This is a demo of broken traceback when reraising exception in python 2.x.

py2-broken-traceback.py

"""
This is a demo of broken traceback
when reraising exception in python 2.x.

In this code, "raise e" makes python
lost original traceback, attached to the
exception. It outputs:

Traceback (most recent call last):
  File "/tmp/test.py", line 19, in <module>
    another()
  File "/tmp/test.py", line 14, in another
    raise e
RuntimeError: blah

But if you'll fix this code and leave only
"raise" alone, then it will show original
traceback:

Traceback (most recent call last):
  File "/tmp/test.py", line 19, in <module>
    another()
  File "/tmp/test.py", line 12, in another
    some()
  File "/tmp/test.py", line 8, in some
    return bar()
  File "/tmp/test.py", line 5, in bar
    return foo()
  File "/tmp/test.py", line 2, in foo
    raise RuntimeError("blah")
RuntimeError: blah
"""


def foo():
    raise RuntimeError("blah")

def bar():
    return foo()

def some():
    return bar()

def another():
    try:
        some()
    except Exception as e:
        raise e


if __name__ == '__main__':
    another()

Давно Бикинг писал статью о вариантах обхода такой ситуации. Одной из идей была, что, ежели e никак не дополняется, то голый raise должен справиться с задачей.