Unicode in code causes parsing to stop early

[Oblynx]

Given a code chunk

Weave.CodeChunk("\nΝᵢₙ= 2\nNₛₚ= 2\nszᵢₙ= (5,5)\nszₛₚ= (6,9)\n", 1, 0, 16, "", Dict{Symbol,Any}(:dpi=>96,:fig=>true,:fig_pos=>nothing,:out_width=>nothing,:label=>nothing,:eval=>true,:wrap=>true,:fig_env=>nothing,:term=>false,:hold=>false…), "", "", AbstractString[], Weave.ChunkOutput[])

which contains Unicode characters, calling Weave.(chunk,report,Main) causes parsing to stop early:

Weave.CodeChunk("\nΝᵢₙ= 2\nNₛₚ= 2\nszᵢₙ= (5,5)\nszₛₚ= (6,9)\n", 1, 0, 16, "", Dict{Symbol,Any}(:dpi=>96,:fig=>true,:fig_pos=>nothing,:out_width=>"\\linewidth",:label=>nothing,:eval=>true,:wrap=>true,:fig_env=>nothing,:term=>false,:hold=>false…), "", "", AbstractString[], Weave.ChunkOutput[ChunkOutput("Νᵢₙ= 2\n", "", "", "", AbstractString[]), ChunkOutput("Nₛₚ= 2\n", "", "", "", AbstractString[]), ChunkOutput("szᵢₙ= (5,5)\n", "", "", "", AbstractString[])])

This is due to incorrect UTF-8 string handling in https://github.com/mpastell/Weave.jl/blob/master/src/run.jl#L259
UTF-8 strings don't always encode each character in a single byte.

length counts the number of UTF-8 characters in the string, whereas pos as updated by Meta.parse counts the number of bytes.
Therefore, https://github.com/mpastell/Weave.jl/blob/master/src/run.jl#L259
should be n = sizeof(input).