RFR (S): 8189797:Fix initializer lists in G1ConcurrentMark and G1CMTask

[Kim Barrett]

> On Oct 23, 2017, at 12:02 PM, Thomas Schatzl <thomas.schatzl at oracle.com> wrote:
> 
> Hi all,
> 
>  can I have quick reviews for this change that fixes initializer lists
> for the aforementioned classes?
> 
> It just rearranges their members so they reflect the actual
> initialization order as mandated by the c++ standard. This is just
> nicer for anyone trying to think about initialization order. There is
> no actual bug fixed.
> 
> CR:
> https://bugs.openjdk.java.net/browse/JDK-8189797
> Webrev:
> http://cr.openjdk.java.net/~tschatzl/8189797/webrev/
> Testing:
> local compilation
> 
> Thanks,
>  Thomas

I like this change.  Some time ago I tried the experiment of turning
on -Wreorder, and got lots of warnings.  I don't remember whether I
filed a bug for that, and JIRA seems to not be responding to me right
now. 

Just a couple of very minor comments.  I don't need a new webrev for
these.

_cm_thread not commented in initializer list as being initialized
later.

When there's an initializer list, especially a long one like this, I
like to put the opening brace for the constructor body at the
beginning of its own line, in order to make it easier to spot where
the initializer list ends and the constructor body begins.