Name: dd4877 Date: 07/21/2004
12345678901234567890123456789012345678901234567890123456789012345678901234567890
Tom R and I exchanged the following e-mails:
>> Tom Rodriguez wrote:
>>
>>> I was looking through some code and noticed that your change for 4820592
>>> made vframeStreamComment::fill_from_frame virtual, so I went looking to
>>> see why that needed to be done. I can't find any overriding of that
>>> method so I was wondering why it was made virtual. It's not a big deal,
>>> I'm just confused by the change.
>>>
>>> tom
>
> Daniel Daugherty wrote:
>
>> Tom,
>>
>> Good catch. I went back through my notes. I made fill_from_frame virtual
>> because I had originally overridden the method in the Forte specific goop.
>> During the test,refine,test process I moved all the code that made the Forte
>> version unique to other places. I removed the overriding method and forgot
>> to remove virtual from the original declaration.
>>
>> Sorry for the mess. Do you want me to remove it or are you changing code
>> in the area?
>>
>> Dan
> I figured it was something like that. If you wouldn't mind cleaning it up
> at some point that would be great. Virtual dispatch kind of sucks on sparc
> processors, so I like to avoid it. I was just reading some assembly from
> a crash trying to make sense of it and was expecting fill_from_frame to be
> be inlined so it confused me for a little while. Thanks!
>
> tom
======================================================================
- relates to
-
JDK-4820592 AsyncGetCallTrace: assert(pc_desc != 0, "scopeDesc must exist")
- Resolved