[Zope-Coders] Re: July Bug Day Roundup

[Tres Seaver]

Chris Withers wrote:

> OK. I still like the idea of assigned stuff going back to pending if no 
> action happens for a month. Is that a possibility?

I would rather address the issue with "nagmail":  each supporter would 
get a periodic (weekly, likely, for the main Zope collector) e-mail 
summarizing the issues to which they are assigned.

>> "Resolved"
>>
>>   The fix or feature has been implemented, along with any necessary 
>> tests.
>>
>>   Tests which would provoke the bug, if it were still present, and/or
>>   demonstrate the correctness of the new functionality are always 
>> required,
>>   unless they are impossible to write.  (Leniency may be warranted for 
>> very
>>   minor changes?)
> 
> 
> ...or in situations where ther eare no tests for the whole area the 
> change is to be made in?

Nope.  *Any* change could be expected to have one or more tests which 
demonstrate either that:

  - the bug existed before before the fix (e.g., the test fails without
    the fix), but no longer exists after (the test passes).

  - the new feature works.

Writing more tests for the unit being changed qualifies as a "work of 
supererogation."  There are occoasionally areas (medusa's guts, for one) 
which are so gnarly that nobody can figure out how to write tests for 
that.  If you plan to resolve an issue without adding test, and want to 
plead "gnarliness" as an excuse, expect to defend the choice. ;)

Tres.
-- 
===============================================================
Tres Seaver                                tseaver at zope.com
Zope Corporation      "Zope Dealers"       http://www.zope.com