[asterisk-dev] Reviewboard and commit policy change

[Kevin P. Fleming]

On 09/05/2011 07:12 AM, Olle E. Johansson wrote:

> I think we should open for smaller commits from experienced developers - like I see in other parts of Asterisk than chan_sip. I have had a review open for weeks for a trivial change that I've proven in running systems that it prevents crashes. I haven't got a "ship it" during over two weeks in the process so at this moment I can't move it forward. It hangs there in reviewboard and bug tracker while systems are crashing out in the user field.

As far as I know, what you are proposing was *already* the policy. An 
experienced developer should certainly feel free to commit changes that 
they feel are trivial (or even non-trivial, but straightforward). After 
all, the changes are going into an SCM, and are emailed out to hundreds 
of people immediately; we have plenty of opportunities for post-commit 
code review, as long as the size of the change is manageable to review 
in a minute or two.

> If things go bad in the svn code, we can fix it. But at this point I feel it's better to fix stuff - even maybe incorrectly - than to wait for a reviewboard process that no one has time to devote to.

There we'll have to agree to disagree; I'd rather have a problem go 
unfixed than have the *wrong* fix go in. We have too many examples of 
*wrong* fixes causing more pain for users, and being harder to track 
down and resolve, than the problem they were supposed to solve.

> Asking for me or others to spend more time in Reviewboard for other persons issues doesn't really help solving this issue - and that is the only response I have gotten so far when trying to discuss this process. In the best of possible worlds we would all have time to work on future stuff, help others with their code and work with bug tracker and reviewboard processes. I don't find myself being able to afford doing that in this world that I live in. But I still want to be able to contribute.

There is a cost to being a productive, collaborative member of a 
development community. This is true in pretty much every open source 
community that I keep tabs on or participate in; people who never 
participate in review of code they didn't write end up seeing fewer (or 
no) reviews of their own code.

> I can accept to use reviewboard for larger changes or additions of new code. But not for code cleanups, smaller bug fixes and such maintenance work. It just adds to the burden of everyone and makes the process take too much time, which means that we won't see this work in subversion. I think that's bad for everyone.

I'm not aware of any policy that status that ReviewBoard must be used 
for 'code cleanups' and 'small bug fixes' or maintenance work. If there 
are open reviews for such changes, posted by experienced members of our 
community, then I'm sorry time has been wasted by posting the reviews, 
because it wasn't necessary.

-- 
Kevin P. Fleming
Digium, Inc. | Director of Software Technologies
Jabber: kfleming at digium.com | SIP: kpfleming at digium.com | Skype: kpfleming
445 Jan Davis Drive NW - Huntsville, AL 35806 - USA
Check us out at www.digium.com & www.asterisk.org