Enable V8BasedStructuredClone by default.

Enable V8BasedStructuredClone by default.

Layout test expected results are changed for the following reasons:

- Some exception messages are formatted slightly differently
  (e.g. to convey the type of the object which failed to be cloned).

- blink::ScriptValueSerializer emitted frequent "object count validation"
  tags (the most frequently seen is 3f00) which are not required by
  either deserializer, but the new one does not.

- Excessive depth throws a RangeError, like JSON.stringify and other
  functions that internally reach excessive depth (this is consistent
  with Safari, and similar to how Edge also throws the same exception as
  for an ordinary stack overflow in script).

- Certain large numbers are encoded differently; specifically, V8 now
  chooses between signed-int and double representations depending on
  whether the value has a small-integer (SMI) encoding in memory (this
  value range varies by CPU architecture). The old implementation tried
  to fit in a signed-int, then tried to fit in an unsigned-int, and only
  used double representation if that failed. Both implementations can
  read all representations, so this is a compatible change. Tests that
  are sensitive to variation by CPU architecture are commented out; there
  are unit tests that ensure that this serialization works correctly.

BUG=148757
Committed: https://crrev.com/19b847a9301f7139b7ff85cb32eb13e514fa1ab1
Cr-Commit-Position: refs/heads/master@{#434685}

[jbroman, 2016-11-19 21:33:49]

The CQ bit was checked by jbroman@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-11-19 21:34:03]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[jbroman, 2016-11-19 22:37:54]

The CQ bit was unchecked by jbroman@chromium.org

[jbroman, 2016-11-20 15:36:09]

The CQ bit was checked by jbroman@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-11-20 15:36:20]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[jbroman, 2016-11-20 15:42:03]

The CQ bit was checked by jbroman@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-11-20 15:42:18]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[jbroman, 2016-11-20 15:53:15]

Description was changed from

==========
Enable V8BasedStructuredClone by default.

This should be good to go for M57.

BUG=148757
==========

to

==========
Enable V8BasedStructuredClone by default.

Layout test expected results are changed for the following reasons:
- some exception messages are formatted slightly differently
  (e.g. to convey the type of the object which failed to be cloned)
- blink::ScriptValueSerializer emitted frequent "object count validation"
  tags (the most frequently seen is 3f00) which are not required by
  either deserializer, but the new one does not
- excessive depth throws a RangeError, like JSON.stringify and other
  functions that internally reach excessive depth (this is consistent
  with Safari, and similar to how Edge also throws the same exception as
  for an ordinary stack overflow in script)
- 

BUG=148757
==========

[jbroman, 2016-11-20 15:54:26]

Description was changed from

==========
Enable V8BasedStructuredClone by default.

Layout test expected results are changed for the following reasons:
- some exception messages are formatted slightly differently
  (e.g. to convey the type of the object which failed to be cloned)
- blink::ScriptValueSerializer emitted frequent "object count validation"
  tags (the most frequently seen is 3f00) which are not required by
  either deserializer, but the new one does not
- excessive depth throws a RangeError, like JSON.stringify and other
  functions that internally reach excessive depth (this is consistent
  with Safari, and similar to how Edge also throws the same exception as
  for an ordinary stack overflow in script)
- 

BUG=148757
==========

to

==========
Enable V8BasedStructuredClone by default.

Layout test expected results are changed for the following reasons:
- some exception messages are formatted slightly differently
  (e.g. to convey the type of the object which failed to be cloned)
- blink::ScriptValueSerializer emitted frequent "object count validation"
  tags (the most frequently seen is 3f00) which are not required by
  either deserializer, but the new one does not
- excessive depth throws a RangeError, like JSON.stringify and other
  functions that internally reach excessive depth (this is consistent
  with Safari, and similar to how Edge also throws the same exception as
  for an ordinary stack overflow in script)

BUG=148757
==========

[commit-bot: I haz the power, 2016-11-20 18:46:48]

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

[commit-bot: I haz the power, 2016-11-20 18:46:49]

Dry run: Try jobs failed on following builders:
  win_chromium_rel_ng on master.tryserver.chromium.win (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.win/builders/win_chromium_rel_...)

[jbroman, 2016-11-21 16:22:20]

The CQ bit was checked by jbroman@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-11-21 16:23:01]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[jbroman, 2016-11-21 16:26:11]

Description was changed from

==========
Enable V8BasedStructuredClone by default.

Layout test expected results are changed for the following reasons:
- some exception messages are formatted slightly differently
  (e.g. to convey the type of the object which failed to be cloned)
- blink::ScriptValueSerializer emitted frequent "object count validation"
  tags (the most frequently seen is 3f00) which are not required by
  either deserializer, but the new one does not
- excessive depth throws a RangeError, like JSON.stringify and other
  functions that internally reach excessive depth (this is consistent
  with Safari, and similar to how Edge also throws the same exception as
  for an ordinary stack overflow in script)

BUG=148757
==========

to

==========
Enable V8BasedStructuredClone by default.

