Fix positioning of Browser Plugins

content/renderer/browser_plugin/browser_plugin.cc

Index: content/renderer/browser_plugin/browser_plugin.cc
diff --git a/content/renderer/browser_plugin/browser_plugin.cc b/content/renderer/browser_plugin/browser_plugin.cc
index 377e64749d7c044c5273e9a6960696674e3543ec..006b7a3d841f1663203af5043903076ffc151655 100644
--- a/content/renderer/browser_plugin/browser_plugin.cc
+++ b/content/renderer/browser_plugin/browser_plugin.cc
@@ -382,7 +382,7 @@ void BrowserPlugin::updateGeometry(const WebRect& plugin_rect_in_viewport,
   if (!attached())
     return;
 
-  if (old_view_rect.size() == view_rect_.size()) {

[Stephen Chennney, 2016/05/10 18:40:23]

This very odd condition is a remnant of
https://chromium.googlesource.com/chromium/src/+/32deec6509deb789265c50ba2943...,
much moved since then. It was revealed by a recent change to prevent double
calls to updateGeometry, which are wasteful.

[lazyboy, 2016/05/10 20:51:58]

Which change was that?
which are wasteful.

The intention here IIRC was when plugin was not resized but its position (x,y)
changed, we wanted to send the rect to browser/ process in this case.
What you have still accomplishes the same and but a bit more.
In case where the size have changed (but x, y didn't), we would send an IPC
through GuestViewContainer in line 380 above:
https://code.google.com/p/chromium/codesearch#chromium/src/extensions/browser...
and then an IPC here, which wasn't necessary.

Maybe instead you need
if (old_view_rect != view_rect_ && old.size() == view_rect_.size())
Would that accomplish our goal here?

+  if (old_view_rect != view_rect_) {
     // Let the browser know about the updated view rect.
     BrowserPluginManager::Get()->Send(new BrowserPluginHostMsg_UpdateGeometry(
         browser_plugin_instance_id_, view_rect_));