Thanks, Dean!
Coleen
On 4/23/20 4:16 PM, Dean Long wrote:
OK, thanks, looks good!
dl
On 4/22/20 7:32 PM, coleen.phillim...@oracle.com wrote:
On 4/22/20 9:00 PM, Dean Long wrote:
It looks like calling the JVMCI getSourceFileName() on an array
would have accessed random memory because it
OK, thanks, looks good!
dl
On 4/22/20 7:32 PM, coleen.phillim...@oracle.com wrote:
On 4/22/20 9:00 PM, Dean Long wrote:
It looks like calling the JVMCI getSourceFileName() on an array would
have accessed random memory because it was expecting an
InstanceKlass. Instead of returning null we
Thanks, Ioi!
Coleen
On 4/23/20 12:43 PM, Ioi Lam wrote:
Hi Coleen,
The changes look good to me.
Thanks
- Ioi
On 4/21/20 1:12 PM, coleen.phillim...@oracle.com wrote:
Summary: moved fields around and some constant fields into ConstantPool
This is a simple change except that I moved some
On 4/22/20 9:00 PM, Dean Long wrote:
It looks like calling the JVMCI getSourceFileName() on an array would
have accessed random memory because it was expecting an
InstanceKlass. Instead of returning null we might want to throw an
exception like in HotSpotResolvedPrimitiveType.
It was never
It looks like calling the JVMCI getSourceFileName() on an array would
have accessed random memory because it was expecting an InstanceKlass.
Instead of returning null we might want to throw an exception like in
HotSpotResolvedPrimitiveType.
dl
On 4/22/20 5:39 PM, Dean Long wrote:
Can you
Can you compare the result to some string, like "Object.java"? That
seems to be what HotSpotResolvedObjectTypeTest.java is doing.
Also, did getSourceFileName() return null for arrays before your change?
dl
On 4/22/20 8:18 AM, coleen.phillim...@oracle.com wrote:
Hi Dean, Thank you for
Hi Dean, Thank you for looking at the JVMCI changes and the suggestion
to add the test. I did this and found a bug. The new test is quite
limited because there's no good test to see if a source file name can
assertNotNull(type.getSourceFileName()), so I couldn't iterate through
the list
Hi Coleen. The JVMCI changes look OK. It looks like there is a Graal
unittest that covers getSourceFileName, but those tests don't always get
run. If it's not too much trouble, could you look into enabling
getSourceFileName() testing in
On 4/21/20 4:41 PM, Chris Plummer wrote:
Hi Coleen,
The SA changes look good. BTW, I've already made the
TestIntConstant.java change in the Loom project.
InstanceKlass::_misc_is_unsafe_anonymous was already changed to "8" in
loom. You might want to sync up with Ron to make sure you aren't
On 4/21/20 4:41 PM, Chris Plummer wrote:
Hi Coleen,
The SA changes look good. BTW, I've already made the
TestIntConstant.java change in the Loom project.
InstanceKlass::_misc_is_unsafe_anonymous was already changed to "8" in
loom. You might want to sync up with Ron to make sure you aren't
Hi Coleen,
The SA changes look good. BTW, I've already made the
TestIntConstant.java change in the Loom project.
InstanceKlass::_misc_is_unsafe_anonymous was already changed to "8" in
loom. You might want to sync up with Ron to make sure you aren't making
conflicting changes in this area.
Summary: moved fields around and some constant fields into ConstantPool
This is a simple change except that I moved some constant fields from
InstanceKlass into the constant pool so they can be shared read-only in
the CDS archive. There are associated repercussions in SA and JVMCI, so
please
12 matches
Mail list logo
