[llvm-dev] RFC #2: Improving license & patent issues in the LLVM community

Chris Lattner via llvm-dev llvm-dev at lists.llvm.org
Fri Nov 11 11:26:07 PST 2016

Hi David,

I think Danny has covered most of the issues that has come up, one thing I want to clarify:

On Nov 3, 2016, at 8:42 AM, David Chisnall via llvm-dev <llvm-dev at lists.llvm.org> wrote:
> For clarification: Is my interpretation incorrect?  If I compile code with GCC, which uses templates from libc++ headers and therefore results in libc++ code being inserted into the resulting binary, am I required to abide by clause 4 of the Apache license and include the libc++ attribution?
>> Yes.
>> But, AFAIK, this is deliberate.  IE the view is that in this case, you *should* be giving attribution.
>> So this is at least "not a bug", regardless of whether it's liked or not.
> I believe that this would be a show-stopper for FreeBSD’s use of libc++ and compiler-rt.  As such, I would strongly oppose this and would not consent to any of my code in libc++ being relicensed under the proposed terms.

DannyB is right that this is what is required by RFC#2, but that was already pointed out as a major problem with #2.  I agree with you that it would be a show-stopper for major LLVM users if libc++/compiler-rt/etc required this sort of attribution.  Our goal is to allow other compilers (including proprietary compilers like ICC, MSVC, etc) to use LLVM runtimes without attribution.  I’m actively iterating with folks (which, as usual, is a very slow process) to fix this.  

I have no believable ETA for an update (but I hope for the next month or two), but I’ll post RFC #3 with our next attempt at revising this.  This is the only open issue that we’re working on at the moment.  I’m hopeful that once we resolve it that we’ll have a final draft - but, of course, some other issue may come up and require a #4.


More information about the llvm-dev mailing list