On Mon, May 27, 2019 at 05:54:05PM -0400, Allison Randal wrote: > On 5/25/19 12:56 PM, Greg KH wrote: > > > > Again, remember we have over 65 thousand files in the kernel source > > tree. Any single file that tries to reference them all, in any form, is > > going to be unworkable. > > Yeah, we wouldn't be looking to track every single license notice change > throughout history, that wouldn't be reasonable. We want to narrow it > down to specific sets of changes that removed license notices and > replaced them with SPDX identifiers. And, ideally, display those with > the most minimal amount of information possible. It might even be > reasonable to generate the page as a list of links to the pretty diff > displays of the relevant commits, like: > > https://github.com/torvalds/linux/commit/fd534e9b5fdcf9bab33c03cb3ade1a1ae5b23c20 > > That's the most faithful capture of the removed license notices we could > possibly provide, and is more accessible than simply saying that they're > in the git history. But, it might not satisfy the most conservative > definitions of "keep intact". > > It seems like we're weighing effort against effectiveness here, but > without a clear definition of what effective means, other than our best > guess at how "keep intact" might be interpreted by someone, somewhere, > sometime. Might it help (and reduce risks) to involve those who clearly hold copyrights to a file, or who at least claim that they hold a copyright? In other words: should these patches be CC'ed / BCC'ed (at least!) to those explicitly listed as copyright holders in the files changed by each of these patches? Thanks, Dominik
