On Sat, Aug 13, 2022 at 12:43 AM Linus Walleij <linus.walleij@xxxxxxxxxx> wrote: > > On Wed, Aug 10, 2022 at 1:41 AM Jakob Hauser <jahau@xxxxxxxxxxxxxx> wrote: > > On 08.08.22 13:47, Andy Shevchenko wrote: > > > > P.S. Do you see now how your series and the end result become better? > > > > The driver improves. Though we kind of get lost in details, I have the > > impression we could go on like this forever. > > I think at one point Jonathan mentioned the phenomenon "perfect is the > enemy of good", which even has its own Wikipedia article: > https://en.wikipedia.org/wiki/Perfect_is_the_enemy_of_good > > When I feel (for some intuitive definition of "feeling") that a patch > series to my subsystem is getting over-reviewed, I usually just > merge it and tell everyone involved to write and/or request > additional patches if they are troubled by the result. It's a fine line, > admittedly, it's not like I can define the trigger point. The problem here is that a new round of review was triggered by a build bot that can't compile it with a Clang which follows C standard, but doesn't have sugar to it. Otherwise series was ready to be merged and can be improved later on. -- With Best Regards, Andy Shevchenko