Layout test expected results are changed for the following reasons:

- Some exception messages are formatted slightly differently
  (e.g. to convey the type of the object which failed to be cloned).

- blink::ScriptValueSerializer emitted frequent "object count validation"
  tags (the most frequently seen is 3f00) which are not required by
  either deserializer, but the new one does not.

- Excessive depth throws a RangeError, like JSON.stringify and other
  functions that internally reach excessive depth (this is consistent
  with Safari, and similar to how Edge also throws the same exception as
  for an ordinary stack overflow in script).

- Certain large numbers are encoded differently; specifically, V8 now
  chooses between signed-int and double representations depending on what
  whether the value has a small-integer (SMI) encoding in memory (this
  value range varies by CPU architecture). The old implementation tried
  to fit in a signed-int, then tried to fit in an unsigned-int, and only
  used double representation if that failed. Both implementations can
  read all representations, so this is a compatible change. Tests that
  are sensitive to variation by CPU architecture are commented out; there
  are unit tests that ensure that this serialization works correctly.

BUG=148757
==========

[jbroman, 2016-11-21 18:57:04]

The CQ bit was checked by jbroman@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-11-21 18:57:12]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-11-21 20:30:32]

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

[commit-bot: I haz the power, 2016-11-21 20:30:33]

Dry run: Try jobs failed on following builders:
  mac_chromium_rel_ng on master.tryserver.chromium.mac (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.mac/builders/mac_chromium_rel_...)

[jbroman, 2016-11-21 20:50:14]

The CQ bit was checked by jbroman@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-11-21 20:50:40]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-11-21 22:48:59]

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

[commit-bot: I haz the power, 2016-11-21 22:49:00]

Dry run: This issue passed the CQ dry run.

[jbroman, 2016-11-21 22:53:10]

jbroman@chromium.org changed reviewers:
+ esprehn@chromium.org, haraken@chromium.org

[jbroman, 2016-11-21 22:53:11]

I don't consider this to be blocked by
https://bugs.chromium.org/p/chromium/issues/detail?id=667260, which is an issue
in V8's unicode handling generally.

[esprehn, 2016-11-21 22:55:41]

Description was changed from

==========
Enable V8BasedStructuredClone by default.

Layout test expected results are changed for the following reasons:

- Some exception messages are formatted slightly differently
  (e.g. to convey the type of the object which failed to be cloned).

- blink::ScriptValueSerializer emitted frequent "object count validation"
  tags (the most frequently seen is 3f00) which are not required by
  either deserializer, but the new one does not.

- Excessive depth throws a RangeError, like JSON.stringify and other
  functions that internally reach excessive depth (this is consistent
  with Safari, and similar to how Edge also throws the same exception as
  for an ordinary stack overflow in script).

- Certain large numbers are encoded differently; specifically, V8 now
  chooses between signed-int and double representations depending on what
  whether the value has a small-integer (SMI) encoding in memory (this
  value range varies by CPU architecture). The old implementation tried
  to fit in a signed-int, then tried to fit in an unsigned-int, and only
  used double representation if that failed. Both implementations can
  read all representations, so this is a compatible change. Tests that
  are sensitive to variation by CPU architecture are commented out; there
  are unit tests that ensure that this serialization works correctly.

BUG=148757
==========

to

==========
Enable V8BasedStructuredClone by default.

Layout test expected results are changed for the following reasons:

- Some exception messages are formatted slightly differently
  (e.g. to convey the type of the object which failed to be cloned).

- blink::ScriptValueSerializer emitted frequent "object count validation"
  tags (the most frequently seen is 3f00) which are not required by
  either deserializer, but the new one does not.

- Excessive depth throws a RangeError, like JSON.stringify and other
  functions that internally reach excessive depth (this is consistent
  with Safari, and similar to how Edge also throws the same exception as
  for an ordinary stack overflow in script).

- Certain large numbers are encoded differently; specifically, V8 now
  chooses between signed-int and double representations depending on
  whether the value has a small-integer (SMI) encoding in memory (this
  value range varies by CPU architecture). The old implementation tried
  to fit in a signed-int, then tried to fit in an unsigned-int, and only
  used double representation if that failed. Both implementations can
  read all representations, so this is a compatible change. Tests that
  are sensitive to variation by CPU architecture are commented out; there
  are unit tests that ensure that this serialization works correctly.

