RFR: Bulk integration from jdk/jdk 2018-04-23
Roman Kennke
rkennke at redhat.com
Mon Apr 23 21:02:33 UTC 2018
Am 23.04.2018 um 23:00 schrieb Aleksey Shipilev:
> On 04/23/2018 10:48 PM, Roman Kennke wrote:
>> Looks like I've messed up the merge (again). I fixed it by hg stripping
>> up to the merge changesets, fixing some whitespace stuff, hg commit
>> --amend the changes right into the same changeset. I think it should be
>> safe to push this.
>>
>> Output of:
>> hg diff -r 49794:50054 src/hotspot/share/gc/shenandoah
>> src/hotspot/cpu/x86/gc/shenandoah
>>
>> http://cr.openjdk.java.net/~rkennke/merge-2018-04-23/shenandoah-changes-02.patch
>
> I see there are debugging leftovers, like:
>
> + // tty->print_cr("RB src.base: %s", src.base()->name());
> + // __ verify_oop(src.base(), "broken oop before RB");
> + /*
> + if (in_heap) {
> + read_barrier_not_null(masm, src.base());
> + }
> + */
> + // __ verify_oop(src.base(), "broken oop before RB");
>
> Other than that, it looks good. We can clean up the debugging stuff after the merge.
Oops. But I'll be poking around in this code later anyway, and this
debugging stuff might turn out to be useful. And the commented out
read_barrier... blah will become used too. So yeah, let's fix this for
real later.
Trying to push now. Crossing fingers. (what a nightmare merge...)
Roman
More information about the shenandoah-dev
mailing list