RFR(S): 8160564: TEST: Add a test to check the implementation of VersionProps.versionNumbers()

Volker Simonis volker.simonis at gmail.com
Thu Jul 7 13:59:13 UTC 2016


can I please have a review for the following change which makes
VersionProps.versionNumbers() testable and the corresponding
regression test:


During the review for "8160457: VersionProps.versionNumbers() is
broken" it was suggested to refactor VersionProps.versionNumbers() in
order to make it testable by a regression test. This change does
exactly that. It extracts the implementation of
VersionProps.versionNumbers() into a new method
parseVersionNumbers(String versionNumber) which can be tested from the
associated regression test.

There are still two points to notice:

- VersionProps.versionNumbers() deliberately doesn't check for badly
formatted version strings because it is assumed it will only get valid
input (see discussion for "JDK-8160000: Runtime.version() cause
startup regressions" at [2]). So we can currently only check that
VersionProps.versionNumbers() accepts all kind of valid version

- I was a little bit surprised that I could reflectively access and
execute java.lang.VersionProps.parseVersionNumbers() where both the
class and the method are package private. Maybe this is related to
Jigsaw issue #ReflectiveAccessToNonExportedTypes [3]? As I'm not a
Jigsaw expert, I'd be graceful to anybody explaining me why this is
still so easily possible with Jigsaw?

Thank you and best regards,

[1] http://mail.openjdk.java.net/pipermail/core-libs-dev/2016-June/042058.html
[2] https://bugs.openjdk.java.net/browse/JDK-8160000
[3] http://openjdk.java.net/projects/jigsaw/spec/issues/#ReflectiveAccessToNonExportedTypes

More information about the core-libs-dev mailing list