interested in?)
But don't know how to tell unicode-math to use computer modern for math instead of Latin Modern. I have tried \setmathfont{Computer Modern} without success
Don't really know how to explore the fonts available on my system and what alphabets they contain. Gnome fonts for sure is kind of unhelpful here
2 hours later…
8:46 PM
@nathdwek You can open the fonts in fontforge.
9:05 PM
Still very curious what font is responsible for this Q on my system


With xelatex
@nathdwek use lualatex and then look at the end of the log.
\show\thefont says latin modern roman, but using unicode-math and setmathfont to latin modern roman gives a different Q
@UlrikeFischer Excellent advice
I'm learning a lot of (probably useless stuff) on fonts today
@nathdwek When you don't load unicode-math you get the symbol from mf-dist/fonts/type1/public/amsfonts/cm/cmsy10.pfb which is computer modern
Yes I just arrived there @AlanMunn
@nathdwek the point of unicode-math is to use unicode math fonts so using it with cmsy10 which is a 7bit font with just 127 characters is more or less the same as not loading unicode-math.
9:17 PM
So I have to say adios to this gorgeous Knuth Q too?
Oy what a sad day
@nathdwek no you can use classic tex math fonts (which is actually the default with xelatex)
Yeah I meant if I want to still benefit from the other advantages of unicode-math
I just tried \setmathfont{cmsy10.pfb}[range=cal] but eh....I have to read that log
@nathdwek try using NewComputerModern which is closer to cm in some ways than latin modern
@nathdwek no don't do that. only use unicode fonts with \setmathfont
Thanks for the suggestion but a bit boring
Or I probably did not choose the right otf file
Oh I did, ignore the above message
Or just take it as an indicator of my ignorance
@DavidCarlisle I think the mathcal Q look the same in latin modern and new computer modern.
9:24 PM
Sure does
Well I find some respite in my grieving of the cmsy10 cal Q using XITS for just math cal
Was not prepared for today being marked by such an emotional loss
Also blackboard R is a sad departure as well with latin modern :'(
Next time a faculty board asks me if I think I'm a good fit for their culture, I can at least point to this as a testament of how reluctant to change I am
Just a perfect match with some subsets of greybeard academia culture
@nathdwek some unicode math Q
texdoc unimath-symbols for the full list
@DavidCarlisle I wonder how Pagella (and a few others) got their calligraphic alphabets.
I don't have the R column on my system, I don't know which font it corresponds to
And also not listed in the key you link
@nathdwek it's the version in texlive 2023 ^^
Hellooooo ConcreteMath
Ah yeah I'm still on Jammy Vanilla through APT
9:39 PM
@nathdwek VVVV
commit 33aa3ebc93282a593ca4c5814e06f45c270f7e85
Author: David Carlisle <d.p.carlisle@gmail.com>
Date:   Wed Sep 14 21:52:34 2022 +0100

    Concrete, XCharter

commit 9e216b57d6d4a24ff471a1489facc34bfe01acfd
Author: David Carlisle <d.p.carlisle@gmail.com>
Date:   Wed Sep 14 15:38:25 2022 +0100

    ncm Book not Regular

commit 13bcc56e7252bc96897665a68e6c48d25ad3398f
Author: David Carlisle <d.p.carlisle@gmail.com>
Date:   Wed Sep 14 14:28:19 2022 +0100

    additional fonts for #596
Got it
I updated that doc with a few extra columns last year
/must not post Emacs bashing to the DANTE list
@Skillmon or anywhere
9:40 PM
@DavidCarlisle nah, here is fine.
Peace at last
Maybe I can get productive now haha
Unless I choose to randomly hyperfocus on some other silly detail as a way to avoid actual work
@DavidCarlisle how else should you come to your senses and join the one true editor?!
Thanks @DavidCarlisle @UlrikeFischer and @mickep
@nathdwek you should try this for the authentic typeset math experience:
A: Use LaTeX to simulate old typewriter written texts

David CarlisleThis is now available as the typewriter package on ctan and texlive etc Improved version with some Greek and mathematics, and avoiding small numbers being written using 2e-5 notation into the pdf (and crashing the pdf reader) This version assumes that the CM Unicode opentype fonts are available....

