I endorse that as the complete, right way to do it, myself. That said I think we'd get no work done if we followed this full process for every change we make. But, we can't work with zero coordination. So the practical answer must be something in between.
I think we can afford to be a bit less formal on a small, and a very young, project. I suppose I don't bother with a JIRA issue for tiny stuff. For example, just saw some code was committed that uses log4j directly. Since we use slf4j this didn't compile, so I just fixed it. And on JIRA issues I assume we should use our judgment -- if Grant concluded from his own work and other email threads it was resolved, seems fine to mark it as such, skipping ahead, technically. I say we use our judgment, err slightly in favor of over-communication, and communicate openly about whether we are being too formal or too informal. Sean On Wed, Jun 17, 2009 at 11:03 AM, Robert Burrell Donkin<[email protected]> wrote: > grant committed the patch in > https://issues.apache.org/jira/browse/MAHOUT-134 (thanks :-) and set > the issue to resolved but not closed > > am i right in thinking that mahout is using the convention that the > reporter of the issue should check that the committed code works ok > then close the issue...? > > (or is there some other convention at work here...) > > - robert >
