github-actions[bot] commented on code in PR #64755:
URL: https://github.com/apache/doris/pull/64755#discussion_r3471604064


##########
regression-test/suites/fault_injection_p0/test_index_compound_directory_fault_injection.groovy:
##########
@@ -178,7 +178,10 @@ suite("test_index_compound_directory_fault_injection", 
"nonConcurrent") {
                 res = sql "select COUNT() from 
${test_index_compound_directory} where request match 'gif'"
                 try_sql("DROP TABLE IF EXISTS 
${test_index_compound_directory}")
             } catch(Exception ex) {
-                assertTrue(ex.toString().contains("failed to initialize 
storage reader"))
+                // _tablet_reader->init() now returns the underlying error 
verbatim and only
+                // appends the tablet/backend context (no longer the misleading
+                // "failed to initialize storage reader" wording).
+                assertTrue(ex.toString().contains("tablet=") && 
ex.toString().contains("backend="))

Review Comment:
   The new message check is still optional because it only runs from this 
`catch`. If the debug-point read unexpectedly succeeds, execution goes through 
the `try_sql("DROP TABLE...")` path and the test never asserts that the 
scanner-open path failed. Even when it does throw, checking only `tablet=` and 
`backend=` would still pass if the old `"failed to initialize storage reader. 
tablet=..., backend=..."` wording came back, since that message also had both 
fields. Since this PR is relying on this regression to lock in the new 
`_open_impl()` wording, please make the failure mandatory and assert the old 
phrase is absent, or assert a stable underlying injected error together with 
the new tablet/backend context.



-- 
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]


---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to