DevTools: introduce bound object on the agent side.

DevTools: Migrate to InspectorController for network and console events.
- introduced bound object on the agent side;
- established remote dispatch of WebInspector calls
- using fake InspectorFrontend for serializing events and sending them over the ipc
- removed net agents from both sides
- moved GetResource stuff to tools agent

Assumes following is landed:
https://bugs.webkit.org/show_bug.cgi?id=26010

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=16980

[yurys, 2009-05-26 11:43:21]

http://codereview.chromium.org/113836/diff/27/1025
File webkit/glue/devtools/bound_object.cc (right):

http://codereview.chromium.org/113836/diff/27/1025#newcode53
Line 53: v8::Context::Scope frame_scope(context_);
rename frame_scope to something more relevant

http://codereview.chromium.org/113836/diff/27/1026
File webkit/glue/devtools/bound_object.h (right):

http://codereview.chromium.org/113836/diff/27/1026#newcode10
Line 10: class BoundObject {
please add some comments

http://codereview.chromium.org/113836/diff/27/1029
File webkit/glue/devtools/js/devtools.js (right):

http://codereview.chromium.org/113836/diff/27/1029#newcode49
Line 49: RemoteToolsAgent.DispatchOnClient =
you can assign to WebInspector.dispatch directly

http://codereview.chromium.org/113836/diff/27/1022
File webkit/glue/devtools/js/inject_dispatch.js (right):

http://codereview.chromium.org/113836/diff/27/1022#newcode41
Line 41: dispatch = function() {
excerpt from the style guide: "Functions that take variable number of arguments
should add an argument called var_args after all the other arguments."

http://codereview.chromium.org/113836/diff/1033/1036
File webkit/glue/webdevtoolsagent_impl.cc (right):

http://codereview.chromium.org/113836/diff/1033/1036#newcode105
Line 105: utility_context_->Global());
run this on RemoteWebInspector

[pfeldman, 2009-05-26 12:43:05]

http://codereview.chromium.org/113836/diff/27/1025
File webkit/glue/devtools/bound_object.cc (right):

http://codereview.chromium.org/113836/diff/27/1025#newcode53
Line 53: v8::Context::Scope frame_scope(context_);
On 2009/05/26 11:43:21, Yury Semikhatsky wrote:
> rename frame_scope to something more relevant

Done.

http://codereview.chromium.org/113836/diff/27/1026
File webkit/glue/devtools/bound_object.h (right):

http://codereview.chromium.org/113836/diff/27/1026#newcode10
Line 10: class BoundObject {
On 2009/05/26 11:43:21, Yury Semikhatsky wrote:
> please add some comments

Done.

http://codereview.chromium.org/113836/diff/27/1029
File webkit/glue/devtools/js/devtools.js (right):

http://codereview.chromium.org/113836/diff/27/1029#newcode49
Line 49: RemoteToolsAgent.DispatchOnClient =
On 2009/05/26 11:43:21, Yury Semikhatsky wrote:
> you can assign to WebInspector.dispatch directly

Not anymore - I need to do apply there. see updated code below.

http://codereview.chromium.org/113836/diff/27/1022
File webkit/glue/devtools/js/inject_dispatch.js (right):

http://codereview.chromium.org/113836/diff/27/1022#newcode41
Line 41: dispatch = function() {
On 2009/05/26 11:43:21, Yury Semikhatsky wrote:
> excerpt from the style guide: "Functions that take variable number of
arguments
> should add an argument called var_args after all the other arguments."

Done.

http://codereview.chromium.org/113836/diff/1033/1036
File webkit/glue/webdevtoolsagent_impl.cc (right):

http://codereview.chromium.org/113836/diff/1033/1036#newcode105
Line 105: utility_context_->Global());
On 2009/05/26 11:43:22, Yury Semikhatsky wrote:
> run this on RemoteWebInspector

I fooled myself and yourself here. I can't run it on RemoteWebInspector.

[yurys, 2009-05-27 07:02:35]

LGTM

http://codereview.chromium.org/113836/diff/94/1140
File webkit/glue/devtools/bound_object.h (right):

http://codereview.chromium.org/113836/diff/94/1140#newcode10
Line 10: // BoundObject lets you map Javascript method calls and property
accesses
Javascript->JavaScript

http://codereview.chromium.org/113836/diff/1159/113
File webkit/glue/devtools/js/devtools.js (right):

http://codereview.chromium.org/113836/diff/1159/113#newcode166
Line 166: var mycallback = function(content) {
you don't need this additional wrapper, just pass opt_callback directly
devtools.Callback.wrap will take care of undefined value

http://codereview.chromium.org/113836/diff/1159/106
File webkit/glue/devtools/js/inject_dispatch.js (right):

http://codereview.chromium.org/113836/diff/1159/106#newcode42
Line 42: if (method == 'inspectedWindowCleared') {
what is that?

