[api,heap] Fix external GC callbacks.

[api,heap] Fix external GC callbacks.

* Add types to properly report what has been executed in the GC
* Unify GCPrologueCallback and GCEpilogueCallback into GCCallback
* Report processing of second round weak handels, either synchronously or asynchronously

BUG=chromium:521946
LOG=N

Committed: https://crrev.com/8aef442917e82491f0fc66426f71f865353d94e2
Cr-Commit-Position: refs/heads/master@{#30218}

[Michael Lippautz, 2015-08-18 09:01:14]

mlippautz@chromium.org changed reviewers:
+ haraken@chromium.org, hpayer@chromium.org, vogelheim@chromium.org

[Michael Lippautz, 2015-08-18 09:01:15]

{GCPrologueCallback} and {GCEpilogueCallback} and their corresponding
{CallbackPair}s are exactly the same, respectively. We never differentiate
between them. From V8's perspective it's a callback and all that matters is the
signature. IMHO we should unify them, wdyt?

https://codereview.chromium.org/1298113003/diff/40001/src/global-handles.cc
File src/global-handles.cc (right):

https://codereview.chromium.org/1298113003/diff/40001/src/global-handles.cc#n...
src/global-handles.cc:848: isolate()->heap()->CallGCPrologueCallbacks(
I am not particularly happy with this dependency. Suggestions?

[haraken, 2015-08-18 09:09:18]

A change to include/v8.h looks good from the binding perspective. Thanks!

[Hannes Payer (out of office), 2015-08-18 09:18:13]

lgtm

https://codereview.chromium.org/1298113003/diff/40001/include/v8.h
File include/v8.h (right):

https://codereview.chromium.org/1298113003/diff/40001/include/v8.h#newcode4963
include/v8.h:4963: kGCTypeIncremental = 1 << 2,
kGCTypeIncrementalMarking maybe?

https://codereview.chromium.org/1298113003/diff/40001/include/v8.h#newcode4978
include/v8.h:4978: typedef void (*GCCallback)(GCType type, GCCallbackFlags
flags);
We should deprecated the corresponding API functions.

[Michael Lippautz, 2015-08-18 10:03:08]

https://codereview.chromium.org/1298113003/diff/40001/include/v8.h
File include/v8.h (right):

https://codereview.chromium.org/1298113003/diff/40001/include/v8.h#newcode4963
include/v8.h:4963: kGCTypeIncremental = 1 << 2,
On 2015/08/18 09:18:13, Hannes Payer wrote:
> kGCTypeIncrementalMarking maybe?

Done.

https://codereview.chromium.org/1298113003/diff/40001/include/v8.h#newcode4978
include/v8.h:4978: typedef void (*GCCallback)(GCType type, GCCallbackFlags
flags);
On 2015/08/18 09:18:13, Hannes Payer wrote:
> We should deprecated the corresponding API functions.

The typedefs are completely (read: 100%) compatible, so I cannot add overloaded
methods. I changed the methods to use the new types and added deprecation
warnings to the old ones.

[vogelheim, 2015-08-18 10:05:51]

lgtm (Looks good, API-wise.)

Please note that we're presently in API freeze until the M46 branch is cut. That
means your change - which is 100% backwards compatible - is fine, but a *use* of
this change by an embedder - which I'd assume is the next step - is not. The
reason is that in the release phase Chrome releng has oftentimes found it
necessary to rollback to an older version of V8; and once a Blink-side change
would make use of the new typedefs or enums that would no longer be possible.

So if you are planning a matching Blink change, please do hold off with that
change until after the branch, next week.

[Michael Lippautz, 2015-08-18 10:32:09]

Thanks a lot! 

Let's wait with the blink changes until M46 branch is cut then.

[Michael Lippautz, 2015-08-18 10:34:53]

The CQ bit was checked by mlippautz@chromium.org

[Michael Lippautz, 2015-08-18 10:34:53]

The patchset sent to the CQ was uploaded after l-g-t-m from hpayer@chromium.org
Link to the patchset: https://codereview.chromium.org/1298113003/#ps60001
(title: "Addressed comments")

[commit-bot: I haz the power, 2015-08-18 10:35:05]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1298113003/60001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1298113003/60001

[commit-bot: I haz the power, 2015-08-18 10:41:15]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2015-08-18 10:41:15]

Try jobs failed on following builders:
  v8_linux_gcc_compile_rel on tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux_gcc_compile_rel/bu...)

[Michael Lippautz, 2015-08-18 11:38:29]

The CQ bit was checked by mlippautz@chromium.org

[Michael Lippautz, 2015-08-18 11:38:30]

The patchset sent to the CQ was uploaded after l-g-t-m from hpayer@chromium.org,
vogelheim@chromium.org
Link to the patchset: https://codereview.chromium.org/1298113003/#ps140001
(title: "Got rid of too many V8::Type vs Isolate::Type casts")

[commit-bot: I haz the power, 2015-08-18 11:38:38]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1298113003/140001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1298113003/140001

[commit-bot: I haz the power, 2015-08-18 12:03:51]

Committed patchset #8 (id:140001)

[commit-bot: I haz the power, 2015-08-18 12:04:03]

Patchset 8 (id:??) landed as
https://crrev.com/8aef442917e82491f0fc66426f71f865353d94e2
Cr-Commit-Position: refs/heads/master@{#30218}

[haraken, 2015-08-19 13:11:21]

https://codereview.chromium.org/1298113003/diff/140001/src/heap/heap.cc
File src/heap/heap.cc (right):

https://codereview.chromium.org/1298113003/diff/140001/src/heap/heap.cc#newco...
src/heap/heap.cc:776: // filtering.

CL uploaded here: https://codereview.chromium.org/1288683005/