Hi Mike,

Do you mind if I conditionalize the mock tests?  Since they are only unit
tests and not required for runtime, it shouldn't hurt anything.

There are already tests in the codebase that have this structure:

if sys.version_info[:2] >= (2, 4):
  #define test case here

Thanks,

Roger

On Mon, Nov 23, 2009 at 4:41 PM, Mike Naberezny <[email protected]>wrote:

> Hi Roger,
>
>
> Roger Hoover wrote:
>
>> I found and fixed one other issue and updated the patch.  Here's the
>> summary:
>> - Added support for setting the ownership and permissions for an FCGI
>> socket
>> - Fixed bug where FCGI socket reference count was not getting decremented
>> on spawn error
>>
>> I updated the unit tests to verify the reference count fix.
>>
>
> Thanks for all your work on this patch.  I've applied it to the trunk.
>
> I ran into one issue with it.  Supervisor still has some users on Python
> 2.3 that we want to support.  I have fixed some other 2.3 incompatibilities
> that we introduced in the 3.0 alphas.  I believe the only 2.3 issue left is
> the tests are currently broken on 2.3 because of the use of the "mock"
> package. The "mock" package uses decorators heavily and these were
> introduced in Python 2.4.
>
> Chris and I both agree with you about all the dummy objects in the unit
> tests.  I also personally like the "mock" package.  We would like to use it
> but not at the expense of 2.3 compatibility.  We decided that we will drop
> compatibility with 2.3 in Supervisor 4.  At that time, we'd be happy to
> revisit this.
>
> Could you please modify your tests to remove the use of "mock"?  I
> apologize that we had previous said it was fine but we didn't realize this
> issue at the time.
>
> Thanks,
> Mike
>
> --
> Mike Naberezny
> Maintainable Software
> http://maintainable.com
>
_______________________________________________
Supervisor-users mailing list
[email protected]
http://lists.supervisord.org/mailman/listinfo/supervisor-users

Reply via email to