content: Remove usage of NOTIFICATION_WEB_CONTENTS_DESTROYED from content

content: Remove usage of NOTIFICATION_WEB_CONTENTS_DESTROYED from content

This is a first CL in a long series of removal of things from chrome.
I prefer to start with a simple CL and work step by step on chrome removals.

BUG=170921
R=avi@chromium.org

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=201348

[jochen (gone - plz use gerrit), 2013-05-10 19:53:33]

https://codereview.chromium.org/14651029/diff/1/content/shell/shell.h
File content/shell/shell.h (right):

https://codereview.chromium.org/14651029/diff/1/content/shell/shell.h#newcode53
content/shell/shell.h:53: public WebContentsObserver,
Shell shouldn't be a WebContentsObserver. Can you use a similar destruction
observer patterns as you use in WCI?

[Avi (use Gerrit), 2013-05-10 20:26:44]

https://codereview.chromium.org/14651029/diff/1/content/shell/shell.h
File content/shell/shell.h (right):

https://codereview.chromium.org/14651029/diff/1/content/shell/shell.h#newcode53
content/shell/shell.h:53: public WebContentsObserver,
On 2013/05/10 19:53:33, jochen wrote:
> Shell shouldn't be a WebContentsObserver. Can you use a similar destruction
> observer patterns as you use in WCI?

What's your reasoning here?

[jochen (gone - plz use gerrit), 2013-05-10 20:31:50]

+jam

I didn't find the discussion after a quick search, maybe John remembers

[Paweł Hajdan Jr., 2013-05-13 20:15:30]

https://codereview.chromium.org/14651029/diff/1/content/shell/shell.h
File content/shell/shell.h (right):

https://codereview.chromium.org/14651029/diff/1/content/shell/shell.h#newcode53
content/shell/shell.h:53: public WebContentsObserver,
On 2013/05/10 19:53:33, jochen wrote:
> Shell shouldn't be a WebContentsObserver. Can you use a similar destruction
> observer patterns as you use in WCI?

Done.

[Avi (use Gerrit), 2013-05-15 19:22:19]

https://codereview.chromium.org/14651029/diff/1/content/shell/shell.h
File content/shell/shell.h (right):

https://codereview.chromium.org/14651029/diff/1/content/shell/shell.h#newcode53
content/shell/shell.h:53: public WebContentsObserver,
Jochen: I will again ask why. Shell needs to observe a single WC, so why this
game of indirection and a proxy object?

https://codereview.chromium.org/14651029/diff/14001/content/browser/web_conte...
File content/browser/web_contents/web_contents_impl.h (right):

https://codereview.chromium.org/14651029/diff/14001/content/browser/web_conte...
content/browser/web_contents/web_contents_impl.h:523: // Creats and adds to the
map a destruction observer watching |web_contents|.
typo

[jochen (gone - plz use gerrit), 2013-05-21 12:09:44]

https://codereview.chromium.org/14651029/diff/1/content/shell/shell.h
File content/shell/shell.h (right):

https://codereview.chromium.org/14651029/diff/1/content/shell/shell.h#newcode53
content/shell/shell.h:53: public WebContentsObserver,
On 2013/05/15 19:22:19, Avi wrote:
> Jochen: I will again ask why. Shell needs to observe a single WC, so why this
> game of indirection and a proxy object?

As I said, I don't remember the discussion we had about this exactly. In
general, Shell is a pretty lame object - a concept like this doesn't exist at
all in chrome/. It's a bit like a bastardized TabContents before we deleted it.

I think we should delete Shell all together (just like we did with TabContents).
Making it implement both WebContentsDelegate and WebContentsObserver will make
it just more difficult.

[Paweł Hajdan Jr., 2013-05-21 18:59:15]

PTAL

Avi, what do you think after seeing Jochen's response?

https://codereview.chromium.org/14651029/diff/14001/content/browser/web_conte...
File content/browser/web_contents/web_contents_impl.h (right):

https://codereview.chromium.org/14651029/diff/14001/content/browser/web_conte...
content/browser/web_contents/web_contents_impl.h:523: // Creats and adds to the
map a destruction observer watching |web_contents|.
On 2013/05/15 19:22:20, Avi wrote:
> typo

Done.

[Avi (use Gerrit), 2013-05-21 19:09:53]

LGTM

I'm not thrilled with that decision, though if we are going to kill Shell it
makes sense. I didn't realize we were going to, though.

Who's doing that?

[jochen (gone - plz use gerrit), 2013-05-21 19:24:23]

On 2013/05/21 19:09:53, Avi wrote:
> LGTM
> 
> I'm not thrilled with that decision, though if we are going to kill Shell it
> makes sense. I didn't realize we were going to, though.
> 
> Who's doing that?

nobody signed up for the task yet. :-/

[Paweł Hajdan Jr., 2013-05-21 20:21:16]

Committed patchset #4 manually as r201348 (presubmit successful).