BUG=148757
==========

[esprehn, 2016-11-21 22:58:50]

I think this probably needs an intent to ship since it's web exposed for the
error messages and the depth change?

lgtm

https://codereview.chromium.org/2517813002/diff/100001/third_party/WebKit/Lay...
File
third_party/WebKit/LayoutTests/fast/dom/Window/script-tests/postmessage-clone-really-deep-array.js
(right):

https://codereview.chromium.org/2517813002/diff/100001/third_party/WebKit/Lay...
third_party/WebKit/LayoutTests/fast/dom/Window/script-tests/postmessage-clone-really-deep-array.js:5:
tryPostMessage('reallyDeepArray', true, null, RangeError);
So this does mean there's a web compat change here?

[haraken, 2016-11-22 00:59:55]

With an intent-to-ship, LGTM! There you can emphasize the great performance
improvement.

[jbroman, 2016-11-22 20:12:12]

Intent thread:
https://groups.google.com/a/chromium.org/d/topic/blink-dev/He73SYod67A/discus...

[jbroman, 2016-11-28 16:28:06]

The CQ bit was checked by jbroman@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-11-28 16:28:29]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-11-28 17:52:14]

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

[commit-bot: I haz the power, 2016-11-28 17:52:15]

Dry run: This issue passed the CQ dry run.

[jbroman, 2016-11-28 17:53:53]

The CQ bit was checked by jbroman@chromium.org

[jbroman, 2016-11-28 17:53:54]

Intent has 3xLGTM, landing.

[commit-bot: I haz the power, 2016-11-28 17:54:18]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-11-28 17:59:57]

CQ is committing da patch.

Bot data: {"patchset_id": 100001, "attempt_start_ts": 1480355633975010,
"parent_rev": "931553a0e16453d8cce9bc6e1bcacdf673ccdbc7", "commit_rev":
"00bfc25976d8cd013f6a576ac81d01fe5a5b377f"}

[commit-bot: I haz the power, 2016-11-28 18:00:22]

Committed patchset #6 (id:100001)

[commit-bot: I haz the power, 2016-11-28 18:06:08]

Description was changed from

==========
Enable V8BasedStructuredClone by default.

Layout test expected results are changed for the following reasons:

- Some exception messages are formatted slightly differently
  (e.g. to convey the type of the object which failed to be cloned).

- blink::ScriptValueSerializer emitted frequent "object count validation"
  tags (the most frequently seen is 3f00) which are not required by
  either deserializer, but the new one does not.

- Excessive depth throws a RangeError, like JSON.stringify and other
  functions that internally reach excessive depth (this is consistent
  with Safari, and similar to how Edge also throws the same exception as
  for an ordinary stack overflow in script).

- Certain large numbers are encoded differently; specifically, V8 now
  chooses between signed-int and double representations depending on
  whether the value has a small-integer (SMI) encoding in memory (this
  value range varies by CPU architecture). The old implementation tried
  to fit in a signed-int, then tried to fit in an unsigned-int, and only
  used double representation if that failed. Both implementations can
  read all representations, so this is a compatible change. Tests that
  are sensitive to variation by CPU architecture are commented out; there
  are unit tests that ensure that this serialization works correctly.

BUG=148757
==========

to

==========
Enable V8BasedStructuredClone by default.

Layout test expected results are changed for the following reasons:

- Some exception messages are formatted slightly differently
  (e.g. to convey the type of the object which failed to be cloned).

- blink::ScriptValueSerializer emitted frequent "object count validation"
  tags (the most frequently seen is 3f00) which are not required by
  either deserializer, but the new one does not.

- Excessive depth throws a RangeError, like JSON.stringify and other
  functions that internally reach excessive depth (this is consistent
  with Safari, and similar to how Edge also throws the same exception as
  for an ordinary stack overflow in script).

- Certain large numbers are encoded differently; specifically, V8 now
  chooses between signed-int and double representations depending on
  whether the value has a small-integer (SMI) encoding in memory (this
  value range varies by CPU architecture). The old implementation tried
  to fit in a signed-int, then tried to fit in an unsigned-int, and only
  used double representation if that failed. Both implementations can
  read all representations, so this is a compatible change. Tests that
  are sensitive to variation by CPU architecture are commented out; there
  are unit tests that ensure that this serialization works correctly.

BUG=148757

Committed: https://crrev.com/19b847a9301f7139b7ff85cb32eb13e514fa1ab1
Cr-Commit-Position: refs/heads/master@{#434685}
==========

[commit-bot: I haz the power, 2016-11-28 18:06:09]

Patchset 6 (id:??) landed as
https://crrev.com/19b847a9301f7139b7ff85cb32eb13e514fa1ab1
Cr-Commit-Position: refs/heads/master@{#434685}