Re: Repository layout

On 06/04/2012 05:46 PM, Robin Berjon wrote:
> On Jun 1, 2012, at 14:22 , James Graham wrote:
>> On 06/01/2012 02:01 PM, Robin Berjon wrote:
>>> On May 31, 2012, at 16:15 , James Graham wrote:
>>>>> Overall I agree that the approval step is a bureaucratic
>>>>> speed bump that is not being helpful. I think that we should
>>>>> move to a commit-then-review model in which people who have
>>>>> an interest in passing a test suite can file bugs against
>>>>> broken tests. Ideally, we would make flagging broken tests
>>>>> easy — I'm thinking about ways of doing that in the framework
>>>>> (suggestions welcome — I wonder if I could just add a flag to
>>>>> "reported as broken" test cases).
>>>>
>>>> Make it possible to link to bug reports?
>>>
>>> Are you thinking of something like<link rel='bug' href='URL of
>>> bug report'>? I think that would work for me.
>>
>> I meant in the metadata. I don't see why you would want to link to
>> a bug about a test inside a test, really.
>
> This *is* metadata :) Storing the information in the test means that
> when you fix the bug, you just remove the link and any system
> processing the test can immediately see that it is flagged as
> bug-free. If you separate the metadata from what it applies to, it is
> likely to go out of sync and we'll end up with tests that have been
> fixed but not flagged as such in the system.

Yes, but it can *also* mean that you break the test, if you aren't very 
careful. Consider a test that does 
document.getElementsByTagName("link")[0], for example.

I wouldn't trust an automated tool to update the test files themselves, 
but I would trust them to update a separate metadata file. This would 
allow integration with bugzilla, for example, so that the test could be 
flagged as "fixed" as soon as the corresponding bug was closed (and 
flagged as broken as soon as a bug was opened/confirmed, without the 
person opening the bug having to take any separate steps).

Received on Monday, 4 June 2012 16:20:14 UTC