gtk test shell

Basic GTK test shell: links in the SConscript and brings up a window with nonfunctioning UI.

[Evan Martin, 2008-09-19 02:11:06]

Here's a sample patch.  From here you could imagine bringing in bits of glue
(I've done a bit already).

[tony, 2008-09-19 06:01:46]

LGTM

http://codereview.chromium.org/2983/diff/1/3
File webkit/tools/test_shell/SConscript (right):

http://codereview.chromium.org/2983/diff/1/3#newcode92
Line 92: if env['PLATFORM'] == 'win32':
Maybe add "TODO(port)"

http://codereview.chromium.org/2983/diff/1/3#newcode117
Line 117: if env['PLATFORM'] == 'win32':
TODO(port)

http://codereview.chromium.org/2983/diff/1/3#newcode147
Line 147: if env['PLATFORM'] == 'win32':
TODO(port)

http://codereview.chromium.org/2983/diff/1/5
File webkit/tools/test_shell/test_shell_gtk.cc (right):

http://codereview.chromium.org/2983/diff/1/5#newcode38
Line 38: m_mainWnd = gtk_window_new(GTK_WINDOW_TOPLEVEL);
Can we update WindowHandle in base/gfx/native_widget_types.h to be a GtkWidget?

http://codereview.chromium.org/2983/diff/1/6
File webkit/tools/test_shell/test_shell_main_gtk.cc (right):

http://codereview.chromium.org/2983/diff/1/6#newcode29
Line 29: //MessageLoopForUI main_message_loop;
Can you add a comment as to why this is commented out?

http://codereview.chromium.org/2983/diff/1/6#newcode42
Line 42: // Default to a homepage if we're interactive.
Maybe this comment should be merged with the comment above it.

http://codereview.chromium.org/2983/diff/1/6#newcode57
Line 57: printf("Would load URI: %s.\n", WideToUTF8(uri).c_str());
Can we use LOG(INFO) here?

[Evan Martin, 2008-09-19 17:01:52]

Please look again.

[tony, 2008-09-19 17:16:09]

LGTM.  You may need to resolve conflicts because I change Port to port.

[Evan Martin, 2008-09-19 17:23:23]

Yeah, I just did.  :\

And I had to delete cairo.h for this to work, which I guess means I should sit
on it until the merge branch lands?

[Evan Martin, 2008-10-02 02:06:59]

Baby steps.

This may conflict with a change from dsh to bring in GTK.