> On July 25, 2013, 1:54 p.m., Alan Conway wrote:
> > trunk/qpid/cpp/src/tests/acl.py, line 1254
> > <https://reviews.apache.org/r/12875/diff/2/?file=326059#file326059line1254>
> >
> >     What's wrong with the original code? I've used this pattern myself, 
> > often. I don't see any benefit to visiting every use of fail() in the 
> > tests!!

Just looked at the JIRA: the example there IS wrong, because "catch Exception" 
will catch the exception thrown by fail(). But this code is correct: 
   except qpid.session.SessionException, e
will not catch the exception thrown by fail().


- Alan


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/12875/#review23841
-----------------------------------------------------------


On July 23, 2013, 6:19 p.m., Chug Rolke wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/12875/
> -----------------------------------------------------------
> 
> (Updated July 23, 2013, 6:19 p.m.)
> 
> 
> Review request for qpid.
> 
> 
> Bugs: qpid-5010
>     https://issues.apache.org/jira/browse/qpid-5010
> 
> 
> Repository: qpid
> 
> 
> Description
> -------
> 
> Acl.py self tests issue self.fail() calls if an expected exception is not 
> thrown. A proposed fix for two patterns is shown in this review.
> 
> Essentially if the expected exception is thrown then pass. If some other 
> exception is thrown then show it's details. If no exception is thrown then 
> issue the self.fail() that was in the try block. Finally, where necessary, 
> run the cleanup handling to recover from the exception.
> 
> 
> Diffs
> -----
> 
>   trunk/qpid/cpp/src/tests/acl.py 1506161 
> 
> Diff: https://reviews.apache.org/r/12875/diff/
> 
> 
> Testing
> -------
> 
> There are 222 self.fail() calls that need to be examined and two of them are 
> changed in this review. The rest need scrutiny and when fixed may reveal 
> failures in the code as in QPID-5011.
> 
> 
> Thanks,
> 
> Chug Rolke
> 
>

Reply via email to