MagicMock 'protocol' methods workaround for 0.7.2?

[GoogleCodeExporter]

What steps will reproduce the problem?
1. m = magicMock()
2. m[1]  # not a magicMock

This is fixed in rev 481, but is there a way to simulate it?  just wait until 
0.8?

Original issue reported on code.google.com by gregg.l...@lolapps.com on 6 Oct 2011 at 9:50

[GoogleCodeExporter]

This should do it:

    m.__getitem__ = MagicMock() 

Original comment by fuzzyman on 6 Oct 2011 at 10:14

• Changed state: Invalid

[GoogleCodeExporter]

Fuzzyman, I was unclear in the statement of the problem!  I was hoping for a 
solution that would go to an arbitrary depth (i.e. m[1][1][1]).  I appreciate 
the answer, and this is the best workaround I have come up with so far:

import mock

def _gettable(*args,**kwargs):
    M = mock.MagicMock()
    M.__getitem__ = _gettable
    return M

m = mock.MagicMock()
m.__getitem__ = _gettable

Original comment by gregg.l...@lolapps.com on 7 Oct 2011 at 2:50

[GoogleCodeExporter]

Ah, indeed. Yours is an interesting and clever solution. The best solution of 
course is to use 0.8.

Original comment by fuzzyman on 7 Oct 2011 at 3:05