StackOverflow when displaying recursive output causes segfault

[pito-svk]

Hi, I am not sure if this is a bug, but it looks like it is.

Simple:
you have an array let's say

a = []
push!(a, a)

• it causes me chain of Segmentation fault errors and terminate REPL and later when I write something in terminal - my input and output is in bold for some time.
  

Please correct me, if this isn't a bug.

Thanks.

[Keno]

It's a stack overflow, but that's fine, it should just throw an error.

[jakebolewski]

This works for me on 0.4. @pito-svk are you on windows? Can you post the output of versioninfo()?

julia> a = []
0-element Array{Any,1}

julia> push!(a, a)
1-element Array{Any,1}:
Error showing value of type Array{Any,1}:
ERROR: StackOverflowError:

[ivarne]

Note that the bug likely is in the printing of the resulting recursive array, not in creating it.

[pito-svk]

@jakebolewski no, ubuntu:

Julia Version 0.4.0-dev+3533
Commit f875a54 (2015-02-22 20:56 UTC)
DEBUG build
Platform Info:
  System: Linux (x86_64-linux-gnu)
  CPU: Intel(R) Core(TM) i7-4510U CPU @ 2.00GHz
  WORD_SIZE: 64
  BLAS: libopenblas (USE64BITINT DYNAMIC_ARCH NO_AFFINITY Haswell)
  LAPACK: libopenblas
  LIBM: libopenlibm
  LLVM: libLLVM-3.3

@ivarne right (in script):

a = []
push!(a, a)
print(a) # this causes error

or this:

a = []
push!(a, a)
@printf("%d", a) # also

[jakebolewski]

Although this is a legit stack overflow bug, we could print something like

julia> a = []
0-element Array{Any,1}

julia> push!(a, a)
1-element Array{Any,1}:
 Any[#= circular reference =#]

here to avoid the overflow just by checking is(x, itr) in show_delim_array

[ihnorton]

This may be specific to llvm 3.3; I can reproduce there, but on llvm 3.5 I get a proper error.

[pito-svk]

@jakebolewski Right. For example python prints this:

[[...]]

...and javascript:

[ [Circular] ]

[jiahao]

Related: #3831

[Keno]

This one should probably be closed already, given that the original issue was fixed to not overflow the stack.