[Gnash-commit] gnash ChangeLog libmedia/FLVParser.cpp

[Sandro Santilli]

CVSROOT:        /sources/gnash
Module name:    gnash
Changes by:     Sandro Santilli <strk>  08/06/04 16:49:23

Modified files:
        .              : ChangeLog 
        libmedia       : FLVParser.cpp 

Log message:
                * libmedia/FLVParser.cpp (getBufferLength): don't rely
                  on cursor to tell how much time the buffer contains;
                  add notes about need to sort out a rationale for
                  what to return when both audio/video buffers exist..

CVSWeb URLs:
http://cvs.savannah.gnu.org/viewcvs/gnash/ChangeLog?cvsroot=gnash&r1=1.6815&r2=1.6816
http://cvs.savannah.gnu.org/viewcvs/gnash/libmedia/FLVParser.cpp?cvsroot=gnash&r1=1.10&r2=1.11

Patches:
Index: ChangeLog
===================================================================
RCS file: /sources/gnash/gnash/ChangeLog,v
retrieving revision 1.6815
retrieving revision 1.6816
diff -u -b -r1.6815 -r1.6816
--- ChangeLog   4 Jun 2008 16:47:03 -0000       1.6815
+++ ChangeLog   4 Jun 2008 16:49:22 -0000       1.6816
@@ -1,5 +1,12 @@
 2008-06-04 Sandro Santilli <address@hidden>
 
+       * libmedia/FLVParser.cpp (getBufferLength): don't rely
+         on cursor to tell how much time the buffer contains;
+         add notes about need to sort out a rationale for
+         what to return when both audio/video buffers exist..
+
+2008-06-04 Sandro Santilli <address@hidden>
+
        * testsuite/misc-ming.all/NetStream-SquareTest.c: don't let
          the buffer loaded bar become > 100%, but keep it at alpha
          50 while not saturated, and bring to 100 when saturated.

Index: libmedia/FLVParser.cpp
===================================================================
RCS file: /sources/gnash/gnash/libmedia/FLVParser.cpp,v
retrieving revision 1.10
retrieving revision 1.11
diff -u -b -r1.10 -r1.11
--- libmedia/FLVParser.cpp      4 Jun 2008 11:48:50 -0000       1.10
+++ libmedia/FLVParser.cpp      4 Jun 2008 16:49:23 -0000       1.11
@@ -126,18 +126,22 @@
 boost::uint32_t
 FLVParser::getBufferLength()
 {
-       if (_video) {
-               size_t size = _videoFrames.size();
-               if (size > 1 && size > _nextVideoFrame) {
-                       return _videoFrames.back()->timestamp; //  - 
_videoFrames[_nextVideoFrame]->timestamp;
-               }
-       }
-       if (_audio) {
-               size_t size = _audioFrames.size();
-               if (size > 1 && size > _nextAudioFrame) {
-                       return _audioFrames.back()->timestamp; //  - 
_audioFrames[_nextAudioFrame]->timestamp;
+       // TODO: figure wheter and why we should privilege
+       //       video frames over audio frames when both
+       //       are available
+       //       I belive the corrent behaviour here would
+       //       be using the smallest max-timestamp..
+
+       if (_video && !_videoFrames.empty())
+       {
+               return _videoFrames.back()->timestamp; 
                }
+
+       if (_audio && ! _audioFrames.empty())
+       {
+               return _audioFrames.back()->timestamp; 
        }
+
        return 0;
 }
 boost::uint16_t