« first day (5070 days earlier)      last day (85 days later) » 

cfr
cfr
05:32
@DavidCarlisle thanks. since bad luck comes in threes anyway, I'll just send karl a third query.
 
6 hours later…
11:06
I pokered too high. I was hoping for an upvote before I get the next tick, now I missed the palindrome by 5 :(
3
@Skillmon rough
11:27
@Skillmon why would you expect an upvote?
@DavidCarlisle why not? People get upvotes all the time, why shouldn't rabbits get them as well?
@Skillmon onlly soft rabbit shaped furry rabbits
11:42
@DavidCarlisle that's transfurphobic!
or xenofurphobic? Is there a word for not-furry-phobic? Transfurphobic isn't correct, since my species doesn't have fur to begin with...
@Skillmon How rude of this users to accept a perfectly working answer :)
@samcarter Exactly! I knew you'd understand :)
@Skillmon you can always retry... give a bounty to @DavidCarlisle 😆
cfr
cfr
12:09
@Skillmon you could write a different answer as community wiki and hope the OP accepts it instead :-).
how common is it for packages to use _ as a letter but not use expl3?
@cfr Rare but not unknown
cfr
cfr
@JosephWright seems confusing.
@cfr mathtools for example does it.
@cfr Some predate expl3, some are format-neutral, some are more recent but the authors have made that call (as underscore is pretty common in programming)
12:24
@UlrikeFischer that one does it especially well and thought out.
cfr
cfr
@UlrikeFischer that was the other package I was thinking of. (indirectly - I actually managed to find your message using the search function, though I'm not sure I could do so again).
@UlrikeFischer I didn't know that.
12:58
@UlrikeFischer yes but mathtools is _really_expl3_but_not_quite:_we_should_fix_that
@DavidCarlisle Yes, we probably should
13:12
@DavidCarlisle _oh_no!
@mickep oh no
@PauloCereda oh_dinner_!
@DavidCarlisle oh no
13:40
@DavidCarlisle mathtools and empheq both uses mhsetup (a prefixed subset of the expl3 syntax).
@daleif yes but it's a subset of expl3 as it was 10 years or so ago and doesn't match current code, it's not really helpful, I think we should lose mhsetup and use standard expl3 command names
@daleif it was always confusing but sort of made sense to avoid having to load expl3 but as expl3 is in the format now it really is just confusing with no advantages
@DavidCarlisle I know, but someone has to sit down and do it, and I do not have the time at the moment. And last I had a look, there is not clear mapping from mhsetup to current expl3 (as you mention), which makes it harder to convert it without making changes in the out put.
BTW: what is the coffing equivalent of \box_resize_to_ht_plus_dp:Nn? Interface3 only mentions resize and scale, and it is not clear if any of them can scale in such a way that the aspect ratio is kept.
@DavidCarlisle The only advange is those extreme users who manually update packages, then they can do that using just 3 files. AFAIK you could do that with archive if your doc used something never than what they had in production. But IMO we cannot let those hold progress back.
14:16
@daleif yes but three files is one too many. as expl3 is in the format you could drop mhsetup and use expl3 commands directly (defining any missing commands in the packages)
@daleif err pass (ask @JosephWright :-)
@daleif yes I know that feeling, there is clearly no rush, but as we start to move tagging support for math in the direction of mathtools having "almost expl3" and "expl3" in close proximity will be an issue I think.
@DavidCarlisle I kinda need a way to convert a coffin into a box. This particular feature, is not used that often, so saving a coffin in a box os probably not too bad.
@daleif I was going to say you could coffin_typeset into a box. there must somewhere be a box register in the coffin structure but I'm not sure if we offer any documented way of reaching it directly
@DavidCarlisle that was what I just did. Seems to work fine.
@DavidCarlisle Only ways that are secret to @PauloCereda
Is there good intro to making graphical test files? I was thinking about making a test suite for mathtools such that we know if some output changed, when we start converting it to expl3. Then one could just go through the manual and make test files for everything documented.
When looking at empheq, I'd kinda like to have the empheq react to a qedhere option, such that it can also add a nicely placed tomestone in math mode like it can with ntheorem. That's a feature feature I've been missing lately in other amsthm based projects.
