You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Right now ucanto assumes that it sends HTTP request with invocations and will get receipts for it in the response. This design does not account for long running tasks where it would be impractical (or sometimes) possible to keep connection open.
We have been working around this limitation in filecoin pipeline and pretty much in every new thing we have worked on since (like blob interface).
This limitation has negative effect on the protocol design as we do things differently to work around the limitation, which translated to poorer developer experience and increased maintenance burden.
Cost
To make an educated estimate we'd need to put some work into design & implementation plan, then we'll be able to tell what needs to change how much work that would be. I suspect coming up and writing up initial design would be about a day of work. Probably me or @vasco-santos would be best candidates for it and after initial design probably few rounds of feedback loop (probably some hours split across across days)
@Gozala - please help me convert this more to the general what/why/cost template. Sounds like primary thing we get is ability to have long running invocations. Is there a specific long running invocation we need to provide right now that is blocked by this?
What
Implement support for async receipt delivery
Why
Right now ucanto assumes that it sends HTTP request with invocations and will get receipts for it in the response. This design does not account for long running tasks where it would be impractical (or sometimes) possible to keep connection open.
We have been working around this limitation in filecoin pipeline and pretty much in every new thing we have worked on since (like blob interface).
This limitation has negative effect on the protocol design as we do things differently to work around the limitation, which translated to poorer developer experience and increased maintenance burden.
Cost
To make an educated estimate we'd need to put some work into design & implementation plan, then we'll be able to tell what needs to change how much work that would be. I suspect coming up and writing up initial design would be about a day of work. Probably me or @vasco-santos would be best candidates for it and after initial design probably few rounds of feedback loop (probably some hours split across across days)
Context
storacha-network/w3up#1326
The text was updated successfully, but these errors were encountered: