Changeset 414

Timestamp:

04/12/07 19:29:25 (1 year ago)

Author:

dconrad

Message:

Apparently returning 0 for frame size with a format that has a variable number of frames per packet is still broken on Intel. XiphQT returns 1 on Intel, so we should as well.

Files:

• trunk/FFissionCodec/FFissionDecoder.cpp (modified) (2 diffs)

trunk/FFissionCodec/FFissionDecoder.cpp

@@ -418 +418 @@
-}
-
+// comment from XiphQT (variable frames per packet means FrameSize should be reported 0, 
+// but apparently needs 1 on Intel?):
+/* The following line has been changed according to Apple engineers' suggestion
+   I received via Steve Nicolai (in response to *my* bugreport, I think...).
+   (Why don't they just implement the VBR-VFPP properly? *sigh*) */
+#ifdef TARGET_CPU_X86
+#define SHOULD_BE_ZERO 1
+#else
+#define SHOULD_BE_ZERO 0
+#endif
-
-void FFissionVBRDecoder::GetProperty(AudioCodecPropertyID inPropertyID, UInt32& ioPropertyDataSize, void* outPropertyData)
@@ -431 +441 @@
-        switch (inPropertyID) {
-                case kAudioCodecPropertyPacketFrameSize:
-0
+SHOULD_BE_ZERO
-                        break;
-