loss of cause??

[Kirk Pepperdine]

Hi Mikael,

I'm already using 8 (my own builds) and I've been waiting for it to settle down before committing time to GC logging. I'm sorry I missed the discussion. I follow as much as I can but some times..... So, while it seems reasonable to not turn on cause in 7, I can't see shops moving to 8 and in the mean time loss of (system) is a regression.

Regards,
Kirk

On 2013-10-10, at 4:40 PM, Mikael Gerdin <mikael.gerdin at oracle.com> wrote:

> Richard,
> 
> On 10/10/2013 04:35 PM, Richard Warburton wrote:
>> Hi,
>> 
>>    We chose to not add the GC cause to all gc log records since we felt
>>    that was too large a change for 7.
>>    The previous behavior was completely inconsistent, all the different
>>    GC:s had different ways of signaling that a full gc was caused by a
>>    Systm.gc() call and System.gc() was the only gc cause which was logged.
>> 
>>    In 8 we decided to enable gc cause printing for all gc log records.
>> 
>> 
>> Thanks for explaining the decision.  Does this mean that every (most)
>> PrintGCDetails related GC Log record type will change by default in Java
>> 8 in order to print the cause?
> 
> Every (most) is a good observation :)
> 
> The CMS concurrent mark/sweep records don't have that information since they are not triggered by a specific gc cause.
> 
> I don't remember all the corner cases off the top of my head, but have a go at a JDK 8 early access build if you want to check it out for yourself.
> 
> https://jdk8.java.net/download.html
> 
> /Mikael
> 
>> 
>> regards,
>> 
>>   Dr. Richard Warburton
>> 
>> http://insightfullogic.com
>> @RichardWarburto <http://twitter.com/richardwarburto>
>