RFR: 8146399: Refactor the BlockOffsetTable classes.
mikael.gerdin at oracle.com
Thu Jan 7 13:25:55 UTC 2016
On 2016-01-07 14:23, Thomas Schatzl wrote:
> On Thu, 2016-01-07 at 12:27 +0100, David Lindholm wrote:
>> Hi Mikael,
>> Thanks for looking at this! I have fixed your comments:
> - the comment for the G1BlockOffsetTable constructor is out of date
> - I would prefer if in g1BlockOffsetTable.inline.hpp, the definition of
> ck_containing_addr_const() the newline after the return type were
> - in the same file, the arguments of block_at_or_preceding() are not
> aligned properly
> - in the same file, in forward_to_block_containing_addr_const(), please
> add braces for the early exit (line 140) at the if-statement.
> - in the same file, forward_to_block_containing() also misses braces in
> the first if-statement
> - g1BlockOffsetTable.cpp, additional newline in line 203
> - in heapRegion.inline.hpp, the change in the name of the
> G1OffsetTableContigSpace causes lots of parameter specifications to be
> not conforming to code style. Could you look over this file again?
> - in case the SA agent is still around, did you check the changes in
> vmstructs_g1 with it?
Just FTR, I ran David's changes and did some cursory SA testing (jmap -heap)
> Thanks a lot for the effort.
More information about the hotspot-gc-dev