On Wed, Feb 9, 2022 at 11:14 AM Jilayne Lovejoy <jlovejoy(a)redhat.com> wrote:
On 2/9/22 9:08 AM, Richard Fontana wrote:
> On Wed, Feb 9, 2022 at 10:53 AM Florian Weimer <fweimer(a)redhat.com> wrote:
>> * Serge Guelton:
>>
>>> the LLVM project has moved to an Apache Software License 2.0 with exception
>>> license, referenced as
https://releases.llvm.org/10.0.0/LICENSE.TXT
>>>
>>> Some more details are available here:
>>>
>>>
https://llvm.org/docs/DeveloperPolicy.html#new-llvm-project-license-frame...
>>>
>>> Does it make sense to have it listed in
>>>
https://fedoraproject.org/wiki/Licensing:Main#Good_Licenses and the
>>> associated short name be compatible with, say, rpmdiff?
>> Isn't our position that the relicensing has not happened yet, that the
>> SPDX identifiers in the sources are incorrect, and that the project
>> still distributes the sources under the old LLVM license (called “NCSA”
>> in the Fedora framework)?
> Separately from that issue, I am aware of one unresolved (for my team
> at Red Hat) longstanding objection to some aspects of the language of
> the LLVM exception (raised by at least one person outside of Red Hat).
> I wouldn't want to classify the exception as "good" without reaching
> some sort of resolution on that issue.
>
> Richard
just saw this after I wrote my email. That's interesting and a bit
surprising - I'd be curious to hear more about what the objection is
(and especially since I know LLVM got a lot of feedback when they were
drafting it!)
I think they got a lot of feedback from a somewhat narrow part of the
universe that didn't include historically non-LLVM-oriented community
members.
To be clear, my position would be that even if the exception embodies
some horrible drafting problem it does not mean LLVM, conceptualized
as being "Apache-2.0" (-only, as it were) or "Apache-2.0 AND NCSA" or
whatever is problematic for Fedora. It sort of goes to the issue of
how important it is to make sure the license metadata incorporates the
details of all exceptions given that normally exceptions are
structured as additional permissions (though sometimes "exceptions"
are additional restrictions and historically some of those have been
problematic). Not so long ago (before we were engaged in the current
iteration of looking into having Fedora migrate from Callaway
Notation™ to SPDX identifiers) I suggested that Fedora abandon all
efforts to track exceptions as being ultimately a waste of time, but I
have a feeling you would disagree with that view. :-)
Richard