Open manashaVetrivelu opened 2 years ago
Attn: @0xdaryl
@dylanjtuttle : please investigate
Brad @BradleyWood, may I ask you to take a look at this problem?
Oops! Sorry for the noise, Brad @BradleyWood - I had intended to assign this one to myself, as I had looked at a problem with a similar symptom in the past.
I thought this might be a duplicate of Issue #15311. However, I am able to reproduce the "Cannot intersect constraints on acalli" assertion failure both with and without the fix contained in eclipse/omr#6619. I will continue to investigate.
It looks like the problem occurs effectively with a call to Class.newInstance()
on Byte.TYPE
, which should lead to an exception since an instance of a primitive type can't be instantiated with Class.newInstance()
.
As this is only affecting Sanity.system
and we are focusing on sanity.functional
for the 0.41 release, I'll move this out.
In the
Sanity.system
test suite, The assert at/home/jenkins/workspace/Build_JDK11_x86-64_linux_Personal/omr/compiler/optimizer/OMRValuePropagation.cpp:659: c
failed when it was compiled withPROD_WITH_ASSUMES
enabled. It was reproducible in all of the tests mentioned below. It failed during the following tests: MauveMultiThrdLoad_5m_0 MauveMultiThrdLoad_5m_1The platform used to build and test was
x86-64_linux
. The link to the Jenkins Build. The link to the Grinder Build.The stack trace for each of the tests are:
MauveMultiThrdLoad_5m_0 with the stack trace:
MauveMultiThrdLoad_5m_1 twice with the trace:
and