Hi Hiltjo, Hiltjo Posthuma wrote on Sun, Dec 23, 2018 at 03:34:12PM +0100:
> Shouldn't these kind of diffs be OK'd by other developers before > committing though? Not necessarily. When a developer focusses on work in a particular area and takes responsibility to fix anything that breaks there, in particular when few or no others work in the area, that developer owns that code. When few others work in an area, requiring the code owner to get OKs for each and every improvement can be detrimental because on the one hand, it can slow down development for little benefit, and on the other hand, it can put undue strees on the few others who occasionally look at that code, too. Both getting and giving OKs can be frustrating and time-consuming. Espie doesn't ask for OKs in pkg_add(1), not even for major changes, unless he is worried whether or not some general direction is desired. Neither do i in mandoc(1). Neither does landry@ in firefox nor robert@ in chrome. And so on. Even though i don't use mg(1), my impression is that lum@ owns that code nowadays, that almost the only other developer occasionally contributing to it is florian@, and that these two have been doing a very decent job during the last few years: i haven't noticed any complaints about their work on mg(1). Unless other *developers* request a backout, the code owner gets to decide about the details of the way forward. And "the current behaviour is already better than what was there last week, and i'll work on it next week for further improvements" is definitely an adequate response. Yours, Ingo
