Implement --enable-web-sockets flag.

Implement --enable-web-sockets flag.

BUG=12497
TEST=none

[darin (slow to review), 2009-10-05 19:42:00]

http://codereview.chromium.org/255075/diff/1/7
File webkit/api/public/WebKit.h (right):

http://codereview.chromium.org/255075/diff/1/7#newcode113
Line 113: WEBKIT_API void enableWebSockets();
nit: please move this up to just below the similar methods for databases.  Can
we just use the term enableSockets / socketsEnabled?  We tend to leave off the
Web prefix when just referring to these things.  For example, it is technically
WebDatabase but we just call it database at this API layer.

http://codereview.chromium.org/255075/diff/1/12
File webkit/glue/webpreferences.cc (right):

http://codereview.chromium.org/255075/diff/1/12#newcode55
Line 55: settings->setExperimentalWebSocketsEnabled(WebKit::webSocketsEnabled()
|| web_sockets_enabled);
please stick to 80 char line length

[ukai, 2009-10-06 04:12:12]

Thanks for review.

http://codereview.chromium.org/255075/diff/1/7
File webkit/api/public/WebKit.h (right):

http://codereview.chromium.org/255075/diff/1/7#newcode113
Line 113: WEBKIT_API void enableWebSockets();
On 2009/10/05 19:42:00, darin wrote:
> nit: please move this up to just below the similar methods for databases.  Can
> we just use the term enableSockets / socketsEnabled?  We tend to leave off the
> Web prefix when just referring to these things.  For example, it is
technically
> WebDatabase but we just call it database at this API layer.

Done.
Do you think I should fix all WebScokets methods/fields in this CL, or just fix
at WebKit API interface?

http://codereview.chromium.org/255075/diff/1/12
File webkit/glue/webpreferences.cc (right):

http://codereview.chromium.org/255075/diff/1/12#newcode55
Line 55: settings->setExperimentalWebSocketsEnabled(WebKit::webSocketsEnabled()
|| web_sockets_enabled);
On 2009/10/05 19:42:00, darin wrote:
> please stick to 80 char line length

Done.

[Yuzo, 2009-10-06 04:16:51]

http://codereview.chromium.org/255075/diff/1/10
File webkit/api/src/WebSettingsImpl.cpp (right):

http://codereview.chromium.org/255075/diff/1/10#newcode254
Line 254: #if ENABLE(WEB_SOCKETS)
This means calling
setExperimentalWebSocketsEnabled(false)
has no effect unless WEB_SOCKET is defined.

Is this what you want?

[ukai, 2009-10-06 04:22:12]

http://codereview.chromium.org/255075/diff/1/10
File webkit/api/src/WebSettingsImpl.cpp (right):

http://codereview.chromium.org/255075/diff/1/10#newcode254
Line 254: #if ENABLE(WEB_SOCKETS)
On 2009/10/06 04:16:53, Yuzo wrote:
> This means calling
> setExperimentalWebSocketsEnabled(false)
> has no effect unless WEB_SOCKET is defined.
> 
> Is this what you want?

Yes.  If WEB_SOCKETS is not enabled, this method is NOP.

[Yuzo, 2009-10-06 04:26:06]

LGTM

[darin (slow to review), 2009-10-06 04:39:12]

http://codereview.chromium.org/255075/diff/1/7
File webkit/api/public/WebKit.h (right):

http://codereview.chromium.org/255075/diff/1/7#newcode113
Line 113: WEBKIT_API void enableWebSockets();
On 2009/10/06 04:12:14, ukai wrote:
> On 2009/10/05 19:42:00, darin wrote:
> > nit: please move this up to just below the similar methods for databases. 
Can
> > we just use the term enableSockets / socketsEnabled?  We tend to leave off
the
> > Web prefix when just referring to these things.  For example, it is
> technically
> > WebDatabase but we just call it database at this API layer.
> 
> Done.
> Do you think I should fix all WebScokets methods/fields in this CL, or just
fix
> at WebKit API interface?

I think it's enough to just worry about the WebKit API.  In Chrome land, it is
probably good to refer them as WebSockets.

Btw, if the enabling/disabling of WebSockets corresponds to methods on
WebCore::Settings, then shouldn't we just have methods for enabling/disabling
WebSockets on WebSettings?

[ukai, 2009-10-06 04:57:03]

On 2009/10/06 04:39:12, darin wrote:
> http://codereview.chromium.org/255075/diff/1/7
> File webkit/api/public/WebKit.h (right):
> 
> http://codereview.chromium.org/255075/diff/1/7#newcode113
> Line 113: WEBKIT_API void enableWebSockets();
> On 2009/10/06 04:12:14, ukai wrote:
> > On 2009/10/05 19:42:00, darin wrote:
> > > nit: please move this up to just below the similar methods for databases. 
> Can
> > > we just use the term enableSockets / socketsEnabled?  We tend to leave off
> the
> > > Web prefix when just referring to these things.  For example, it is
> > technically
> > > WebDatabase but we just call it database at this API layer.
> > 
> > Done.
> > Do you think I should fix all WebScokets methods/fields in this CL, or just
> fix
> > at WebKit API interface?
> 
> I think it's enough to just worry about the WebKit API.  In Chrome land, it is
> probably good to refer them as WebSockets.
> 
> Btw, if the enabling/disabling of WebSockets corresponds to methods on
> WebCore::Settings, then shouldn't we just have methods for enabling/disabling
> WebSockets on WebSettings?

Hmm, yes, just having methods on WebSettings is enough.
Removed WebKIT API changes.

[darin (slow to review), 2009-10-06 05:02:40]

LGTM!