REPL: history search crash with TAB

[cmcbride]

Found accidentally, but it looks like history search and tab completion don't play well together. I'm assuming this is a bug...

When using REPL history search (CTRL-R), hitting TAB will throw an error that ends the REPL session (see below). It seems consistent and TAB does not have to be the first character.

julia> versioninfo()
Julia Version 0.3.1-pre+23
Commit 77c15f9* (2014-08-31 17:39 UTC)
Platform Info:
  System: Darwin (x86_64-apple-darwin13.3.0)
  CPU: Intel(R) Core(TM) i5-4288U CPU @ 2.60GHz
  WORD_SIZE: 64
  BLAS: libopenblas (USE64BITINT DYNAMIC_ARCH NO_AFFINITY Haswell)
  LAPACK: libopenblas
  LIBM: libopenlibm
  LLVM: libLLVM-3.3

(reverse-i-search)`': ERROR: `input_string` has no method matching input_string(::SearchState)
 in complete_line at REPL.jl:294
 in complete_line at LineEdit.jl:988
 in complete_line at LineEdit.jl:139
 in anonymous at LineEdit.jl:1100
 in prompt! at ./LineEdit.jl:1397
 in run_interface at ./LineEdit.jl:1372
 in run_interface at /Users/cameron/code/julia/julia/usr/lib/julia/sys.dylib
 in run_frontend at ./REPL.jl:819
 in run_repl at ./REPL.jl:170
 in _start at ./client.jl:399
 in _start at /Users/cameron/code/julia/julia/usr/lib/julia/sys.dylib

[flpf]

Hi.

I think I found a similar problem.
In a Julia-REPL without loaded packages tab-completion of functions results in

julia> writedlm( [...Hitting Tab gives..]

writedlm(io::IO,a::Union(AbstractArray{T,1},AbstractArray{T,2}),dlm) at datafmt.jl:502
writedlm{T}(io::IO,a::AbstractArray{T,0},dlm) at datafmt.jl:518
writedlm(io::IO,a::AbstractArray{T,N},dlm) at datafmt.jl:521
writedlm(io::IO,itr,dlm) at datafmt.jl:532
writedlm(fname::String,a,dlm) at datafmt.jl:549
writedlm(io,a) at datafmt.jl:554

After loading a package (I've tried PyPlot, Gadfly, Winston) julia crashes on tab-completion

julia> using Winston

julia> writedlm( [...Hitting Tab gives...] ERROR: type UnionType has no field body
 in show at show.jl:80 (repeats 2 times)
 in print_to_string_3B_504 at /usr/bin/../lib/julia/sys.so
 in argtype_decl at methodshow.jl:18
 in arg_decl_parts at methodshow.jl:30
 in show at methodshow.jl:36
 in print_to_string at string.jl:24
 in string at string.jl:31
 in complete_methods_3B_1834 at /usr/bin/../lib/julia/sys.so
 in completions_3B_1827 at /usr/bin/../lib/julia/sys.so (repeats 2 times)
 in complete_line at REPL.jl:279
 in complete_line at LineEdit.jl:141
 in complete_line at LineEdit.jl:139
 in anonymous at LineEdit.jl:1175
 in anonymous at LineEdit.jl:1197
 in prompt_21__3B_1846 at /usr/bin/../lib/julia/sys.so
 in run_interface_3B_1845 at /usr/bin/../lib/julia/sys.so (repeats 2 times)
 in run_frontend_3B_1797 at /usr/bin/../lib/julia/sys.so
 in run_repl_3B_1797 at /usr/bin/../lib/julia/sys.so
 in _start_3B_1707 at /usr/bin/../lib/julia/sys.so (repeats 2 times)

julia  16.55s user 0.23s system 20% cpu 1:23.12 total

I tried a working .4 release and the stable .3 both giving me a similar error.

julia> versioninfo()
Julia Version 0.3.0
Commit 7681878 (2014-08-20 20:43 UTC)
Platform Info:
  System: Linux (x86_64-unknown-linux-gnu)
  CPU: Intel(R) Core(TM) i5-2467M CPU @ 1.60GHz
  WORD_SIZE: 64
  BLAS: libblas
  LAPACK: liblapack
  LIBM: libm
  LLVM: libLLVM-3.3

I am not sure if this is a problem of the packages or a REPL bug so I added it here.

[ivarne]

@flpf I think that issue is beeing tracked at #8163. I tried your example on yesterdays master, and it behaves as you described.

[flpf]

@ivarne I tried the same with a couple of other (smaller) packages like AudioIO, PyCall, DataStructures.... and did not get that error. So I guess the error occurs only with some
packages.

I don't know if this helps in any way but I thought I let you know.
If there is anything I can do to debug that error some more please let me now. But from reading
#8163 I think that this problem is not really traceable for me without diving into julias JITcompile
system.

[flpf]

@ivarne, @cmcbride.
I did some debugging and found out that my problem originates in the FixedPointNumbers.jl package.
When commenting out the show function in the source file "ufixed.jl" (lines 156-165),
the error is gone.

# Show
function show(io::IO, x::Ufixed)
    print(io, "Ufixed", nbitsfrac(typeof(x)))
    print(io, "(")
    showcompact(io, x)
    print(io, ")")
end
showcompact(io::IO, x::Ufixed) = show(io, round(convert(Float64,x), iceil(nbitsfrac(typeof(x))/_log2_10)))

show{T<:Ufixed}(io::IO, ::Type{T}) = print(io, "Ufixed", nbitsfrac(T))

Maybe you can check if this is the source of the error in your systems as well.
I will file an issue under the FixedPointNumbers.jl package.

[flpf]

For me the error is fixed after updating FixedPointNumbers.jl.

[Keno]

I can't reproduce this anymore and the REPL code has had a lot of work.