Remote assistance on Chrome OS Part IV - It2MeHost

remoting/host/chromeos/aura_desktop_capturer.cc

Index: remoting/host/chromeos/aura_desktop_capturer.cc
diff --git a/remoting/host/chromeos/aura_desktop_capturer.cc b/remoting/host/chromeos/aura_desktop_capturer.cc
index caf6837430d03df22dfc29824a13bc3bbc89ea1d..f6d85b4c59174e10feb67c3b5bcae562a2d60438 100644
--- a/remoting/host/chromeos/aura_desktop_capturer.cc
+++ b/remoting/host/chromeos/aura_desktop_capturer.cc
@@ -40,8 +40,8 @@ SkiaBitmapDesktopFrame* SkiaBitmapDesktopFrame::Create(
     scoped_ptr<SkBitmap> bitmap) {
 
   webrtc::DesktopSize size(bitmap->width(), bitmap->height());
-  DCHECK_EQ(kRGBA_8888_SkColorType, bitmap->info().colorType())

[rmsousa, 2014/10/14 00:18:25]

Out of curiosity: was this broken before? Are we the first feature to exercise
this code?

[kelvinp, 2014/10/15 23:03:09]

Yes, I created this file two weeks ago but it is not being exercised other than
unittest.  I create the DCHECK based on the comments on webrtc::DeskopFrame,
which is apparently incorrect.

-      << "DesktopFrame objects always hold RGBA data.";
+  DCHECK_EQ(kBGRA_8888_SkColorType, bitmap->info().colorType())
+      << "DesktopFrame objects always hold BGRA data.";

[Jamie, 2014/10/14 01:18:41]

There seem to be a couple of changes that flip the endianness, but there's
nothing in the CL description. If it's an unrelated bugfix, consider moving it
to another CL. If not, please mention it in the description.

[kelvinp, 2014/10/15 23:03:09]

This is required otherwise the DCHECK will fail after the remoting session is
connected. I will update the CL description.

 
   uint8_t* bitmap_data = reinterpret_cast<uint8_t*>(bitmap->getPixels());