edburns%acm.org 0176a48d48 This checkin provides a sketch of how I plan to re-implement the ...

NewWindowListener mechanism.  Next step will be to continue to try to
get it working.

A test/automated/src/classes/org/mozilla/webclient/WindowCreatorTest.java
A test/automated/src/test/WindowCreatorTest0.html
A test/automated/src/test/WindowCreatorTest1.html

- new test, currently fails.

M build-tests.xml

- add new test, non-running

M classes_spec/org/mozilla/webclient/impl/wrapper_native/EventRegistrationImpl.java

- prep for NewWindowListener

M src_moz/WindowCreator.cpp

- sketch implementation plan.  In WindowCreator.CreateChromeWindow2:

+        /*
+          Block this thread.
+
+          Call back into java and ask the user to create a top level
+          window and hand it, or an added child of it, to us.  Call this
+          thing the userWindow.
+
+          Create a new BrowserControl, get its BrowserControlCanvas and
+          make it be a child of the userWindow.
+
+          Set the userWindow and the BrowserControlCanvas to visible ==
+          true.  This is necessary to get the cause the underlying
+          mozilla window to be created.
+
+          java returns the C++ nativeBrowserControl to us.  Cast it to a
+          native NativeBrowserControl C++ object instance.  If the
+          nsIURI is non-null, cause the new window to navigate to that
+          URI.  Return the NativeBrowserControl's EmbedWindow instance,
+          which is an impl of nsIWebBrowserChrome.
+
+          I'm not sure if it's safe to do all this on the same thread on
+          which mozilla calls us.  I hope so.
+        */


git-svn-id: svn://10.0.0.236/trunk@167522 18797224-902f-48f8-a5cc-f745e15eee43

2005-01-11 07:02:12 +00:00

3.9 KiB

Raw Blame History

View Raw