liunan-ms opened a new pull request, #239:
URL: https://github.com/apache/ant/pull/239

   ## Summary
   Make the ZIP compression-level test compare explicit levels 1 and 9 instead 
of default compression against level 9.
   
   `testCompressionLevel` depends on `test6` to create the baseline archive and 
asserts that the level-9 archive is strictly smaller. With Microsoft OpenJDK 25 
in our environment, the default-level and level-9 archives have the same size, 
so the existing assertion fails. A higher compression level does not guarantee 
smaller output for every input.
   
   This applies the downstream test-only patch unchanged. It sets `level="1"` 
on the baseline archive and retains the size assertion. It does not change 
Ant's production compression behavior. This remains a fixture-dependent size 
check, not a guarantee of size ordering for arbitrary data.
   
   ## Validation
   Using Microsoft OpenJDK 25 (`25.0.4.1+1-LTS`) on Fedora 43 with system 
zlib-ng:
   - Bootstrapped upstream Ant from `master`.
   - Ran the original `testCompressionLevel` target and reproduced the failure.
   - Applied the patch, cleaned the test output, and reran the target 
successfully.
   
   Command: `JAVA_HOME=/usr/lib/jvm/msopenjdk-25 bootstrap/bin/ant -q -f 
src/etc/testcases/taskdefs/zip.xml tearDown setUp testCompressionLevel tearDown`
   
   The full JUnit suite was not run. The same change was previously validated 
in an Azure Linux Ant 1.10.15 package build with Microsoft OpenJDK 25.
   
   Microsoft's JDK links to system zlib in the environments inspected, while 
Fedora's JDK packaging defaults to bundled zlib. This may explain the differing 
results, but the library difference has not been conclusively isolated as the 
cause.
   


-- 
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