Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
  • Jan 02 23:58
    @SethTisue banned @fakubishes:nerdsin.space
  • Dec 15 2021 05:01
    som-snytt commented #12516
  • Dec 15 2021 04:38
    SethTisue edited #1312
  • Dec 15 2021 04:38
    SethTisue opened #2273
  • Dec 15 2021 04:31
    jackkoenig opened #12516
  • Dec 15 2021 04:29
    SethTisue edited #1312
  • Dec 15 2021 04:28
    SethTisue edited #1312
  • Dec 15 2021 04:27
    SethTisue labeled #9831
  • Dec 15 2021 04:27
    scala-jenkins milestoned #9831
  • Dec 15 2021 04:27
    SethTisue labeled #9831
  • Dec 15 2021 04:27
    SethTisue opened #9831
  • Dec 15 2021 03:35
    som-snytt commented #11339
  • Dec 15 2021 03:27
    som-snytt labeled #12494
  • Dec 15 2021 03:07
    SethTisue edited #1312
  • Dec 15 2021 03:07
    SethTisue edited #1312
  • Dec 15 2021 03:05
    SethTisue edited #1312
  • Dec 15 2021 03:05
    SethTisue edited #1312
  • Dec 15 2021 03:05
    SethTisue edited #1312
  • Dec 15 2021 02:58
    SethTisue edited #1312
  • Dec 15 2021 02:58
    SethTisue synchronize #1312
Klas Segeljakt
@segeljakt
lol
They put a JVM in the JVM?
Rob Norris
@tpolecat
Hm I think Scala is not very good for metaprogramming, depending on how you define it I guess.
The Scala macro system is a garbage fire.
I think it's pretty good for embedded DSLs though.
Seth Tisue
@SethTisue
What metaprogramming support exists in Rust?
Long Cao
@longcao
Rust has macros and derivation, not sure how that lines up with what's available in Scala https://doc.rust-lang.org/book/second-edition/appendix-04-macros.html
Gabriel Claramunt
@gclaramunt
give me a collapsing tower of interpreters :)
Klas Segeljakt
@segeljakt
Rust’s macros are not on the same level as Scala’s I think
the macros define meta-variables which can be substituted for token-trees
Separators, like (), {}, [] create token trees
Rob Norris
@tpolecat
@gclaramunt o/
Harrison Houghton
@hrhino
Sounds kinda like Dotty LMS, then?
Elikar Ibn Mtsouko
@mtsouko_twitter
What is a good rule of thumb to follow when creating a trait vs abstract class? It seems there is a good amount of overlap in functionality ...
Li Haoyi
@lihaoyi
Python’s metaprogramming is the best :D
it even has hygienic quasiquotes
i think ichoran was talking about runtime java source code generation and compilation
although i have put a jvm inside a jvm and it’s awesome(ly slow)
Li Haoyi
@lihaoyi
spark does runtime java source generation and it’s not bad; compiles fast enough, JITs down to fast assembly and is easier for most people to debug than malformed blobs of bytecode
Klas Segeljakt
@segeljakt
What does hygienic mean?
with quasiquotes :D
Li Haoyi
@lihaoyi
it means when you call a function foo in a quasiquote, the macro generated code calls the function foo that is lexically in scope where the quasiquote was defined, not whatever random foo may be in scope at the macro callsite
Gabriel Claramunt
@gclaramunt
@mtsouko_twitter since you can only mix one abstract class and many traits as you want, I try to use traits as much as I can... abstract classes is for when you really want OO and is a reasonable hierarchy
Justin du Coeur, AKA Mark Waks
@jducoeur
Yeah -- I tend to favor traits by default, unless I want constructor parameters or have some other reason to specifically want a class.
Fabio Labella
@SystemFw
abstract classes can have parameters, that's a big point (it will change though). They are also better for bincompat in 2.11, if you're writing a library
Rob Norris
@tpolecat
Once trait parameters exist get ready for "why do we even have classes?"
Elikar Ibn Mtsouko
@mtsouko_twitter
What were the design reasons behind not giving traits parameters in the first place?
Rob Norris
@tpolecat
Same reasoning as class/interface in Java I imagine.
nafg
@nafg
@klassegeljakt if the reason is JVM startup time, you may want to look into scala-native or Graal VM
Klas Segeljakt
@segeljakt
Ok I'll take a look
If you know any more libs please let me know 🙏
Rob Norris
@tpolecat
I guess I should own up to writing a cursed DSL. Take it as a warning https://github.com/tpolecat/basic-dsl
Miguel Pérez Pasalodos
@mpasa
That is just great :smile:
Hanns Holger Rutz
@Sciss

Is there a way to constrain a Double to "just smaller than 1.0" such that (aDouble * aPositiveInt).toInt < aPositiveInt is guaranteed to hold? I.e. I don't want to perform that check after each multiplication. Like I can empircally get a very long literal:

assert (0.999999999999999944488848768742172978818416595 < 1)

assert ((0.999999999999999944488848768742172978818416595 * Int.MaxValue).toInt < Int.MaxValue)

But I wasn't patient enough to keep looking for the digits (seems there are more)

(Plus I'm afraid this might rely on the particular JVM, OS, or something)
Hanns Holger Rutz
@Sciss
This looks better: java.lang.Double.longBitsToDouble(0x3fefffffffffffffL) - comments?
Otto Chrons
@ochrons
yea, use longBitsToDouble, that's what I would use :)
Martijn Hoekstra
@martijnhoekstra
I think what you're asking is equivalent to the largest double smaller than 1.0. That's java.lang.Math.nextAfter(1.0, -1.0)
Otto Chrons
@ochrons
funny enough, I had exactly the same concern in my scheduling algorithm yesterday as it takes in 0.0 <= x < 1.0
and I needed to scale that to 0..N-1 where N is the number of input elements
Luciano
@lJoublanc

I'm writing a typeclass that should have an instance for a type refinement (using shapelss' Aux trick). However, I keep on getting this kind of compiler error when trying to instantiate the extension ops:

<console>:21: error: inferred kinds of the type arguments (Int,Int,res1.F[a],res1.O[k],res
1.U) do not conform to the expected kinds of the type parameters (type K,type V,type F,typ
e O,type U).
res1.F[a]'s type parameters do not match type F's expected parameters:
type F has one type parameter, but type F has one, res1.O[k]'s type parameters do not matc
h type O's expected parameters:
type O has one type parameter, but type O has one
       Series.Indexable.ops.toAllIndexableOps(res1)
                            ^

It seems nonsensical to me. Anybody got ideas how to fix this?

The above occurs when calling e.g. Series.Indexable.ops.toAllIndexableOps(res1)
Jan Bessai
@JanBessai
does it need to be called with rest1.F instead of res1.F[a] maybe?
can you try and make an example with minimal context?
abstract class A {
  protected def foo(): Unit
}

object X {
  def proxyA(x: A): A = new A { def foo() = x.foo() }
}
any idea why this yields
Test.scala:6: error: method foo in class A cannot be accessed in A
 Access to protected method foo not permitted because
 prefix type A does not conform to
 <$anon: A> where the access takes place
  def proxyA(x: A): A = new A { def foo() = x.foo() }
I feel this is confusing, since access happens within a subclass of A
Jan Bessai
@JanBessai
this is fixable with protected [somePackage], but I'm not sure I understand why it doesn't work without that
Martijn Hoekstra
@martijnhoekstra
I think (I'm not sure) that what you wrote is the moral equivalent to object X = { def foo() = x.foo }; new A with X
at least, that's the only way I can make sense of it, and it sounds sort of kind of consistent
no, that's nonsense