9:43 PM
@nathdwek my 2nd highest answer on the site tex.stackexchange.com/a/139878/1090
There is a question (and package?) somewhere out there about an xkcd like font I think
That I used a couple in times in slides when trying to handwave a point across
Also for proper "hand-drawn" "organic potato" venn diagrams
I don't know if the "potato" imagery carries across, but in French (or at least when I was in school, at the school I was), we called venn diagrams potatoes
@nathdwek there are packages for loads of silly stuff (and I have to sadly admit that I'm not responsible for all of them)
because they had to be "not-perfectly-circular" enough to emphasize meaning of geometry
@Skillmon if you used an editor that allowed you to save files you would have written more
@DavidCarlisle ZZ, you should be familiar with it, seems to be your most loved key combination at least for macro names.
@nathdwek we had a very nice Venn diagram in this chat a couple of days ago.
9:47 PM
@Skillmon sadly I stopped using Z, because of real life stuff
Nov 29 at 20:52, by Paulo Cereda
user image
@DavidCarlisle oh no, what happened?
You open a newspaper last year?
Or that would be my guess at least
@nathdwek yes
@DavidCarlisle really? That's keeping you from using \ZZ as macro names?
@Skillmon you can only do what you can do
@Skillmon fortunately Unicode does have other letters
9:50 PM
If I can chime in in support (or at least to say that it does cross the mind of some people), I was pondering making a joke about ZZ three messages above before @DavidCarlisle mentioned they did not use it anymore
@nathdwek I try to not let myself be affected by such stuff, anyone's opinion of whom thinks I'd be in favour of such abhorrent stuff because I use \ZZ in an example code isn't worth taking into consideration.
That is an argument indeed, but in those debates I always tend to side with the side of "but then again, what does it cost to take action X"
@nathdwek well, it invalidates 99% of @DavidCarlisle's code, if \ZZ is now off the charts :)
I have people in my lab arguing with slides and numbers and public figures about how sorting is useless in my country (Belgium) because (and they say numbers support it), the recycling facilities cannot process the collected amount of sorted waste
And it is very technically correct argument
but then again I prefer not looking at the numbers and sorting anyways because it does not cost me anything anyway
@Skillmon I think no one here is pushing to invalidate the existing code or break compatibility
But why not take this moment as an opportunity to explore other fun naming conventions for new "clean-slate" stuff
@nathdwek :) I know. Still it (obviously, for proof see above) never would've crossed my mind to change my habits of coding because of, well, that.
9:55 PM
I've never (that I am aware) been in that position so can't really comment
@nathdwek right, @DavidCarlisle could use the equivalent \wq instead of \ZZ.
That would be a fun thing
It's a world of opportunity
(and under \ExplSyntaxOn he could use \:wq even)
(now we have to make carriage return letter, and he could even go \:wq<cr>)
10:33 PM
@DavidCarlisle and @nathdwek -- the cmsy mathcal{Q} isn't there at all. The Q in column R is closer to the Zapf Q from eusm. (Take a look at the font charts in the amsfonts user guide, amsfndoc; cmbsy is on p.34; eusm is a few pages earlier.)
@Skillmon -- there are more people than @DavidCarlisle who are decidedly wedded to emacs. Sone of them, however, refuse to indulge in editor war battles.
11:00 PM
@barbarabeeton yes although that's only a kind of snapshot, opengtype fonts can hide away whole alphabet styles behind font features so to really check what each font has you need to check the documentation of each, or open in a font editor or....
11:15 PM
@DavidCarlisle -- I know that, but was trying to get a definitive "historical" identification. And since I was familiar with those two, I looked where I knew I could check accurately.

« first day (4788 days earlier)      last day (74 days later) »