Edit: meaning, I think, that "miners" (same as other full nodes) would accept/ignore codes they don't know IF they are included in a block, but would not accept them for inclusion into their own blocks.
+1 .. this is soft fork
That was my basic understanding, but I'm glad to get confirmation from ~odalv on this.
Nothing has been stated which refutes my logic. If the old version has the functionality that luigi1111, Odalv, and tcbcof have reaffirmed is their assumption, then the dire outcome I explained is the potential result. Perhaps there are other possible mitigations to such an attack.
Somebody needs to bring this to the attention of the Core devs. Appears on the surface of it, they may have a serious lapse in logic (but not having seen their rebuttal, I might be missing something).
I did not think it refuted your logic, beyond that you need at least one miner in your pocket (or your own miner, whatever) to get started.
On the surface it seems like you may have something, but I've not thought about it in depth (and indeed, am no expert).
