proposal for blanket authorized orange fixes for mail & mailnews meeting specific constraints
asutherland at asutherland.org
Fri Apr 1 03:58:17 UTC 2011
On 03/31/2011 08:11 PM, Justin Wood (Callek) wrote:
> The deal-breaker for me is "module owner or peer of code in question."
> since there are *some* people who could abuse this power if that
> wasn't the case.
Did you mean to type "deal-maker" here? It feels like this statement
wants to agree with the previous statement. In case it's a specific
point you do disagree with, let me elaborate on my logic for that point:
- I am assuming the module owners/peers have the best understanding of
what the test is actually doing or actually intending to do. It is very
easy to break a test so that it neither fails nor tests what it is
supposed to test.
- I believe we want the people theoretically responsible for a module to
know what is going on inside of it. If the list of module owners/peers
does not correspond with reality, it should be amended. The pragmatic
example of this is that if I am writing buggy tests and someone keeps
fixing them without me noticing, I may just keep on writing buggy tests :)
More information about the tb-planning