14:33
@mickep being extinct you would know about extracting from coffins
@daleif depends on how you define "graphical" a standard l3build test would normally be enough using \showbox or \showoutput log format of the typeset result
@DavidCarlisle yes, but the documentation is rather hard to understand. So I was just looking for auitable examples, that is used to compare the typeset output.
Is \hbox_overlap_right:n equivalent to \rlap? I'm kinda missing a hackers note (so one can search for \rlap)
@daleif almost all l3build tests are doing this look eg in latex2e/required/tools/testfiles:
$ grep showoutput -l testfiles/*lvt
testfiles/github-0042.lvt
testfiles/github-0053.lvt
testfiles/github-0072.lvt
testfiles/github-0100.lvt
testfiles/github-0102-tools.lvt
testfiles/github-0131.lvt
testfiles/github-0152.lvt
testfiles/github-0173.lvt
testfiles/github-0237.lvt
testfiles/github-0270.lvt
testfiles/github-0297-rollback.lvt
testfiles/github-0297.lvt
testfiles/github-0614.lvt
testfiles/github-0766.lvt
testfiles/github-1128.lvt
testfiles/github-1160.lvt
testfiles/github-1245.lvt
testfiles/github-1377.lvt
ahh, the rlap def is the same as that one
@DavidCarlisle thanks
@DavidCarlisle But the ducks don't...
And smash is \vbox_to_zero:n? Hmm not quite TbT says that's a hbox.
14:47
@daleif without the math tests, possibly nearest 2e equiv would be \parbox{\hsize}[0pt] to force zero height
@DavidCarlisle Math tests? Oh, the ones amsmath uses? Here is only use it for boxed data.
@daleif even without amsmath, latex (and plain tex) define \smash in math mode to have math content in a \mathchoice
I'd like to see how far I can take this letter class to expl3 conversion, so as few standard TeX/LaTeX constructions as possible.
@DavidCarlisle Oh, I did not even check source2e
I still need to replace all parboxes and minipages (same thing). I have one or two places where I use varwidth, which I probably cannot replace without reimplementing varwidth.
@daleif yes there isn't really an L3 typesetting layer yet so for things like varwidth just using as is makes sense, although probably some new templated paragraph trial module will arrive one day
@samcarter good comment: not my fault :-)
In l3doc, in the function env, can one list just one macro as TF and another as not? I have a macro that creates macros. It pulls keys out of a given prop var into a var it is given. Sometimes you just want the data (the non TF version) and sometimes you'd like to use it inside a TF branch.
Ands BTW, \cmd\par does not work ;-)
15:04
@DavidCarlisle That's the important thing :)
@samcarter of course
16:00
@daleif Well no, \cs{par}
16:40
@UlrikeFischer can I just trust you? (xcolor)
@DavidCarlisle :)
Otherwise dinner time, I'll look later.
@DavidCarlisle I trust my older self here (I made the code 2 years ago in a test branch and forgot it ...)
@DavidCarlisle but I think we should handle it. \usepackage[cmyk]{xcolor}\usepackage{tabularray} simply errors (I wonder why it is isn't reported more often ...).
@UlrikeFischer yes the test files look good but I haven't checked all the internal xcolor contortions
17:42
@DavidCarlisle there was a party:
user image
4
@UlrikeFischer toasting the main course before cooking it?
@DavidCarlisle we need @PauloCereda for the oh no.
18:00
@DavidCarlisle ooh no
(force of rabbit -- eeh habit)
 
2 hours later…
20:23
Any VSCode users here? Do you have a good scheme for expl3 syntax highlighting? The default from LaTeXWorkshop doesn't seem to do much, but maybe I'm not understanding how to configure it.
user image
7
Yay! Finally!
2
@Rmano and with such a nice squirrel.
@AlanMunn Er, yes I use VS Code, but not for TeX - that's what TeXworks is for ;)
20:41
@JosephWright you misspelled VIM twice. Though I don't understand the "but not for TeX" part.
@Rmano only 383000 to go!
@Skillmon s/vim/emacs/g
@DavidCarlisle g/emacs/d
@DavidCarlisle oh no
@Skillmon <3
@PauloCereda the advantage of posting "ooh no" is that it counts both as a message containing "ooh" and as "oh no".
@PauloCereda Also: Yay, carrots! Here, have one, too: <3
@Skillmon ooh rabbits are very good at efficiency
@Skillmon ooh ε>
20:57
@PauloCereda Oh, a Greek carrot
@JasperHabicht ooh :)
21:35
@JosephWright No, that's what TeXShop is for... But since I'm spending more time in VSCode these days, I'm contemplating using it for TeX too.

« first day (5070 days earlier)      last day (85 days later) »