Jeff King <peff@xxxxxxxx> writes: > On Wed, Mar 17, 2021 at 12:38:42PM -0700, Junio C Hamano wrote: > ... >> You may want to join the list discussion and stop whatever UI change >> you find undesirable before it materializes next time. > > If we discover later that some portion of users prefer an older > behavior, it may be reasonable to build on top with a config option to > allow selecting the old or the new. True. I do not think this one passes the bar, as the porcelain UI is subject to change, and it is not worth flipping and flopping every few years. > - how big is the code burden to support both behaviors? In this case, > I don't think it's too bad; it's restoring the old newlines with a > conditional Accumulated little cuts will start hurting someday, though. I think the change was done to make the output consistent between two cases (with or without "use X to do Y" hints), so it won't be sufficient to conditionally revert with a switch. An option to add these blank lines back would also have to add extra blank lines to places where there was (and is) no blank line, so it is more like a new development than a partial reversion. cf. https://lore.kernel.org/git/?q=d%3A20190415..20190615+f%3Ajohnlinp%40gmail.com > - how many people would actually care enough to set the option? Even > without a lot of code, each option is _some_ burden to carry, both > in the code and in the overall complexity of the interface. I'm less > convinced in this particular case that a lot of people care (given a > single comment after many releases), but perhaps those interested in > the change could produce data (note I said "data", not "argue more > vigorously" ;) ). ;-)
