my gut says that so long as my patched pds is logging the dpop jti values the replay issue will magically fix itself

just you wait im going to catch you https://pds.kelinci.net/xrpc/com.atproto.sync.getBlob?did=did:plc:ia76kvnndjutgedggx2ibrem&cid=bafkreicvu7kjr3e4buqhrmyokwvv6uq3udpjrkjo5pk2c6re55z4pkptb4

i could try debugging this a lot more closely by attaching a lua script to my nginx proxy so it would log jti values and the timestamp but something tells me this is not going to be worth doing

🤏 i'm so close to just giving up on oauth entirely, long live app passwords

the only thing i've concluded with me getting "dpop replay error" is that mobile browsers are cursed, i don't know why it's happening

if I do prereleases with changesets on a separate unstable branch, do I exit out of it *before* merging to main? the docs doesn't seem clear on this besides recommending that I should do separate unstable branch

spun up a dumb xrpc service that implements com.atproto.identity.resolve* endpoints since not even the reference pds has them for some reason [identitas.kelinci.deno.net/xrpc/com.atpro…](https://identitas.kelinci.deno.net/xrpc/com.atproto.identity.resolveIdentity?identifier=mary.my.id ) [github.com/mary-ext/kl-id…](

GitHub

kl-identitas/mod.ts at trunk · mary-ext/kl-identitas

Contribute to mary-ext/kl-identitas development by creating an account on GitHub.

)

what https://pds.kelinci.net/xrpc/com.atproto.sync.getBlob?did=did:plc:ia76kvnndjutgedggx2ibrem&cid=bafkreigm4xxbq5e2cysgwotkjxowjfoljp3na7ojmfmtq23wcaefbjs52e

[music.youtube.com/watch?v=QmFhKm…]( ) [Kono Sekai De Mitsuketamono]( )

Next.js' parallel and intercepting routes seem nice, I wish other frameworks has them