Interesting. I've been under impression that once you licence your code, you cannot "take it back" (for that particular version or release).
https://stackoverflow.com/questions/5419923/can-gpl-be-re-li...
As an example of the difficulty: U.S. Copyright gives authors absolute termination rights.[1][2] An author can unilaterally choose to terminate a license or transfer 35 years subsequent to the grant, but no more than 40 years after. A contract, promise, or some other contrivance executed to nullify the right is void, just like you can't willingly contract yourself into slavery. Which is to say, a court won't punish you in any way, shape, or form for statutory termination; and won't hear any arguments about reliance or other arguments intended to coax a court to enjoin termination.
However, statutory termination requires notice to be given at least 2 years prior (but not more than 10 years prior) to termination. There are unresolved questions regarding notice and its effect. Many observers (including Stallman) are of the opinion that statutory termination rights are of little or no practical consequence to FOSS because of the technical requirements of notice, but arguably that's motivated thinking on their part.
The types of termination being discussed in the context of the CoC involve more context-dependent questions of law and fact. If the practical effects of the most clear-cut and categorical route to termination are still left open, you can bet that other routes to termination pose significant unresolved questions and implications.
[1] Statutory Termination: https://www.copyright.gov/title17/92chap2.html#203
[2] Copyright Office's rules for effective notice: https://www.copyright.gov/title37/201/37cfr201-10.html
Hell, just reissue with GPLv3. That'll still kill it pretty well.