Mandoline DevTools service: remote debugging protocol over HTTP/WebSocket.

Mandoline DevTools service: remote debugging protocol over HTTP/WebSocket.

This CL:
- adds service-side DevTools agent representation (DevToolsAgentHost) and registration for DevTools agents.
- supports requesting the list of available agents using HTTP request and communicating with them using WebSocket.

BUG=478249
TEST=None

Committed: https://crrev.com/4f1226d35342914dbbf3decb806d0592838b49c6
Cr-Commit-Position: refs/heads/master@{#333804}

[yzshen1, 2015-06-09 18:38:54]

yzshen@chromium.org changed reviewers:
+ jam@chromium.org

[yzshen1, 2015-06-09 18:38:55]

Hi, John.

Would you please take a look? Thanks!

[jam, 2015-06-09 22:21:21]

lgtm

https://codereview.chromium.org/1164383002/diff/20001/components/devtools_ser...
File components/devtools_service/devtools_agent_host.cc (right):

https://codereview.chromium.org/1164383002/diff/20001/components/devtools_ser...
components/devtools_service/devtools_agent_host.cc:17: DCHECK(agent_);
nit: this is redundant since the next line will give as useful of a signal

https://codereview.chromium.org/1164383002/diff/20001/components/devtools_ser...
components/devtools_service/devtools_agent_host.cc:48: DCHECK(delegate_);
ditto

[pfeldman, 2015-06-10 09:28:20]

pfeldman@chromium.org changed reviewers:
+ pfeldman@chromium.org

[pfeldman, 2015-06-10 09:28:21]

https://codereview.chromium.org/1164383002/diff/20001/components/devtools_ser...
File components/devtools_service/devtools_http_server.cc (right):

https://codereview.chromium.org/1164383002/diff/20001/components/devtools_ser...
components/devtools_service/devtools_http_server.cc:7: #include <string.h>
Why do we build a new universe here? Is there a way to reuse the
devtools_http_handler component?

[pfeldman, 2015-06-10 09:30:27]

pfeldman@chromium.org changed reviewers:
+ yurys@chromium.org

[yzshen1, 2015-06-10 15:34:39]

https://codereview.chromium.org/1164383002/diff/20001/components/devtools_ser...
File components/devtools_service/devtools_agent_host.cc (right):

https://codereview.chromium.org/1164383002/diff/20001/components/devtools_ser...
components/devtools_service/devtools_agent_host.cc:17: DCHECK(agent_);
On 2015/06/09 22:21:21, jam wrote:
> nit: this is redundant since the next line will give as useful of a signal

Done.

https://codereview.chromium.org/1164383002/diff/20001/components/devtools_ser...
components/devtools_service/devtools_agent_host.cc:48: DCHECK(delegate_);
On 2015/06/09 22:21:21, jam wrote:
> ditto

Done.

https://codereview.chromium.org/1164383002/diff/20001/components/devtools_ser...
File components/devtools_service/devtools_http_server.cc (right):

https://codereview.chromium.org/1164383002/diff/20001/components/devtools_ser...
components/devtools_service/devtools_http_server.cc:7: #include <string.h>
On 2015/06/10 09:28:21, pfeldman wrote:
> Why do we build a new universe here? Is there a way to reuse the
> devtools_http_handler component?

Thanks for the comment!

I feel hard to reuse devtools_http_handler directly, because we use a mojo
HTTP/WebSocket API to avoid depending on net/, and also we don't want to depend
on content/.

I am planning to reuse the protocol parsing code. It requires some refactoring
of the existing Chrome DevTools code. I will make a proposal and ask for review
from you and other DevTools experts when I start that part of work.

This document talks about the purpose of my current work:
https://docs.google.com/document/d/1HUiqvtmO5_Qfdr2DVL7MLZFu-8i3HeCG1ZNSBJUxJ...

(I have chatted with Dmitry and Yury before about this work.)

And please consider this DevTools work for Manoline is "experimental". It hasn't
reached production-level quality yet and it has no impact on the Chrome DevTools
for now. I am trying to setup the basic things so that we can run some perf
tests for Mandoline as soon as possible.

I would be happy to have a chat if you would like to know more about this work
or have more suggestions. Thanks!

[pfeldman, 2015-06-10 16:25:50]

That sounds fair, thanks for the details.

[yzshen1, 2015-06-10 20:20:25]

The CQ bit was checked by yzshen@chromium.org

[yzshen1, 2015-06-10 20:20:25]

The patchset sent to the CQ was uploaded after l-g-t-m from jam@chromium.org
Link to the patchset: https://codereview.chromium.org/1164383002/#ps60001
(title: " ")

[commit-bot: I haz the power, 2015-06-10 20:20:52]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1164383002/60001

[commit-bot: I haz the power, 2015-06-10 20:33:18]

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

[commit-bot: I haz the power, 2015-06-10 20:33:19]

Try jobs failed on following builders:
  linux_chromium_gn_chromeos_rel on tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_...)

[yzshen1, 2015-06-10 20:50:06]

The CQ bit was checked by yzshen@chromium.org

[yzshen1, 2015-06-10 20:50:07]

The patchset sent to the CQ was uploaded after l-g-t-m from jam@chromium.org
Link to the patchset: https://codereview.chromium.org/1164383002/#ps80001
(title: " ")

[commit-bot: I haz the power, 2015-06-10 20:51:45]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1164383002/80001

[commit-bot: I haz the power, 2015-06-10 21:27:07]

Committed patchset #5 (id:80001)

[commit-bot: I haz the power, 2015-06-10 21:28:01]

Patchset 5 (id:??) landed as
https://crrev.com/4f1226d35342914dbbf3decb806d0592838b49c6
Cr-Commit-Position: refs/heads/master@{#333804}