You can’t do that hygenically because that specification is not hygenic. But it should be pretty easy by making that require have the context of the (module ...) form.

@samdphillips have you done aoc day 2?

I think the answer is yes, because he is #1 now on the leaderboard

It’s K? Really (I’m not disbelieving you… I’m just astonished I haven’t spotted that)

thanks all

what leaderboard?

Is this a private leaderboard?

Yes, you can find it in the pinned message in #advent-of-code

hi all, how much is my compile-time performance going to hurt if i implement basically a whole grammar using syntax-class and splicing-syntax-class? am I looking at a greater-than-linear/constant slowdown if loads of patterns use each other, and themselves recursively?

obviously, recursive/nested patterns will take longer as appropriate. I guess I am asking if under the hood somewhere a >=n2 complexity is going to bite me.

@ryanc ^

@pezi_pink It depends on the grammar. For a linear grammar, the kind that you would normally feed to a LALR(1) parser generator (or similar), it could go very badly. For example, a left-recursive splicing syntax class (with the recursive case first) will just loop forever, I think. For a tree-structured grammar that you could parse with match if not for the syntax nodes, it should work well, but you might need to put #:commit declarations on your syntax classes (but not the splicing syntax classes).

@ryanc thanks for the tips, much appreciated. Ps. I watched one of your videos recently on this stuff, I didn’t know about (~?) and (~@) which have made my life a lot easier, thanks!

Is David’s “Felt” and “Brush” languages the same? (See the two top links at https://racket-stories.com/ )

Yeah I’m on Pacific Time so I have a bit of an advantage. Generally by the time I get to the 10th I cannot complete them as quickly.

I’m holding my GitHub pushes at least until 8am ish Pacific.

I am confused by the output of this: > (regexp-match #px"([:word:])(.*)" "foo") '("oo" "o" "o") I was expecting the f to show up in the result as well.

What am I missing?

To clarify: did running xattr -d on the individual executables not work? Or is it just more convenient to do the whole directory at once?

FWIW, I’ve been (ab)using syntax-parse to do something like what you’re asking in https://github.com/Bogdanp/racket-review/blob/master/lint.rkt#L520 . I haven’t encountered any perf. issues thus far

I missed that [:word:] belongs inside a [ ] …

The solution to everything is more parenthesis!

I’m on pacific time too, but I always participated at 3am…

I’ve discovered from the programminglanguages discord server that it is possible to bridge discord server channels to IRC(freenode) is this desirable? (I’m assuming slack/irc bridge gone forever)

haha

Can you execve(2) from a Racket program?

Like is there a good builtin/library to do it.

Yes, see system and process and other functions with similar names

Those don’t quite fit what I’m looking for (unless there is an option I’m missing), I need to exec a new process and racket would no longer be running.

Does someone have a link to the motivation behind creating the CS VM for Racket?

Never mind, found it: https://groups.google.com/forum/#!msg/racket-dev/2BV3ElyfF8Y/4RSd3XbECAAJ

find ./ -name \*.c