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
{{ message }}
This repository has been archived by the owner on Dec 18, 2023. It is now read-only.
[10:46 AM] fabrice:
why is "no-cache" used for the default cache header value?
[10:59 AM] Arqu:
Lots of those bits are still not fully fleshed out
[10:59 AM] Arqu:
Technically we cache everything if the store is present
[11:00 AM] Arqu:
But we dont handle the etags and cache headers fully so it doesnt make sense we advertise anything else
[11:00 AM] Arqu:
The blessing with content addressed storage is you can cache a cid indefinitely technically as the contents wont change
[11:01 AM] Arqu:
And you never cache ipns type reqs
I'm not sure if it's worth exposing cache headers as configurable. As @Arqu says, behavior needs to different between /ipfs and /ipns. For raw and car response, we should return Cache-Control: immutable.
Thoughts?
The text was updated successfully, but these errors were encountered:
Spec is already being followed, in the next update the misleading flags will be removed. The only exception is that for the IPNS based Last-Modified date we follow what kubo does instead of taking the spec recommendation to either do dns ttl based dates or not at all.
From discord:
I'm not sure if it's worth exposing cache headers as configurable. As @Arqu says, behavior needs to different between
/ipfs
and/ipns
. Forraw
andcar
response, we should returnCache-Control: immutable
.Thoughts?
The text was updated successfully, but these errors were encountered: