x86/takdsp: use arithmetic shift instructions

[ffmpeg] / doc / APIchanges

diff --git a/doc/APIchanges b/doc/APIchanges
index 7cde366b0893fe407d73c690479e79b13ce73931..03ffda8c3705c7450863b390b3fd54dea8344600 100644 (file)
--- a/doc/APIchanges
+++ b/doc/APIchanges
@@ -15,6 +15,17 @@ libavutil:     2015-08-28
 
 API changes, most recent first:
 
+2015-09-29 - xxxxxxx - lavc 57.3.100 / lavc 57.2.0 - avcodec.h
+  Change type of AVPacket.duration from int to int64_t.
+
+2015-09-17 - xxxxxxx - lavc 57.3.100 / lavc 57.2.0 - d3d11va.h
+  Add av_d3d11va_alloc_context(). This function must from now on be used for
+  allocating AVD3D11VAContext.
+
+2015-09-15 - lavf 57.2.100 - avformat.h
+  probesize and max_analyze_duration switched to 64bit, both
+  are only accessible through AVOptions
+
 2015-09-15 - lavf 57.1.100 - avformat.h
   bit_rate was changed to 64bit, make sure you update any
   printf() or other type sensitive code