GPLv2 forbids the additional non-free restrictions added in GPLv3 so they can't be mixed together.
It isn't permitted to use Linux kernel code in GNU projects or vice versa.
GPL is why Linux users don't have a nice mainline ZFS implementation.
This hardly qualifies as freedom.
Conversation
Replying to
While I don't agree with your statement wrt to the licenses, the reason Linux users don't have a mainline ZFS implementation was Sun and is now Oracle. They explicitly want ZFS to be incompatible with Linux.
That is neither the fault of the kernel community nor the GPL.
1
1
Replying to
It is the GPL at fault. GPL is similarly incompatible with MPL 1.0 in the same way for the same reason. It's incompatible with other licenses for similar reasons.
Even if the talking point that Sun chose the license for that reason was true (doubtful), it's still GPL's fault.
1
Replying to
Sun literally wrote the CDDL to be incompatible with the GPL a decade after Linus chose it for the Kernel. They could've done otherwise...
Oracle could still relicense/dual-license, but they don't want to.
1
Replying to
[citation needed]
Hearing the claim on hacker news from GPL / FSF advocates doesn't make it true.
Even if it is true, it's still the fault of the GPL that it's a restrictive non-free license incompatible with licenses doing patent grants (GPLv2) or MPL 1.0 style file licensing.
1
Replying to
Lets build a new incompatible license and blame the old existing one for incompatibility. Right...
Lawyers at Mozilla and Sun knew exactly what they did and what outcome it would have. Mozilla choose to use multiple licenses, because of this. Sun didn't.
1
Replying to
Based on your reasoning, GNU projects use GPLv3 specifically to be incompatible with the Linux kernel.
Apache 2 is a more recent license than GPLv2 too.
Explicitly giving patent grants while allowing defensive usage is clearly something very useful for any modern license.
Most modern software licenses including permissive ones will be incompatible with GPLv2 because they deal with patents.
Speculation about Sun's motivations for using a more granular copyleft license dealing with patents (like any more modern license) doesn't shift the blame.
1
Replying to
Yeah, it kinda does. They never bothered about Linux and didn't even try. It's simply just their fault, now Oracle's, that ZFS is not mainline. Not the license per se.
They choose to not find a solution wrt to patents.
You just can't relicense the kernel in any practical way.
1
Show replies
Replying to
Giving patent licenses sure is useful. (Sun) Oracle could still just donate Patents for ZFS (if there are any) and dual-/relicense ZFS to make it work in mainlane. Microsoft did this with Exfat, Oracle could too.
They refuse to do this for a decade now.
1
1
Replying to
That would not accomplish the same thing as the patent grant system in a license. That applies to all future code that's contributed by others under the license too. It also doesn't require fairly arbitrarily deciding which patents are applicable enough to it in advance.
1
Show replies