HIVE-29397: Apply masking to TestBeeLineDriver qtest goldens when overwrite enabled#6609
Conversation
…rwrite enabled Mask output before overwriteResults in CoreBeeLineDriver, matching QTestUtil.checkCliDriverResults().
|
That ticket seems already assigned to @thomasrebele We shouldn't be picking up tickets like that, unless we confirm the assignee isn't planning to work on it |
Hi @ayushtkn thanks for flagging this. Yesterday I was checking with @thomasrebele for the open Hive Jira's to be worked on, he suggested HIVE-29397(along with a few others) as a good candidate for me to try. Happy to double-check with him or adjust if needed. Thanks! |
|
If you have confirmed with him, that fair enough |
|
I confirm that I have no current plans to work on HIVE-29397. I thought that I'll look into it when I had created the ticket, but then I got busy with other tickets. Sorry for the inconvenience. Thank you, @cyanzheng2926, for picking it up! |
|
The failed test (TestReplicationScenariosAcidTables#testSizeOfDatabaseReplicationViaDistCp) looks unrelated to this PR and could be flaky. @ayushtkn can you please help to re-trigger the pre-commit when you have a moment? Thanks! |
|
thomasrebele
left a comment
There was a problem hiding this comment.
Thank you for the PR, @cyanzheng2926! I've checked it with the test case from the description of HIVE-29397. I confirm it works as expected.
| System.err.println(">>> FILTERED " + qFile.getName() + ": " + (filterEndTime - queryEndTime) | ||
| + "ms"); | ||
| qOutProcessor.maskPatterns(qFile.getOutputFile().getPath()); | ||
|
|
There was a problem hiding this comment.
There's an extraneous linebreak on line 241.



What changes were proposed in this pull request?
Moved
qOutProcessor.maskPatterns()underCoreBeeLineDriver.runTest()outside of theif (!overwrite)path so it runsbefore both the compare and overwrite paths, matching
QTestUtil.checkCliDriverResults().Why are the changes needed?
Previously masking only run when
test.output.overwrite=false. With-Dtest.output.overwrite=true, unmasked output was generated to.q.out, causing normal test runs to fail after masking was applied.Does this PR introduce any user-facing change?
No.
How was this patch tested?
Reproduced and tested as in the Jira description for HIVE-29397 with
desc_table_formatted.q: