Description of problem:

When pythonflow output traces, TypeError happens.


Version-Release number of selected component (if applicable):

bcc-0.6.1-2.fc28.x86_64
bcc-doc-0.6.1-2.fc28.noarch
python3-bcc-0.6.1-2.fc28.x86_64
bcc-lua-0.6.1-2.fc28.x86_64
bcc-tools-0.6.1-2.fc28.x86_64


How reproducible:

Always


Steps to Reproduce:
1. $ python
2. $ sudo /usr/share/bcc/tools/pythonflow <PID>
3. type Enter in python REPL

Actual results:

$ sudo ./pythonflow -C '<stdin>' 26904
Tracing method calls in python process 26904... Ctrl-C to quit.
CPU PID    TID    TIME(us) METHOD
Traceback (most recent call last):
  File "_ctypes/callbacks.c", line 234, in 'calling callback function'
  File "/usr/lib/python3.6/site-packages/bcc/table.py", line 572, in raw_cb_
    callback(cpu, data, size)
  File "./lib/uflow", line 187, in print_event
    ("  " * (depth - 1)) + direction + event.clazz + "." + event.method))
TypeError: must be str, not bytes

(and this traceback repeats)

Expected results:

$ sudo ./pythonflow -C '<stdin>' 27058
Tracing method calls in python process 27058... Ctrl-C to quit.
CPU PID    TID    TIME(us) METHOD
3   27058  27058  3.246    -> b'<stdin>'.b'<module>'               
3   27058  27058  3.246    -> b'<stdin>'.b'<module>'               

Additional info:

In lib/uflow's print_event function, event.clazz and event.method are bytes.
Putting str() for casting works. 

    print("%-3d %-6d %-6d %-8.3f %-40s" % (cpu, event.pid >> 32,
        event.pid & 0xFFFFFFFF, time.time() - start_ts,
        ("  " * (depth - 1)) + direction + str(event.clazz) + "." + str(event.method)))

ctypes.cast maps char * into bytes in python3.  
I expect this problem was caused by Python 2/3 difference, and some other tools have similar problems.