At the time method declarations evaluated to nothing. Because of this, we changed it so method declarations evaluate to the generic function that the method is being added to.
On Fri, Mar 13, 2015 at 10:31 AM, Matteo Fasiolo <[email protected]> wrote: > Ok, but the closure (the first version) in the 2012 post was not working > > julia> typeof(Rrss) > Nothing > > while now it does > > typeof(gridfgen([1.; 3.], [.2; 4.])) > Function > > or maybe I am missing something. > > > > On Friday, March 13, 2015 at 2:02:34 PM UTC, Stefan Karpinski wrote: >> >> Closures have always worked, they just have some performance issues still. >> >> On Fri, Mar 13, 2015 at 9:05 AM, Matteo Fasiolo <[email protected]> >> wrote: >> >>> Hi All, >>> >>> This (2012) discussion <https://gist.github.com/dmbates/3939427> seems >>> to suggest that generic functions cannot be used within closures in Julia. >>> But this seems to work now >>> >>> function creator(y) >>> >>> function power(x) return x .^ y end >>> >>> function myNorm(x) return sum( power(x) ) end >>> >>> return power, myNorm >>> >>> end >>> >>> julia> (a, b) = creator(2.) >>> (power,myNorm) >>> >>> julia> a(2) >>> 4.0 >>> >>> julia> b([1. 2.]) >>> 5.0 >>> >>> I use generic functions because they support keyword optional arguments >>> and I want to do things >>> such as >>> >>> function a(y, extra...) function b(x, extra...; z = true) "something" >>> end end >>> >>> which seems to work in terms of not "mixing up" the ... with the >>> optional arguments. >>> >>> Notice that I am mainly an R user, currently experimenting with Julia, >>> so my approach might be completely wrong. >>> >>> [1]: https://gist.github.com/dmbates/3939427 >>> >> >>
