RFR/RFC: Pick up aarch64-shenandoah-jdk8u222-b02 to sh/jdk8

[Andrew John Hughes]

On 22/05/2019 07:56, Aleksey Shipilev wrote:
> Ping Andrew Hughes.
> 
> On 5/20/19 12:26 PM, Aleksey Shipilev wrote:
>> Hi,
>>
>> This is our first ever merge from aarch64-port/jdk8u-shenandoah to sh/jdk8. This picks up
>> aarch64-shenandoah-jdk8u222-b02 tag and merges it into sh/jdk8.
>>
>> Changesets (only hotspot, others are trivial):
>>   http://cr.openjdk.java.net/~shade/shenandoah/merges/jdk8-8u222-b02/changesets.txt
>>
>> Webrev (only hotspot, others are trivial):
>>   http://cr.openjdk.java.net/~shade/shenandoah/merges/jdk8-8u222-b02/webrev.01/
>>
>> Testing: hotspot_gc_shenandoah {fastdebug,release}
>>
>> Andrew Hughes, we need to agree on process here. The way it is done right now, we are doing this:
>>  a. Pull aarch64-port/jdk8u-shenandoah -> sh/jdk8, merge
>>  b. Test it, pile on our backports
>>  c. Repeat (a)-(b) until Shenandoah is stable
>>  d. RFR the pull sh/jdk8 -> aarch64-port/jdk8u-shenandoah, tag it appropriately
>>
>> Since aarch64-port/jdk8u-shenandoah and sh/jdk8 are related now, this allows us to push/pull between
>> them without any Mercurial voodoo. The caveat there is that the pull at (d) would include a few
>> local "merge" changesets from (a), are you fine with that? I also prefer not to introduce new
>> sh/jdk8-specific tags to avoid contaminating aarch64-port/jdk8u-shenandoah with them, so those
>> merges would be untagged.

That doesn't sound problematic. I'd prefer they were "Merge <tag>"
rather than "Merge" if possible. I've never understood why jcheck only
allows the latter.

>>
>> You could ask why don't we just push the backports to aarch64-port/jdk8u-shenandoah, but the trouble
>> is that 8u is a mess GC-interface-wise, and we would like to have sh/jdk8 as additional sandbox to
>> test for ourselves before integrating. It would be better as we backport more simplifications like
>> LRB and nofwdptr features into sh/jdk8u, which would get Shenandoah upstream exposure down
>> significantly. But so far it is what it is.
>>
> 

I'm happy with whatever way you're most comfortable.

Thanks,
-- 
Andrew :)

Senior Free Java Software Engineer
Red Hat, Inc. (http://www.redhat.com)

PGP Key: ed25519/0xCFDA0F9B35964222 (hkp://keys.gnupg.net)
Fingerprint = 5132 579D D154 0ED2 3E04  C5A0 CFDA 0F9B 3596 4222
https://keybase.io/gnu_andrew