Re: Test case review

On May 10, 2011, at 2:33 AM, James Graham wrote:

> On 05/10/2011 11:14 AM, James Graham wrote:
> 
>> * Contributor makes a number of commits
>> * They create a review request for those commits
>> * Any number of reviewers can create responses to the review request
>> where they provide comments on a specific set of lines in a specific
>> revision of the changed files
>> * Contributor makes any necessary changes, makes a new commit, and adds
>> those commits to the review
>> * Once the reviewers are happy with the changes, the review is marked as
>> approved, which causes the commits to be considered approved
> 
> Just to follow up here, the biggest requirement for me is that review is 
> easy and can be done incrementally. For HTML we have found that getting 
> people to do something relatively interesting like write and submit 
> tests is substantially easier than getting people to spend their time 
> doing something boring like review tests. This means that the tools 
> should make review as easy as possible, which means having a good diff 
> viewer, a simple way to make comments, all review state automatically 
> saved even before the reviewer is done with their response, etc.

Agreed on all points.