traeak commented on code in PR #12408:
URL: https://github.com/apache/trafficserver/pull/12408#discussion_r2282243835
##########
plugins/slice/server.cc:
##########
@@ -507,17 +489,6 @@ handleNextServerHeader(Data *const data)
}
return true;
-
- } break;
- // Refetch the reference slice
- case BlockState::ActiveRef: {
Review Comment:
ActiveRef mean active slice refetch. This is no longer needed as non first
slice refetches are entirely handled by the CRR plugin. Comment must be wrong?
The process is as follows:
- slice plugin requests first slice. This establishes the identifier (etag
or lm).
- slice plugin requests second slice, passing along the first slice
identifer.
- CRR plugin during cache lookup compares the identifier to the cached one
and either returns it (meaning match with first slice slice) or marks that
slice stale and refetches.
- slice plugin: if the refetched slice ident mismatches the first slice, the
transaction is cancelled and the first slice is re-requested with the refetched
slice's identifier, causing the reference slice to be refetched as "stale".
I'll look at doing a better job updating the docs. With this change there's
no need for the slice plugin itself to deal with refetching anything other than
the first (reference) slice anymore.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscr...@trafficserver.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
