It would depend on the communications made by the copyright holder, or
ratifications made by him (if any). Estopple arises when a grantee relys
on the grantors' communication that he will not rescind, and the grantor
takes an action based on that communication.

The clause in version 3 of the GPL (aswell as the CC licenses etc)
furnishes a clear possibility of such a defense.

Version 2 of the GPL lacks such a clause, so one must look to the
communications of the contributing copyright holder's statements, etc.

(Another problem is that the contributor may have never communicated to
the people he might want to rescind from, which adds another hurdle to
an estopple defense)

For instance, if the contributor was queried on the topic and gave a
response publicly, then the next question would be "did the defendant
know of and rely on that statement".

The linux-kernel's ... lassie-fair ... attitude when it comes to formal
matters and the sheer volume of contributors makes one wonder if
anything of the sort was done.

It is, indeed, a problem for Free-Software in the United States, as
copyright is simply "alienable in all ways property is"; and a licensor,
barring an attached interest or contract terms or communications that
would give rise to estopple, has an absolute right to rescind a license
regarding his property at his will; which is one feature that
distinguishes a license from an easement, servitude, real covenant, etc.

(Note: My projects dual license GPLv2 and GPLv3 for this reason: v3
furnishes an estoppel defense, v2 makes everyone happy because it's
"what they know", the linux-kernel never made any real attempts to adopt
the new license and omitted the "any-later-version codicil" from the
very beginning)

On 2018-09-19 03:38, Richard Stallman wrote:

> [[[ To any NSA and FBI agents reading my email: please consider    ]]]
> [[[ whether defending the US Constitution against all enemies,     ]]]
> [[[ foreign or domestic, requires you to follow Snowden's example. ]]]

>
> > One is rescission of the license they granted regarding their code,
> and
> > then a lawsuit under copyright if/when the rescission is ignored.
> > The others are breach of contract, libel, false light, etc.
>
> If "rescission" is really a possibility, it would cause greast trouble
> for the free software community. We would need to take steps to make
> sure it cannot happen.
>
> However, that goes against everything I have been told by others.