On Mon, 23 Feb 2026 14:36:15 GMT, David Beaumont <[email protected]> wrote:
>> test/jdk/jdk/internal/jrtfs/ClassResourcesParityTest.java line 74:
>>
>>> 72: String relPath = jrtClassFile.subpath(2,
>>> jrtClassFile.getNameCount()).toString();
>>> 73: String fqn = relPath.substring(0, relPath.length() -
>>> CLASS_SUFFIX.length()).replace('/', '.');
>>> 74: String baseName = fqn.substring(fqn.lastIndexOf('.') + 1);
>>
>> This is fine, just FYI that JAVA_BASE.relativize(jrtClassFile) is another
>> slicker way to get the relative path from /module/java.base to the resource.
>
> Hmm, good point. However I don't have the `/modules/<name>` path to hand
> here, so I still need to subpath() something. So it's
>
> jrtClassFile.subpath(0, 2).relativize(jrtClassFile).toString();
>
> vs
>
> jrtClassFile.subpath(2, jrtClassFile.getNameCount()).toString();
>
> Marginal win for readability.
Actually, that caused an exception:
Path modRoot = jrtClassFile.subpath(0, 2);
String relPath = modRoot.relativize(jrtClassFile).toString();
results in:
java.lang.IllegalArgumentException: Incorrect filesystem or path:
/modules/java.base/com/sun/crypto/provider/AEADBufferedStream.class
at java.base/jdk.internal.jrtfs.JrtPath.relativize(JrtPath.java:223)
at java.base/jdk.internal.jrtfs.JrtPath.relativize(JrtPath.java:59)
at
ClassResourcesParityTest.testMatchesGetResourceAsStream(ClassResourcesParityTest.java:73)
So I'll stick with what I've got.
-------------
PR Review Comment:
https://git.openjdk.org/valhalla/pull/2032#discussion_r2841250185
