RFR: several small changes including 8207040: [lworld] TestLWorld fails with "Meet Not Symmetric" with -Xcomp

[Roland Westrelin]

>> http://cr.openjdk.java.net/~roland/8207040/webrev.00/
>
> Looks good to me.

Thanks for the review.

>> - checkcast buffers values when it's not needed. The change delays
>>   allocation. I actually wonder if the path where we allocate is seen in
>>   practice and if it is if we need to allocate at all.
>
> Did you try to add an assert and check if it's triggered with our test?

I tried it. That code path is executed at parse time.

>> - the Parse::merge_common sometimes buffers a value when there's no
>>   merge point which is unnecessary so I changed that logic too.
>
> Looks like the checks you are added are the ones used further below (line 1767) but without the
> control()->is_Region() check. Do you know why it's not needed?

I'm not sure but from the comment it doesn't seem to be part of the
logic that detects multiple predecessors.

> Please run all compiler tests with and without -XX:CompileTreshold=100 before pushing.

Tests pass except for TestNewAcmp.java that fails with
InvocationTargetException but that's a known issue.

Roland.