Hello! I was looking at Slingshot while scheming, and wondered:
- are you planning to support for "
listRecords"-like endpoints or is Slingshot for individual records? - and/or, what about "wildcard" queries within a specific repo—like all records of a specific
NSID, or all those matching a partial one (e.g.app.bsky.*)?
Not need for either to come quick, just curious about scope 😃
Hello Ms Boba!!
listRecords: yes! probably as part of a general proxy-to-PDS xrpc endpoint (slingshot can usually query PDSes faster because they are pre-resolved and sometimes already have an active keepalive connection), though we might be able to do neat cache-correct things with listRecords specifically.
as a record cache slingshot itself doesn't actually have a complete list of records per account, so it can't directly service queries by NSID or partial matches itself, unfortunately.
but!!! i'm going to implement partial repo exports in hubble -- sync1.1 already pretty much specced it out but it's never been implemented as far as I know. IIRC bryan and retr0id were aligning on a range-bounds-like query API. I was thinking of also doing prefix and/or collection shortcuts. In any case, these might get exposed publicly for hubble through slingshot in front, so your questions are all relevant here!