On 5/04/2010, at 9:43 PM, Bob Morley wrote: > > Is it reasonable for a non-commiter to review commits? To be honest, I never > even considered that I should be doing that.
Reasonable? It would be fantastic, we have 166 subscribers and very little feedback. > However, two other things occurred to me ... > > Got the distinct feeling that this points to a not very good practice when > it comes to unit testers. Buildbot has been reporting failures since yesterday: http://ci.apache.org/waterfall?show=ofbiz-trunk The problem wasn't missed, I just think Adam was a little peeved that no one noticed his simple mistake in such as small commit. > I definitely understand the pain of attempting to > force a "must have tester" policy but having a project code-coverage goal > might be something to strive towards (with published metrics). Moreover, I > would guess in this case a unit-tester would have likely caught the inverted > condition check (if it was compiling). > > I am definitely not a fisheye expert, but I would hope with its JIRA > integration that something could be automated that would drive reviews > targeted at the commiters. Again having a code-review goal (published) > would be something the project could strive towards. > > It would be pretty awesome to have Ofbiz - 2+ million lines of code, 100% > code reviewed, 90% unit-test code coverage. This isn't a dig at you, but talk is cheap :-) Write some tests and I'll commit them. Regards Scott
smime.p7s
Description: S/MIME cryptographic signature