I suggest updating the README as well, so that it discusses this just like it does SupportNegativeIndices.

@evanphx Could you check this? Thanks!

Is this in the RFC and I just missed it?

@evanphx This is not part of the RFC. It is a part of the proposal found here: json-patch/json-patch2#18. We need this functionality ourselves, and since you already have some non standard features (negative indices) we thought that you might want to merge our changes, so that others also can benefit. If not, it's OK. We'll just keep using our fork. (I can explain why having the possibility to delete by value is very useful if you'd like.)

I would prefer not not stray from the RFC any further because it is surprising for future package users. Especially if they upgraded and suddenly an API that didn't previously allow for deletion now did.

@evanphx I'm reapplying this change to v5 for our internal use. Seeing how json-patch now uses extended ApplyOptions, would you maybe reconsider merging this feature (I'd create a new PR)? I believe this simple feature is especially useful for REST APIs o avoid extra GET requests and race conditions, and so other users of this library could find it useful. Of course, if you still want to avoid further divergence from the RFC, that's understandable, and we'll just continue using a fork.