iloveeyjafjalla opened a new pull request, #11378:
URL: https://github.com/apache/rocketmq/pull/11378

   ### Which Issue(s) This PR Fixes
   
   - Fixes #11377
   
   ### Brief Description
   
   `VerifyMessage` answered `MESSAGE_CORRUPTED` for every outcome that was not 
`CR_SUCCESS`, although the consumption outcome is
   graded on both sides of the call:
   
   - `ClientActivity#buildConsumeMessageDirectlyResult` maps a gRPC client's 
answer: `Code.OK` → `CR_SUCCESS`,
     `Code.FAILED_TO_CONSUME_MESSAGE` → `CR_LATER`, `Code.MESSAGE_CORRUPTED` → 
`CR_RETURN_NULL`;
   - when nothing is consumed, the relay result itself carries the reason (e.g. 
`SYSTEM_ERROR` with
     "client … is not connected to this cluster").
   
   So a client that refused to consume (or was offline) made the surface report 
a healthy message as corrupted, and a caller
   that reacts to `MESSAGE_CORRUPTED` did the wrong thing instead of retrying 
or fixing the client.
   
   The change picks the code from the reported outcome: `MESSAGE_CORRUPTED` 
only for `CR_RETURN_NULL` — the single outcome a
   client reports for a corrupt message (`Code.MESSAGE_CORRUPTED`) — and 
`FAILED_TO_CONSUME_MESSAGE` (the proto's generic code for
   failed consumption) for the others, keeping the relay's remark in the 
message when the client never answered.
   
   ### How Did You Test This Change?
   
   `ProxyAdminGrpcServiceTest`:
   
   - `verifyMessageReportsFailedConsumeWhenTheClientAsksForRetryTest` 
(`CR_LATER`) and
     `verifyMessageReportsFailedConsumeWhenTheClientThrowsTest` 
(`CR_THROW_EXCEPTION`, this replaced
     `verifyMessageReportsCorruptedOnFailureTest`, which asserted the old 
collapse) → `FAILED_TO_CONSUME_MESSAGE`;
   - `verifyMessageReportsCorruptedWhenTheClientReturnsNullTest` 
(`CR_RETURN_NULL`) → `MESSAGE_CORRUPTED`;
   - `verifyMessageReportsFailedConsumeWhenTheClientNeverAnsweredTest` (no 
consume result, relay remark
     "client … is not connected") → `FAILED_TO_CONSUME_MESSAGE` and the remark 
is carried into the message.
   
   Results on Temurin JDK 8 (same JDK as CI), base `9ea2ccdc2` plus this patch:
   
   - before the fix: `expected:<FAILED_TO_CONSUME_MESSAGE> but 
was:<MESSAGE_CORRUPTED>`
   - after the fix: `ProxyAdminGrpcServiceTest — Tests run: 63, Failures: 0, 
Errors: 0, Skipped: 0`
   - `mvn -B -ntp -pl proxy package` with checkstyle + spotbugs enabled: `BUILD 
SUCCESS`, spotbugs `BugInstance size is 0`
   
   Note for the reviewer: the removed test asserted the old 
one-code-for-everything behavior (its name said so); the three new
   tests cover the outcomes it lumped together.
   
   This change was prepared with AI assistance (analysis, patch and test 
drafting); the diff and every result above were
   reproduced locally by running the module's tests against the base commit and 
the patched tree.
   


-- 
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