Varint decoding (in Ruby) fails with "invalid byte sequence in UTF-8"

[ghost]

Decoding a GTFS-realtime message fails for me with a ProtocolBuffers::DecodeError, "error parsing message". Debugging using "ruby -rdebug" (see below) reveals varint.decode is failing trying to read from the input stream using io.getc (varint.rb:30).

Looks to me the code here should be trying to read a byte and not a UTF-8 character, so the error makes sense. Should this line be "io.read(1).ord" instead? That change fixes the problem for me and the message is parsed successfully. (Seems both Ruby 1.8 and 1.9 have an IO.read method, and the method is provided by the LimitedIO class as well.)

---

/usr/local/lib/ruby/gems/1.9.1/gems/ruby-protocol-buffers-1.3.2/lib/protocol_buffers/runtime/varint.rb:30: ' (NilClass) from /usr/local/lib/ruby/gems/1.9.1/gems/ruby-protocol-buffers-1.3.2/lib/protocol_buffers/runtime/message.rb:274:inparse'
from /usr/local/lib/ruby/gems/1.9.1/gems/ruby-protocol-buffers-1.3.2/lib/protocol_buffers/runtime/message.rb:280:in parse' from /usr/local/lib/ruby/gems/1.9.1/gems/ruby-protocol-buffers-1.3.2/lib/protocol_buffers/runtime/field.rb:601:indeserialize'
from /usr/local/lib/ruby/gems/1.9.1/gems/ruby-protocol-buffers-1.3.2/lib/protocol_buffers/runtime/decoder.rb:57:in decode' from /usr/local/lib/ruby/gems/1.9.1/gems/ruby-protocol-buffers-1.3.2/lib/protocol_buffers/runtime/message.rb:274:inparse'
from /usr/local/lib/ruby/gems/1.9.1/gems/ruby-protocol-buffers-1.3.2/lib/protocol_buffers/runtime/message.rb:280:in parse' from ./test.rb:6:inblock in

'
from ./test.rb:5:in open' from ./test.rb:5:in'
/usr/local/lib/ruby/gems/1.9.1/gems/ruby-protocol-buffers-1.3.2/lib/protocol_buffers/runtime/varint.rb:30: byte = io.getc.ord
(rdb:1) c
/usr/local/lib/ruby/gems/1.9.1/gems/ruby-protocol-buffers-1.3.2/lib/protocol_buffers/runtime/decoder.rb:88: invalid byte sequence in UTF-8' (ArgumentError) from /usr/local/lib/ruby/gems/1.9.1/gems/ruby-protocol-buffers-1.3.2/lib/protocol_buffers/runtime/message.rb:280:inparse'
from /usr/local/lib/ruby/gems/1.9.1/gems/ruby-protocol-buffers-1.3.2/lib/protocol_buffers/runtime/field.rb:601:in deserialize' from /usr/local/lib/ruby/gems/1.9.1/gems/ruby-protocol-buffers-1.3.2/lib/protocol_buffers/runtime/decoder.rb:57:indecode'
from /usr/local/lib/ruby/gems/1.9.1/gems/ruby-protocol-buffers-1.3.2/lib/protocol_buffers/runtime/message.rb:274:in parse' from /usr/local/lib/ruby/gems/1.9.1/gems/ruby-protocol-buffers-1.3.2/lib/protocol_buffers/runtime/message.rb:280:inparse'
from ./test.rb:6:in block in <main>' from ./test.rb:5:inopen'
from ./test.rb:5:in `'
/usr/local/lib/ruby/gems/1.9.1/gems/ruby-protocol-buffers-1.3.2/lib/protocol_buffers/runtime/decoder.rb:88: raise(DecodeError, "error parsing message")

[codekitchen]

Oh wow, I wonder how in the world that's slipped through for so long. I'm guessing we must typically get the strings ascii/binary encoded, so there are no multi-byte characters. Thanks for the report, I'll work on a fix soon if you don't.