ozra
Feature request: Splat support in block arguments.
a-function(...args) -> yield args
a-function(1, "", false) ~> |...args| puts args
Currently not possible in Crystal.
Slight correction to the example (and generally, I'd recommend using "say" for informal stdout shout outs - the name reflect the use better imo, but no guns held to the head here ;-) ):
a-function 1, "", false, (...args) ~> say args
Anyway, I'd be happy to see some more "practical" examples, because I can't see myself in use of that. Care to enlighten? Also - if you've already posted issues, for instance in crystal issues, links there are acceptable, since I'll undoubtedly make contributions upstream too.
Logically, if you need variadic methods, then you also need variadic blocks.
Looking over there, my instant reaction was "that should be a macro", which I then saw you did use.
It seems you're coding in a very "dynamic lang mindset". All things are solvable by attacking it from a new perspective.
This is a language primarily targeted at performant code. Statically type checked - because it generates faster code, and is scientifically more productive and safer. Those are the pillars.
Trying to "dynamify" that is like building a ladder reaching around a wall sideways instead of just opening the door and go through.
soft lambdas != methods, so there's really no logic "carry over".
If there really is a good use case, that isn't already solvable equally or better, then it's time to:
- spend all those hours away from family, friends and life matters (needed to code any feature)
- increasing complexity of the compiler, (unavoidable with any feature)
- reducing free syntactical combinations (that might turn out to be better needed for another feature in the future.)
- etc.
There are always costs! And there are many choices of where to spend, so we require real motivation, not implied logics.