isdefined fails for Int32 index

[joachimdahl]

This is probably unintended behaviour.

julia> versioninfo()
Julia Version 1.0.3
Commit 099e826241 (2018-12-18 01:34 UTC)
Platform Info:
  OS: Linux (x86_64-pc-linux-gnu)
  CPU: Intel(R) Xeon(R) CPU E5-2470 v2 @ 2.40GHz
  WORD_SIZE: 64
  LIBM: libopenlibm
  LLVM: libLLVM-6.0.0 (ORCJIT, ivybridge)

julia> isdefined(Array{Any}(undef,1),Int64(1))
false

julia> isdefined(Array{Any}(undef,1),Int32(1))
ERROR: TypeError: in isdefined, expected Symbol, got Int32
Stacktrace:
 [1] top-level scope at none:0

[andreasnoack]

Indices are restricted to Int in many functions. That limits the number of specializations. We have discussed this a couple of times and I'm not sure what our current thinking is wrt to allowing more input types.

[JeffBezanson]

See the description of isdefined:

    isdefined(m::Module, s::Symbol)
    isdefined(object, s::Symbol)
    isdefined(object, index::Int)

Tests whether an assignable location is defined. The arguments can be a module and a symbol
or a composite object and field name (as a symbol) or index.

In other words, it only works on modules or object fields, not array elements. Arrays use isassigned.

[mbauman]

Right, it won't do what you want even if you passed it Int indices, either:

julia> isdefined(Any[1], 1)
false

IMO this is all the more reason it shouldn't do the array-like to_indices conversion of that index.
Perhaps the docstring should point to isassigned to make this a bit clearer?

[JeffBezanson]

+1 to linking docs to isassigned.