Recherche avancée

Sur d’autres sites (7277)

• Using ffmpeg with C++ : Undefined symbols for architecture x86_64

  31 mars 2017, par JavaRunner

  I’m dealing with one example of FFMPEG called is "muxing.c".

  So I tried to compile that file with two Makefiles. The first one is for C-language :

  FFMPEG_LIBS=    libavdevice                        \
  
                  libavformat                        \
  
                  libavfilter                        \
  
                  libavcodec                         \
  
                  libswresample                      \
  
                  libswscale                         \
  
                  libavutil                          \
  
  
  
  CFLAGS := -Wall -g $(shell pkg-config --cflags $(FFMPEG_LIBS)) $(CFLAGS)
  
  LDLIBS := $(shell pkg-config --libs $(FFMPEG_LIBS)) -lm $(LDLIBS)
  
  
  
  SOURCES=        muxing                             \
  
  
  
  OBJS=$(addsuffix .o,$(SOURCES))
  
  
  
  all: $(OBJS) $(SOURCES)

  That makefile compiles well. But if I trying to move this code to C++ : I rename muxing.c to muxing.cpp and use this Makefile :

  CC=g++
  
  
  
  FFMPEG_LIBS=    libavdevice                        \
  
                  libavformat                        \
  
                  libavfilter                        \
  
                  libavcodec                         \
  
                  libswresample                      \
  
                  libswscale                         \
  
                  libavutil                          \
  
  
  
  CPPFLAGS := -Wall -g $(shell pkg-config --cflags $(FFMPEG_LIBS)) $(CPPFLAGS)
  
  LDLIBS := $(shell pkg-config --libs $(FFMPEG_LIBS)) -lm $(LDLIBS)
  
  
  
  SOURCES=        muxing                             \
  
  
  
  OBJS=$(addsuffix .o,$(SOURCES))
  
  
  
  all: $(OBJS) $(SOURCES)

  I get a bunch of error like this :

  $ g++   muxing.o  -L/usr/local/Cellar/ffmpeg/3.2.4/lib -lavdevice -lavformat -lavfilter -lavcodec -lswresample -lswscale -lavutil  -lm -o muxing
  
  Undefined symbols for architecture x86_64:
  
    "av_dict_set(AVDictionary**, char const*, char const*, int)", referenced from:
  
        _main in muxing.o
  
    "av_strerror(int, char*, unsigned long)", referenced from:
  
        av_make_error_string(char*, unsigned long, int) in muxing.o
  
    "avio_closep(AVIOContext**)", referenced from:
  
        _main in muxing.o
  
    "swr_convert(SwrContext*, unsigned char**, int, unsigned char const**, int)", referenced from:
  
        write_audio_frame(AVFormatContext*, OutputStream*) in muxing.o
  
    "av_dict_copy(AVDictionary**, AVDictionary const*, int)", referenced from:
  
        open_video(AVFormatContext*, AVCodec*, OutputStream*, AVDictionary*) in muxing.o
  
        open_audio(AVFormatContext*, AVCodec*, OutputStream*, AVDictionary*) in muxing.o
  
    "av_dict_free(AVDictionary**)", referenced from:
  
        open_video(AVFormatContext*, AVCodec*, OutputStream*, AVDictionary*) in muxing.o
  
        open_audio(AVFormatContext*, AVCodec*, OutputStream*, AVDictionary*) in muxing.o
  
    "av_rescale_q(long long, AVRational, AVRational)", referenced from:
  
        write_audio_frame(AVFormatContext*, OutputStream*) in muxing.o
  
    "av_compare_ts(long long, AVRational, long long, AVRational)", referenced from:
  
        _main in muxing.o
  
        get_video_frame(OutputStream*) in muxing.o
  
        get_audio_frame(OutputStream*) in muxing.o

  Is that mean those libs are not ready to work with x64 ? I strongly need to mix that code (muxing.c) with my C++ code because I have to render a video on the fly using ImageMagick (my code was written with C++) for the frames.

  How can I mix C++ and C in this case ?

  My env is : OS X 10.12.3

• avcodec/pixlet : fix architecture-dependent code and values

  7 mars 2017, par Vittorio Giovara

  avcodec/pixlet : fix architecture-dependent code and values
  The constants used in the decoder used floating point precision,
  
  and this caused different values to be generated on different
  
  architectures.
  So, eradicate floating point numbers and use fixed point (32.32)
  
  arithmetics everywhere, replacing constants with precomputed integer
  
  values.
  Signed-off-by : Vittorio Giovara <vittorio.giovara at gmail.com>
  
  Signed-off-by : Paul B Mahol <onemda@gmail.com>
  

  • [DH] libavcodec/pixlet.c

• pixlet : Fix architecture-dependent code and values

  2 mars 2017, par Vittorio Giovara

  pixlet : Fix architecture-dependent code and values
  The constants used in the decoder used floating point precision,
  
  and this caused different values to be generated on different
  
  architectures. Additionally on big endian machines, the fate test
  
  would output bytes in native order, which is different from the one
  
  hardcoded in the test.
  So, eradicate floating point numbers and use fixed point (32.32)
  
  arithmetics everywhere, replacing constants with precomputed integer
  
  values, and force the pixel format output to be the same in the fate
  
  test.
  Signed-off-by : Vittorio Giovara <vittorio.giovara@gmail.com>
  

  • [DBH] libavcodec/pixlet.c
  • [DBH] tests/fate/video.mak
  • [DBH] tests/ref/fate/pixlet