bneradt commented on PR #13447:
URL: https://github.com/apache/trafficserver/pull/13447#issuecomment-5955724973

   Rebased onto current master, `8e3c77e4b8`, preserving #13695 in full. Its 
atomic policy storage, reload callbacks, per-connection policy snapshot, both 
reload test files, and reload documentation are present. The reload 
implementation and test files have no diff against master.
   
   One history clarification: the reviewed head `c1db6d4c1b` had parent 
`c8935543e2`, which predates `7e2560376b`. It did not delete the reload tests 
or reverse that commit: #13695 was not yet in its ancestry. Comparing the old 
head directly with newer master shows those changes as absent, but the PR's 
patch against its own base did not remove them. The rebase applied without 
conflicts, and `git range-diff` showed the patch unchanged before the 
documentation addition below.
   
   I also documented `proxy.process.http.origin.reuse_fail` as per-probe 
decisions, including two increments if both probes fail and overlap with a 
successful global reuse or a global miss. It is explicitly not a failed-request 
or distinct-transaction count.
   
   Validation on the rebased branch: build/install, formatting, docs build, 109 
focused CTest cases, and all 10 selected AuTests passed. These include all 29 
`http2_flow_control` scenarios (including the cited policy-2 inbound/H2-origin 
scenario), `http2_flow_control_reload`, initial receive windows, 
inbound/outbound SETTINGS ACK handling, continuation, payload validation, 
half-close, pool reuse, and serialized safe retries.
   
   The cited Jenkins build's API and console URLs currently return 404 to the 
configured authenticated reader, so I cannot establish the cause of that 
earlier failure. The passing local test does not by itself prove the rebase 
fixed it. Fresh CI still needs to validate this revision.
   


-- 
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: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to