On Jun 13, 2013, at 11:16 AM, Joe Bowser <bows...@gmail.com> wrote: > No offense, but have you ever run all the manual tests? Doing this is a > time consuming process and is unnecessary in most cases.
(It takes a lot to offend me, you're fine.) Agree on being time consuming, compared to the automated tests. And sometimes it's not entirely clear when a manual test succeeds/fails. Yeah, the full manual tests are probably overkill on a per-commit basis. So when should they get run, just as a final pass when the RC is cut? Or some subset of "applicable" ones that intersect with the function that is being touched in a commit? Or some other trigger? > You can write what you want on the wiki but > it doesn't mean the people writing the code will follow it (myself > included). My goal here is not to put verbage on the wiki, but to get consensus on the standard procedure on this topic. The reason I'm asking these questions is it sounds like they haven't been talked about at this level of detail before, otherwise I would be getting schooled right now. ;-) -- Marcel Kinard
