On Thu, Nov 28, 2024 at 08:53:06AM +0900, Junio C Hamano wrote: > Jeff King <peff@xxxxxxxx> writes: > > > But it seems like a very deep rabbit hole to start adding in > > shared-library line validators, because I think it ends in "now > > compile this before I agree to apply the patch". > > I am not sure I understand your conclusion. Who is telling that to > whom? Somebody sends a patch that creates a file that requires a > special validator and the maintainer gives the validator and tells > the contributor to go use it to make sure their addition passses > before resubmitting? > > I was hoping that the ability to add extra validators is more of an > enabler (than requirement and hindrance) for those who choose to be > extra careful. It is similar to CFLAGS in our Makefile that allows > you to use options to enable more strict compiler warnings than what > other developers usually use, to notice certain class of problems > others may miss. Yes, we who introduce the mechanism to create plug-ins do not have to worry about writing those plug-ins ourselves. But we do have to maintain the plug-in interface, and respond to complaints when it is not rich enough to do what people want to do. So I was merely pessimistically foreseeing where this may end up. ;) Of course... > Shared-libraries and plug-ins remain to be solution in search of > problem at least for this project. I do not really need CSV comma > counter, but I thought it may give a good excuse for those who want > to play with Rust and other stuff ;-) ...if playing with the plug-in interface is the point, none of that may matter. :) -Peff
