Super weird that signmessage is being used as the standard of "not bitcoin anymore"-- we created signmessage in 2011, nothing like it was originally part of bitcoin(*). It's always been a kind of half hearted thing: gpg is much better for message signing. It's never worked with multisig or p2sh, for example, and it's usage seems to really depend on address reuse which is harmful to the whole ecosystem.   It would be trivial to extend the signmessage command to common single key bc1q addresses but there just isn't much interest in doing that (though I would also be surprised if something like electrum or whatever hasn't done it), because doing so still wouldn't let it work with multisig or other script types and wouldn't make it a useful alternative to gpg.

I think a lot of people would be more interested in a new message signing format that was based on signing an invalid (forever timelocked or whatever) transaction, because that could be generic for _all_ addresses including multisig and script using ones... Perhaps it could be based on the PSBT message format... but there too it doesn't seem like there is enough interest for anyone to go about implementing it either.

*So we can assume Bitcoin SV is going to remove it, just like the p2sh support needed to pay to the address in their signature...