More refactoring of PluginProcessHost

Have ChildProcessInfo contain a list of all running child processes (i.e. instead of Service and other child process service maintain it).  In a future change I'll start moving some of the code from PluginProcessHost to ChildProcessInfo.

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=9804

[Matt Perry, 2009-02-13 21:04:38]

lgtm

http://codereview.chromium.org/24017/diff/1015/2030
File chrome/browser/browser.cc (right):

http://codereview.chromium.org/24017/diff/1015/2030#newcode135
Line 135: // thread because that thread manages the plugin process collection.
s/plugin/child

http://codereview.chromium.org/24017/diff/1015/2026
File chrome/common/child_process_info.cc (right):

http://codereview.chromium.org/24017/diff/1015/2026#newcode16
Line 16: std::list<ChildProcessInfo*>* g_children;
use base::Singleton?

http://codereview.chromium.org/24017/diff/1015/2025
File chrome/common/child_process_info.h (right):

http://codereview.chromium.org/24017/diff/1015/2025#newcode106
Line 106: ChildProcessInfo() { }
Since you defined non-default constructors, C++ won't create a default one for
you, so I don't think this is necessary.

[jam, 2009-02-13 22:58:54]

(updated the subject)

http://codereview.chromium.org/24017/diff/1015/2030
File chrome/browser/browser.cc (right):

http://codereview.chromium.org/24017/diff/1015/2030#newcode135
Line 135: // thread because that thread manages the plugin process collection.
On 2009/02/13 21:04:38, Matt Perry wrote:
> s/plugin/child

Done.

http://codereview.chromium.org/24017/diff/1015/2026
File chrome/common/child_process_info.cc (right):

http://codereview.chromium.org/24017/diff/1015/2026#newcode16
Line 16: std::list<ChildProcessInfo*>* g_children;
On 2009/02/13 21:04:38, Matt Perry wrote:
> use base::Singleton?

neat, I didn't know I could do that.

http://codereview.chromium.org/24017/diff/1015/2025
File chrome/common/child_process_info.h (right):

http://codereview.chromium.org/24017/diff/1015/2025#newcode106
Line 106: ChildProcessInfo() { }
On 2009/02/13 21:04:38, Matt Perry wrote:
> Since you defined non-default constructors, C++ won't create a default one for
> you, so I don't think this is necessary.

good point, done

[Matt Perry, 2009-02-13 22:59:58]

looks good (fwiw, I'm not an expert on this code, but the change seems simple
enough)