Some additional work needs to do in WebKit2 for this feature.
Created attachment 109861 [details] patch 1
Comment on attachment 109861 [details] patch 1 View in context: https://bugs.webkit.org/attachment.cgi?id=109861&action=review > Source/WebKit2/WebProcess/WebPage/WebPage.cpp:1707 > +#if ENABLE(PASSWORD_ECHO) > + settings->setPasswordEchoEnabled(true); > + settings->setPasswordEchoDurationInSeconds(1); > +#endif Password echo turned on unconditionally if it’s compiled at all? That just doesn’t seem right. If that was the policy then I’d expect to see that policy in WebCore.
(In reply to comment #2) > (From update of attachment 109861 [details]) > View in context: https://bugs.webkit.org/attachment.cgi?id=109861&action=review > > > Source/WebKit2/WebProcess/WebPage/WebPage.cpp:1707 > > +#if ENABLE(PASSWORD_ECHO) > > + settings->setPasswordEchoEnabled(true); > > + settings->setPasswordEchoDurationInSeconds(1); > > +#endif > > Password echo turned on unconditionally if it’s compiled at all? That just doesn’t seem right. If that was the policy then I’d expect to see that policy in WebCore. By default, PASSWORD_ECHO is not enabled. If the feature is chosen to turn on, a build time flag is required. An alternative way is to enable it in runtime. Alexey and I have discussed on whether the application(such as MiniBrowser) should take care of this (during runtime) but we chose to control this in WebKit. WebCore seems a place to provide this option in the settings.
I remember discussing that every application on a platform should not be required to do something to follow default platform behavior for password echo. Details of proposed solution are a bit vague - this was a while ago.
(In reply to comment #4) > I remember discussing that every application on a platform should not be required to do something to follow default platform behavior for password echo. Details of proposed solution are a bit vague - this was a while ago. It seems strange to have something defaulting off at the WebCore level then turned on unconditionally at the WebKit level.
(In reply to comment #5) > (In reply to comment #4) > > I remember discussing that every application on a platform should not be required to do something to follow default platform behavior for password echo. Details of proposed solution are a bit vague - this was a while ago. > > It seems strange to have something defaulting off at the WebCore level then turned on unconditionally at the WebKit level. The "condition" is the build flag. In fact, symbian has already defined ENABLE_PASSWORD_ECHO=1 in WebCore/config.h. And other platforms are disabled by default. In anther way, we can set different default value using #if PLATFORM() flag in WebCore. Maybe we should do this. Btw, just got the sad news.
We don't need this any more. The compile time is removed in bug 69647 so we don't need any special treatment in WK2 code. The unskipped tests are included there, too.