http://codereview.chromium.org/113836/diff/1159/123
File webkit/glue/inspector_client_impl.cc (right):

http://codereview.chromium.org/113836/diff/1159/123#newcode42
Line 42: 
remove the line

http://codereview.chromium.org/113836/diff/1159/123#newcode53
Line 53: if (inspected_web_view_->GetWebDevToolsAgentImpl())
use {} ?

http://codereview.chromium.org/113836/diff/1159/123#newcode105
Line 105: if (inspected_web_view_->GetWebDevToolsAgentImpl())
ditto

http://codereview.chromium.org/113836/diff/1159/123#newcode118
Line 118: if (inspected_web_view_->GetWebDevToolsAgentImpl())
ditto

http://codereview.chromium.org/113836/diff/1159/123#newcode130
Line 130: if (inspected_web_view_->GetWebDevToolsAgentImpl())
ditto

http://codereview.chromium.org/113836/diff/1159/104
File webkit/glue/webdevtoolsagent_impl.cc (right):

http://codereview.chromium.org/113836/diff/1159/104#newcode106
Line 106: ic->setWindowVisible(false, false);
add comment why this is needed

http://codereview.chromium.org/113836/diff/1159/104#newcode218
Line 218: int identifier) {
it should be unsigned long, isn't it?

[pfeldman, 2009-05-27 09:04:13]

http://codereview.chromium.org/113836/diff/94/1140
File webkit/glue/devtools/bound_object.h (right):

http://codereview.chromium.org/113836/diff/94/1140#newcode10
Line 10: // BoundObject lets you map Javascript method calls and property
accesses
On 2009/05/27 07:02:35, Yury Semikhatsky wrote:
> Javascript->JavaScript

Done.

http://codereview.chromium.org/113836/diff/1159/113
File webkit/glue/devtools/js/devtools.js (right):

http://codereview.chromium.org/113836/diff/1159/113#newcode166
Line 166: var mycallback = function(content) {
On 2009/05/27 07:02:36, Yury Semikhatsky wrote:
> you don't need this additional wrapper, just pass opt_callback directly
> devtools.Callback.wrap will take care of undefined value

Done.

http://codereview.chromium.org/113836/diff/1159/106
File webkit/glue/devtools/js/inject_dispatch.js (right):

http://codereview.chromium.org/113836/diff/1159/106#newcode42
Line 42: if (method == 'inspectedWindowCleared') {
On 2009/05/27 07:02:36, Yury Semikhatsky wrote:
> what is that?

Done.

http://codereview.chromium.org/113836/diff/1159/123
File webkit/glue/inspector_client_impl.cc (right):

http://codereview.chromium.org/113836/diff/1159/123#newcode42
Line 42: 
On 2009/05/27 07:02:36, Yury Semikhatsky wrote:
> remove the line

Done.

http://codereview.chromium.org/113836/diff/1159/123#newcode53
Line 53: if (inspected_web_view_->GetWebDevToolsAgentImpl())
On 2009/05/27 07:02:36, Yury Semikhatsky wrote:
> use {} ?

I am being consistent.

http://codereview.chromium.org/113836/diff/1159/123#newcode105
Line 105: if (inspected_web_view_->GetWebDevToolsAgentImpl())
On 2009/05/27 07:02:36, Yury Semikhatsky wrote:
> ditto

ditto

http://codereview.chromium.org/113836/diff/1159/123#newcode118
Line 118: if (inspected_web_view_->GetWebDevToolsAgentImpl())
On 2009/05/27 07:02:36, Yury Semikhatsky wrote:
> ditto

ditto

http://codereview.chromium.org/113836/diff/1159/123#newcode130
Line 130: if (inspected_web_view_->GetWebDevToolsAgentImpl())
On 2009/05/27 07:02:36, Yury Semikhatsky wrote:
> ditto

ditto

http://codereview.chromium.org/113836/diff/1159/104
File webkit/glue/webdevtoolsagent_impl.cc (right):

http://codereview.chromium.org/113836/diff/1159/104#newcode106
Line 106: ic->setWindowVisible(false, false);
On 2009/05/27 07:02:36, Yury Semikhatsky wrote:
> add comment why this is needed

Done.

http://codereview.chromium.org/113836/diff/1159/104#newcode218
Line 218: int identifier) {
On 2009/05/27 07:02:36, Yury Semikhatsky wrote:
> it should be unsigned long, isn't it?

It used to be int in the net_agent. Rpc does not handle long longs...