Not if we took the {} syntax for empty dict, but this is becoming a rather
disruptive change.On Fri, May 2, 2014 at 11:48 AM, Mike Nolta <[email protected]> wrote: > But then it's unclear if (A,(B,C))[] is Dict{A,Dict{B,C}} or Dict{A,(B,C)}. > > -Mike > > On Fri, May 2, 2014 at 11:26 AM, Stefan Karpinski <[email protected]> > wrote: > > Yeah, that I've got nothing for unless we did this: > > > > (K,V)[ k => f(v) for (k,v) in d ] > > > > > > On Fri, May 2, 2014 at 9:35 AM, Carlo Baldassi <[email protected]> > > wrote: > >> > >> But what's your suggestion about typed dict comprehensions? > >> > >> > >> On Thursday, May 1, 2014 7:11:59 PM UTC+2, Stefan Karpinski wrote: > >>> > >>> Is be more in favor of deprecating the (K=>V)[k=>v] syntax and just > using > >>> keyword args like this Dict{K,V}(k=v). Having so many syntaxes for > this is > >>> confusing and it's not like the (K=>V)[k=>v] syntax is a thing of > beauty. > >>> > >>> On May 1, 2014, at 12:50 PM, Jameson Nash <[email protected]> wrote: > >>> > >>> The => is special here for the parser. What you want is > >>> (Int=>Dict{Int, Int})[ ] > >>> > >>> However, it's possible your alternative syntax could be made to work. > >>> > >>> On Thursday, May 1, 2014, thom lake <[email protected]> wrote: > >>>> > >>>> Both of these work > >>>> > >>>> julia> Dict{Int,Int}() > >>>> Dict{Int64,Int64}() > >>>> > >>>> julia> (Int=>Int)[] > >>>> Dict{Int64,Int64}() > >>>> > >>>> So does this > >>>> > >>>> julia> Dict{Int,Dict{Int,Int}}() > >>>> Dict{Int64,Dict{Int64,Int64}}() > >>>> > >>>> This doesn't > >>>> > >>>> julia> (Int=>(Int=>Int))[] > >>>> ERROR: unsupported or misplaced expression => > >>>> > >>>> Any particular reason? Am I doing something silly? > >>>> > > >
