https://bugs.kde.org/show_bug.cgi?id=469782
--- Comment #6 from Mark Wielaard <m...@klomp.org> --- (In reply to Mikhail Gorodetsky from comment #4) > So, I need to > * Clarify with zstd upstream why they use three slightly different licenses Yes, we do know why, the code come from three different "upstreams". Each using a slightly different BSD 2-clause vs 3-clause license. They "problem" is that they use identical header text for all three, all referencing the same LICENSE file, but that license file only contains one variant. So the question is why don't they list all three variants in that LICENSE file or have slightly different, but more accurate, copyright headers explicitly referencing the specific LICENSE.variant. > * Clarify if they use GPLv2 or GPLv2+ Yes, I'll comment on your other comment why I think they should be explicit in the header instead of relying just on the COPYING file text. > * Add LICENSE.ZSTD, LICENSE.FSE and LICENSE.XXHASH to coregrind/m_debuginfo > with licenses for zstd, FSE library and xxHash Yes, or simply add the text of those to the zstd.h file (they are short enough and imho it is slightly more convenient to just have them inline where possible). BTW. Is this zstd.h file "generated" from the upstream sources? Then we might have to adjust the (upstream) generator program (and document clearly how to regenerate to update the zstd version when needed. > No problem, will do. And my apologies. All this is kind of being pendantic. We can actually use the code since the BSD 2 and 3 clause licenses are compatible with GPLv2+. It would just be good to have proper documentation of the origin. I think including it with the license headers as they currently are is slightly misleading and might cause trouble for someone in the future who is reviewing the origins of this code and not knowing the exact background. -- You are receiving this mail because: You are watching all bug changes.
