Allow multianewarray for 1D arrays on Z - #17714
Conversation
|
Personal build 17174 passed sanity.functional I used a test case with asm to exercise multianewarray for a 1D array and logged the compile. The following line in the log indicates the new code is active:
Bytecode for the test method: The generated code was: |
There was a problem hiding this comment.
Do we need a comment here? I think text in the Assert message is clear about disabling the inline allocation for multianewarray and also the comment on function makes it clear that for <1 dimensional array it calls helper.
There was a problem hiding this comment.
thanks, deleted
r30shah
left a comment
There was a problem hiding this comment.
Thanks a lot @JamesKingdon for fixing this on Z. Changes looks good to me. Left a comment about small nit-pick.
@joransiu can I request you to launch test on this PR and merge this ?
joransiu
left a comment
There was a problem hiding this comment.
LGTM. I agree the comment Rahil highlighted is unnecessary.
The existing multianewarray evaluator assumes at least 2 dimensions which is consistent with normal usage. However the spec permits 1D arrays and this is exploited by some languages, e.g. Groovy.
b77b2d9 to
b67e428
Compare
|
Jenkins test sanity zlinux jdk11,jdk17 |
The existing multianewarray evaluator assumes at least 2 dimensions which is consistent with normal usage. However the spec permits 1D arrays and this is exploited by some languages, e.g. Groovy.
See #17360 for more background and #17358 for the corresponding change on X