Uploaded image for project: 'Apache QPID Interoperability Test Suite'
  1. Apache QPID Interoperability Test Suite
  2. QPIDIT-68

Amqp Types test for binary sends only printable strings

    Details

    • Type: Bug
    • Status: Open
    • Priority: Major
    • Resolution: Unresolved
    • Affects Version/s: 0.1.0
    • Fix Version/s: None
    • Component/s: AMQP Types Test
    • Labels:
      None

      Description

      In the Amqp Types test the binary data is defined in Json. One of the defs is huge but the first four are defined as:

              'binary': [bytes(),
                         bytes(12345),
                         b'Hello, world',
                         b'\\x01\\x02\\x03\\x04\\x05abcde\\x80\\x81\\xfe\\xff',
      

      Adding a little code to instrument the cli argument containing the Json and running a wireshark snoop reveals how these four strings are handled by ProtonCpp.

      Note that jira makes it hard to render a backslash character. In these tables the <backslash> token represents on ascii character.

      Json command line arg

      The json generated by python arrives at ProtonCpp arg four as a string

      position character hex
      0 '[' 0x5b
      1 '"' 0x22
      2 '"' 0x22
      3 ',' 0x2c
      4 ' ' 0x20
      5 '"' 0x22
      6 '1' 0x31
      7 '2' 0x32
      8 '3' 0x33
      9 '4' 0x34
      10 '5' 0x35
      11 '"' 0x22
      12 ',' 0x2c
      13 ' ' 0x20
      14 '"' 0x22
      15 'H' 0x48
      16 'e' 0x65
      17 'l' 0x6c
      18 'l' 0x6c
      19 'o' 0x6f
      20 ',' 0x2c
      21 ' ' 0x20
      22 'w' 0x77
      23 'o' 0x6f
      24 'r' 0x72
      25 'l' 0x6c
      26 'd' 0x64
      27 '"' 0x22
      28 ',' 0x2c
      29 ' ' 0x20
      30 '"' 0x22
      31 'backslash' 0x5c
      32 'backslash' 0x5c
      33 'x' 0x78
      34 '0' 0x30
      35 '1' 0x31
      36 'backslash' 0x5c
      37 'backslash' 0x5c
      38 'x' 0x78
      39 '0' 0x30
      40 '2' 0x32
      41 'backslash' 0x5c
      42 'backslash' 0x5c
      43 'x' 0x78
      44 '0' 0x30
      45 '3' 0x33
      46 'backslash' 0x5c
      47 'backslash' 0x5c
      48 'x' 0x78
      49 '0' 0x30
      50 '4' 0x34
      51 'backslash' 0x5c
      52 'backslash' 0x5c
      53 'x' 0x78
      54 '0' 0x30
      55 '5' 0x35
      56 'a' 0x61
      57 'b' 0x62
      58 'c' 0x63
      59 'd' 0x64
      60 'e' 0x65
      61 'backslash' 0x5c
      62 'backslash' 0x5c
      63 'x' 0x78
      64 '8' 0x38
      65 '0' 0x30
      66 'backslash' 0x5c
      67 'backslash' 0x5c
      68 'x' 0x78
      69 '8' 0x38
      70 '1' 0x31
      71 'backslash' 0x5c
      72 'backslash' 0x5c
      73 'x' 0x78
      74 'f' 0x66
      75 'e' 0x65
      76 'backslash' 0x5c
      77 'backslash' 0x5c
      78 'x' 0x78
      79 'f' 0x66
      80 'f' 0x66
      81 '"' 0x22
      82 ',' 0x2c

      Over the wire

      The ProtonCpp does its work and sends messages over the wire.

      • The messages are encoded with type A0 (binary:vbin8) and an appropriate length.
      Message Content
      0 "" <empty message>
      1 "12345" <five bytes with ascii encodings of the numbers>
      2 "Hello, world"
      3 "<backslash>x01<backslash>x02<backslash>x03<backslash>x05<backslash>x05 abcd <backslash>x80<backslash>x81<backslash>xfe<backslash>xff

      The intent of the test appears to have binary, non-printable bytes (0x01, 0x02, 0xff) in the messages but that's not what happens.

      Maybe this is something simple like having the double backslashes in the test python code.

        Attachments

          Activity

            People

            • Assignee:
              kpvdr Kim van der Riet
              Reporter:
              chug Chuck Rolke
            • Votes:
              0 Vote for this issue
              Watchers:
              1 Start watching this issue

              Dates

              • Created:
                Updated: