1. 14 Feb, 2004 1 commit
    • gbazin's avatar
      · d09106d2
      gbazin authored
      * ALL: small portability fixes.
      d09106d2
  2. 13 Feb, 2004 11 commits
  3. 12 Feb, 2004 4 commits
    • gbazin's avatar
      · d2e4386e
      gbazin authored
      * modules/stream_out/transcode.c: fixed color inversion when transcoding from raw RV24.
      d2e4386e
    • Laurent Aimar's avatar
      * all: use the data_packet_count from the asf header when the server · ddffbfbc
      Laurent Aimar authored
      gives an invalid one. (I hope it doesn't break anything)
      ddffbfbc
    • Laurent Aimar's avatar
      * v4l: patch from Carsten Gottbeh�t (gottbehuet at active-elements dot de) · dc0a3896
      Laurent Aimar authored
       -> Untested.
      
       Original message:
          Hi there,
      
          I have done another small modification. If you remove a hot-pluggable video
          device, like an usb web cam, during capturing, vlc floods the driver with
          ioctls. This leads to Kernel Ooops in some drivers.
      
          The modified v4l.c module does the following now:
          "GrabVideo" returns "VLC_ETIMEOUT" if the question "Did we wait long enough"
          can be answered with yes. If there is a real error in "GrabCapture" or
          "GrabMJPEG", the code "VLC_EGENERIC" is returned. So the calling function
          can distinguish between these two situations. "VLC_EGENERIC" was always
          returned before.
          The function "Read" uses this information to leave the while-loop in case of
          an error, and it returns a "-1" to the calling function. "Read" calls
          "msleep", as it was before, in case of a VLC_ETIMEOUT.
      
          The brightness-control stuff described below is still in the code. Please
          feel free to use, modify or ignore this file. A short feedback would be
          nice.
      
          Bye,
      dc0a3896
    • Eric Petit's avatar
  4. 11 Feb, 2004 6 commits
  5. 10 Feb, 2004 10 commits
  6. 09 Feb, 2004 8 commits