« first day (5151 days earlier)      last day (47 days later) » 

09:52
@cfr :)
@cfr Lucky you (!)
@mickep Double underscore means 'private/internal' - only intended for the person writing the package
@JosephWright And if a user needs to do some arithmetic, should they use \dim_set:Nn?
@mickep Well they should be using document level structures, so \setlength or some key =<dim-value> interface
E.g. siunitx has a setting for table column width, which is a key called ... table-column-width and takes a dimexpr, but the user just provides the expression, no \dimexpr involved for them
@JosephWright OK, but \setlength did not use \dimexpr, so...
So, isn't \dimexpr 1cm + 2mm \relax really the easiest for users?
@mickep Ideally, interfaces get updated so users can directly use expressions, but we have \dimeval{...} for places that doesn't apply
10:12
@JosephWright Ah, and that uses a \relax at the end...
@mickep \dimeval is a wrapper around \dimexpr .. \relax, yes - but with LaTeX syntax
@JosephWright Does that impact performance? You have to pickup an argument. Do you internally use \dimexpr or do you in fact use that \dimeval thing in your low level code?
> \dimeval=\long macro:
#1->\dim_use:N \__dim_eval:w #1\__dim_eval_end: .
@mickep ^^^
@mickep Which are all \let to primitives - \the\dimexpr#1\relax
@mickep There's always a performance hit for a wrapper, but it's a question of the degree - here, you'd not see it at all
(In the same way, an expl3 'tl' op like \tl_set:Nn is \edef#1{\unexpanded{#2}} and very slightly slower than \def#1{#2}, but more flexible)
@JosephWright OK, and it does not add up having all lowlevel thingies like that? I assume you have tested it and that my fear is just wrong.
10:27
@mickep oh yes, we test! Expansion is fast, so a one level wrapper is not an issue
@mickep We worry about tight loops where that’s important: Frank is very hot on this!
@JosephWright OK, very good!
(I just compiled my book, and \dimexpr is used over 200000 times...)
@mickep Sure, and like I said, for expl3 we have expressions universally - we talk a lot about consistency vs performance, but when you trace, etc., these bits of consistency add only ms to run times and make user life easier, so are definitely the way to go
 
7 hours later…
17:11
Has anyone used VSCode Live as a way to collaboratively edit LaTeX documents? It seems like it should be possible and then integrate compilation via GitHub actions.
 
2 hours later…
19:14
@AlanMunn AFAIK you can simply use LaTeX Workshop and compile the document online, I didn't try it personally though
@EmanueleNardi What do you mean by "compile the document online"?
 
1 hour later…
20:37
I just connected my brain we are talking about two different things
20:59
@EmanueleNardi ok that's what I thought. :)
Reporting back on my question: you can definitely use VSCode Live Share to collaboratively edit LaTeX documents, but with one important caveat: the sharer can't view the compiled PDF file, which makes things a lot worse. But for people who aren't too concerned about that, it's nice to know that collaborative editing is possible with a local TeX distribution. The inability to display PDFs or other binary files in Live seems to be a bug or missing feature.
cfr
cfr
21:42
@JosephWright it wouldn't have been so bad if I'd not been in the middle of a (landline) phone call.
among 'loose items' the Met Office recommend you secure in case of a red weather warning for wind, they include sheds. are we supposed to tie them down with string or what?
 
1 hour later…
22:53
@cfr I do. Can I help?

« first day (5151 days earlier)      last day (47 days later) »