Closed milt closed 2 years ago
Closing this, in light of https://github.com/yetanalytics/lrs/pull/81: if deduplicated attachments are accepted, there is no way to tie a statement attachment reference 1-1 with an attachment without the expectation of denormalization.
LRS-70 We match and validate statement data to attachments, but throw out the resulting pairing.
It would be useful for impls to have this for a variety of reasons, chiefly addressing attachment storage
Also examine current behavior: We require that there be a (possibly duplicate) attachment present for each reference, even if the SHA2 is the same. Figure out if this is intended + correct behavior.