On Tue, 19 Dec 2017 20:11:04 -0600 R0b0t1 <r03...@gmail.com> wrote: > I forgot most files were mirrored. So the infrastructure that is the > answer to my question is already in place. Consequently, I don't think > there's any reason to argue against this, unless it ultimately ends up > being a ton of work to package small files (which I can't comment on).
Biggest downsides I see of pushing patches to distfiles mirrors is it greatly desynchronizes the state of tree, similar to how you'd get desynchronization if parts of gentoo.git were sharded into different repositories. Comprehending changes made downstream require additional steps and additional tooling. Correlating patch changes against ebuild changes is even more effort. The only upside is it makes it slightly harder to abuse patch-reuse and have unintentional retroactive patching to existing ebuilds without -r bumps. But ... that's a double edged sword if that sort of thing is occasionally useful and sane. I don't know what the solution is here, but I don't think either strategies of "discourage it" or "encourage it" are the ultimate way forward. Some other strategy must exist. But for now, sensible limits are an acceptable compromise.
pgpNcakDxjHJ5.pgp
Description: OpenPGP digital signature