Re-evaluating parameterized type definitions erros

[mauro3]

I noticed this inconsistency (both in 0.4 and 0.5):

julia> type B
       a
       end

julia> type B
       a
       end

julia> type A{T}
       a
       end

julia> type A{T}
       a
       end
ERROR: invalid redefinition of constant A

Would be nice if the second would not error.

[KristofferC]

A little bit of discussion here: https://groups.google.com/forum/?nomobile=true#!topic/julia-users/OM-NJMkkVoY

[mauro3]

I thought this must have been discussed somewhere.

[cstjean]

@vtjnash It seems that this code now fails:

               _
   _       _ _(_)_     |  A fresh approach to technical computing
  (_)     | (_) (_)    |  Documentation: http://docs.julialang.org
   _ _   _| |_  __ _   |  Type "?help" for help.
  | | | | | | |/ _` |  |
  | | |_| | | | (_| |  |  Version 0.6.0-dev.2778 (2017-02-12 11:20 UTC)
 _/ |\__'_|_|_|\__'_|  |  Commit 71f6bc9 (0 days old master)
|__/                   |  x86_64-apple-darwin13.4.0

julia> immutable X{T}
          a::T
       end

julia> immutable X{T}
          a::T
       end
ERROR: invalid redefinition of constant X

Is that a regression? It's a bummer for cstjean/ClobberingReload.jl#1 (comment) cc. @rdeits

[rdeits]

awesome, thanks @JeffBezanson and @cstjean

[cstjean]

Some definitions still cannot be reevaluated:

julia> const TupleOf{T} = NTuple{N, T} where {N}
Tuple{Vararg{T,N}} where N where T

julia> const TupleOf{T} = NTuple{N, T} where {N}
ERROR: invalid redefinition of constant TupleOf

[yuyichao]

That's not a type definition.

[cstjean]

Then parametric type aliases cannot be reevaluated. MWE: MyVector{N} = Vector{N}. Should this be a new issue? I couldn't find anything relevant.

[JeffBezanson]

We allow re-assigning constants if the new value is identical. We could expand the definition of "identical" for this case.

[cstjean]

Similar to the type-alias case, this type definition cannot be reevaluated:

struct GG{T<:(Vector{U} where U <: Integer)} end