6 comments

[ 5.0 ms ] story [ 28.6 ms ] thread
I wonder how viable this tactic is?

Require all requestors to identify themselves as individual open source hobbyist or closed source commercial.

Refuse the closed/commercial work unless they make a monetary contribution.

The (A)GPL is a source code copyright. It doesn't force you to fix anyone's problem or even interact with them.

If the requestor is a business, it'll be an undesirable development burden to carry their own fork with their own patches. It probably makes business sense to pay instead of hiring their own compiler developer.

Might not make for the best community interaction though.

Seems like a palatable workaround. I think you may need to enforce it once or twice to have teeth but it’s not unheard of. I heard that you can’t release GPL software on the Apple Store due to similar issues.
> I think you may need to enforce it once or twice to have teeth but it’s not unheard of.

That's assuming the courts agree. I haven't read the linker code, but it's very possible that whatever code the linker includes would not be found "creative" enough to be copyrightable. In contrast, libgcc, which gcc links into generated binaries, contains non-trivial routines[0].

[0] https://github.com/gcc-mirror/gcc/blob/master/libgcc/libgcc2...

> I think you may need to enforce it once or twice to have teeth

I really don't think that is the point. This seem to be targeted at scaring legal departments to ease over the process of having corps fork over the money for something they might think could be obtained for free.

I'd obviously prefer mold to be purely OS licensed, but this is a great compromise that will hopefully get Rui paid.
Eh, claiming that AGPL linker results must also be AGPL is extremely iffy and if courts agree is a terrible precedent. I understand that he wants to be able to continue doing this and needs to make money, but I think the previous idea, as unviable as it was, is better than this