Jeremy Taylor

> For your particular use case- how far would it get you if Nippy core included something like a public nippy/type-ids, maybe with explicit length in bytes? > Seems that'd...

Good question! It would at least definitely make an excellent demo. @sidorares funny you should mention encryption, I was dreaming about getting GMSS working from http://www.flexiprovider.de/#PQCProvider

@wotbrew is this closeable now?

Also worth noting that whilst one can `put` a tombstone (as it appears via `open-tx-log`) using the regular method and it ~works (ignoring the indexing mentioned above), the tombstone document...

Ah yep that should have been a with-open, my mistake. In which case maybe this whole repro I attempted to cook up is a red herring. The report still stands...

Hey @FiV0 I just wanted to say thanks again for reflecting on your project requirements and suggesting the change here! I know @jarohen will offer more feedback in due course...but...

Thanks @hukka - I edited a bit to clarify what I think the most realistic scope would be here (although that can still be debated :slightly_smiling_face:).

Hey @ferdinand-beyer I can confirm that this is an expected (though unfortunate) related effect of #1510, where map IDs similarly don't follow `=` compatibility. > The issue is that select-keys...

The subsequent problem to solve is a "missing docs" error, and that looks to be because the ChunkedSeq serialization is losing the edn tag, and therefore `fetch-docs` is being passed...

Hey @mallt - on reflection I think it may be simplest to perform a one-time migration of the underlying tx-log by simply converting the `clojure.lang.PersistentVector$ChunkedSeq` instances to vectors, without modifying...