[prev in list] [next in list] [prev in thread] [next in thread] 

List:       openjdk-hotspot-runtime-dev
Subject:    request for review
From:       Keith.McGuigan () Sun ! COM (Keith McGuigan)
Date:       2008-04-28 14:54:19
Message-ID: 4815E51B.9010505 () sun ! com
[Download RAW message or body]

Hello,

I'd appreciate any review on this small change that I'd like to get into 
HS13 to resolve a P2:


Fixed 6692246: Regression : JDK 6u4 b01 fails two JCK tests when 
fallback is switched off

The type-checking verifier incorrectly rejected classfiles in which it 
could
determine statically that a 'null' was passed to an arraylength opcode.
According to the type-checking specification, 'null' is a valid type to
arraylength (and in fact, the inference verifier accepts this, which is
why these tests only failed when fallback was switched off).

The fix is rather trivial, just add a check for the 'null' type when 
examining
the arraylength operation and don't throw an exception in that case.

Webrev: http://webrev.invokedynamic.info/kamg/6692246/

Testing:
 + JPRT
 + NSK runtime
 + JCK vm tests
 + runThese full

Reviewed by:

Fix verified (y/n)? y
Verified by:
  + CR-attached JCK test cases



[prev in list] [next in list] [prev in thread] [next in thread]