Move the async write handling from QuicChromiumClientStream to the Handle

Move the async write handling from QuicChromiumClientStream to the Handle
in order to put all the async logic in the same place.

Does not actually change semantics of the Handle's write methods,
just changes the implementation.

BUG=716563
Review-Url: https://codereview.chromium.org/2880643004
Cr-Commit-Position: refs/heads/master@{#475261}
Committed: https://chromium.googlesource.com/chromium/src/+/2ab4d878e867d169a3cae4c1f006955ce0439357

[Ryan Hamilton, 2017-05-12 21:06:54]

rch@chromium.org changed reviewers:
+ xunjieli@chromium.org

[xunjieli, 2017-05-15 15:48:37]

LGTM mod two nits on the documentation.

https://codereview.chromium.org/2880643004/diff/1/net/quic/chromium/quic_chro...
File net/quic/chromium/quic_chromium_client_stream.h (right):

https://codereview.chromium.org/2880643004/diff/1/net/quic/chromium/quic_chro...
net/quic/chromium/quic_chromium_client_stream.h:227: // Writes |data| to the
peer, loses the write side if |fin| is true and
nit: loses/closes

https://codereview.chromium.org/2880643004/diff/1/net/quic/chromium/quic_chro...
net/quic/chromium/quic_chromium_client_stream.h:227: // Writes |data| to the
peer, loses the write side if |fin| is true and
minor nit: Maybe break the sentence after "and" or remove "Returns false
otherwise." 

e.g. 

Writes |data| to the peer and closes the write side if |fin| is true.
Returns true if the data have been fully written.

https://codereview.chromium.org/2880643004/diff/1/net/quic/chromium/quic_chro...
net/quic/chromium/quic_chromium_client_stream.h:228: // returns true if the data
could actually be written. Returns false
nit: "could actually be written" isn't as precise. Maybe consider rephrasing?

maybe s/could actually be written/have been fully written?

Could you also mention that if the data cannot be fully written synchronously,
OnCanWrite() will be invoked later?

[Ryan Hamilton, 2017-05-20 22:24:54]

Thanks

https://codereview.chromium.org/2880643004/diff/1/net/quic/chromium/quic_chro...
File net/quic/chromium/quic_chromium_client_stream.h (right):

https://codereview.chromium.org/2880643004/diff/1/net/quic/chromium/quic_chro...
net/quic/chromium/quic_chromium_client_stream.h:227: // Writes |data| to the
peer, loses the write side if |fin| is true and
On 2017/05/15 15:48:36, xunjieli wrote:
> nit: loses/closes

Done.

https://codereview.chromium.org/2880643004/diff/1/net/quic/chromium/quic_chro...
net/quic/chromium/quic_chromium_client_stream.h:227: // Writes |data| to the
peer, loses the write side if |fin| is true and
On 2017/05/15 15:48:36, xunjieli wrote:
> minor nit: Maybe break the sentence after "and" or remove "Returns false
> otherwise." 
> 
> e.g. 
> 
> Writes |data| to the peer and closes the write side if |fin| is true.
> Returns true if the data have been fully written.

Done.

https://codereview.chromium.org/2880643004/diff/1/net/quic/chromium/quic_chro...
net/quic/chromium/quic_chromium_client_stream.h:228: // returns true if the data
could actually be written. Returns false
On 2017/05/15 15:48:37, xunjieli wrote:
> nit: "could actually be written" isn't as precise. Maybe consider rephrasing?
> 
> maybe s/could actually be written/have been fully written?
> 
> Could you also mention that if the data cannot be fully written synchronously,
> OnCanWrite() will be invoked later?

Done.

[xunjieli, 2017-05-22 12:27:51]

lgtm

[Ryan Hamilton, 2017-05-28 14:04:01]

The CQ bit was checked by rch@chromium.org

[commit-bot: I haz the power, 2017-05-28 14:04:08]

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, 2017-05-28 15:15:43]

CQ is committing da patch.

Bot data: {"patchset_id": 40001, "attempt_start_ts": 1495980241139090,
"parent_rev": "4a1a83be8f109eb727dd9e8724cce49801279faf", "commit_rev":
"2ab4d878e867d169a3cae4c1f006955ce0439357"}

[commit-bot: I haz the power, 2017-05-28 15:15:56]

Description was changed from

==========
Move the async write handling from QuicChromiumClientStream to the Handle
in order to put all the async logic in the same place.

Does not actually change semantics of the Handle's write methods,
just changes the implementation.

BUG=716563
==========

to

==========
Move the async write handling from QuicChromiumClientStream to the Handle
in order to put all the async logic in the same place.

Does not actually change semantics of the Handle's write methods,
just changes the implementation.

BUG=716563

Review-Url: https://codereview.chromium.org/2880643004
Cr-Commit-Position: refs/heads/master@{#475261}
Committed:
https://chromium.googlesource.com/chromium/src/+/2ab4d878e867d169a3cae4c1f006...
==========

[commit-bot: I haz the power, 2017-05-28 15:15:59]

Committed patchset #3 (id:40001) as
https://chromium.googlesource.com/chromium/src/+/2ab4d878e867d169a3cae4c1f006...