Re: More Python induced QA pain

[Ken McDonell <kenj@xxxxxxxxxxxxxxxx>]

On 26/07/16 00:07, Frank Ch. Eigler wrote:
> Ken McDonell <kenj@xxxxxxxxxxxxxxxx> writes:
>
> > qa/662 this time (pmwebd testing from python test code) with a variety
> > of fatal Python exceptions. [...]
> > Traceback (most recent call last):
> >    File 
> > "/usr/lib64/python3.4/site-packages/requests/packages/urllib3/connectionpool.py",
> >  line 376, in _make_request
> >      httplib_response = conn.getresponse(buffering=True)
> > TypeError: getresponse() got an unexpected keyword argument 'buffering'
> > [...]
> > During handling of the above exception, another exception occurred:
> This is a problem in a system python libraries, which must have
> changed version on your VMs.  I guess the 662 test case
> (src/test_webapi.py) could do some sort of smoke test to see if
> the requests module is at all alive.  Or just ignore this and
> file a bug report with your distro.
Thanks Frank, but I'm no closer to resolving the issue.

The particular problem you're referring to seems to be related to a 
broken implementation of chunked transfer-encoding in Python 2.6.1 and 
2.6.2 (at least) as far as I can tell from here ... 
https://github.com/kennethreitz/requests/issues/2568
But this test case is failing on Python 3.5.2 in at least one case which 
is the latest and greatest Python as far as I can tell.
And is the test expected to work at all with Python2?  I ask because the 
failure signatures are different with Python2.
The fact is that for a variety of failure signatures this test is not 
working ON ANY of the 30-odd platforms in my QA Farm, so I'm a little 
loathe to write it off as a Python version issue ... it seems more 
broken than that.