A more hands on review process

[Stephen Kelly]

Aaron J. Seigo wrote:

> On Thursday 31 July 2008, Stephen Kelly wrote:
>> I propose a review process based on review criteria instead of time.
>> Reviewers would 'tick the boxes' to confirm that they have reviewed it,
>> and what they looked for. Multiple reviewers could tick only some of the
>> boxes each if necessary (I'd have no idea if something works on solaris
>> or has security flaws for example), and together create a complete
>> review.
> 
> would this take the form of a web-based collaboration app? or?

I was thinking email on k-c-d. If there's a suitable web-app for this kind
of thing which actually makes it easier, then that's cool too. Is
review-board suitable for applications, or mostly just patches (I have no
idea).

Thanks for the comments on the list. It did occur to me while making it that
some of it is specific only to "Public Library", or "Application" etc. I'll
break it out when I incorporate all this feedback.

>> Portability
>> * The new code has no unnecessary platform specific code, such as
>> reading /proc etc. This does not apply to new code which is relevant only
>> on specific platforms.
>> * The application compiles and runs on Windows, Mac, BSD, Solaris.
> 
> this one is going to be very hard to determine as few people have all
> these systems and afaik even fewer of those on the fledgling win/mac
> platform teams follow kdereview. this would be a nice future goal though.
> perhaps it would be good enough for now for these to be optional
> checkboxes =)

Sure. It's all optional. In the future the compilation at least might be
part of a cross-platform dirkboard-like system, or a continuous automated
CDash board. I'll try to get some kind of list together of things to avoid
(reading /proc, the happy constructor pattern etc) for the list.