1. 09 Dec, 2017 1 commit
  2. 08 Dec, 2017 5 commits
    • Tim-Philipp Müller's avatar
      alsamidisrc: unlock create function on shutdown · 0ef44cef
      Tim-Philipp Müller authored
      Fixes occasional hang in generic/states unit test or
      
      GST_CHECKS=test_state_changes_up_and_down_seq make forever
      0ef44cef
    • Edward Hervey's avatar
      oggdemux: Check encoder name is valid · a28880b4
      Edward Hervey authored
      Encoder names should be valid utf-8, if not just ignore them
      a28880b4
    • Edward Hervey's avatar
      typefind: Avoid overflow calculation (image/quicktime) · aa10e822
      Edward Hervey authored
      The qt typefinder uses guint64 values for offset and size calculation
      but the typefinder system only supports gint64 values.
      
      Make sure we don't end up using potentially overflowing values.
      aa10e822
    • Edward Hervey's avatar
      typefind: Avoid overflow calculation · 1df9b05f
      Edward Hervey authored
      The qt typefinder uses guint64 values for offset and size calculation
      but the typefinder system only supports gint64 values.
      
      Make sure we don't end up using potentially overflowing values.
      1df9b05f
    • Akinobu Mita's avatar
      ximagesink, xvimagesink: fix incorrect type conversion of pointer position · 6e770e0e
      Akinobu Mita authored
      I'm currently playing with modified ximagesink that does XGrabPointer()
      in order to receive the mouse events occurred outside of the window and
      send them to the navigation interface.
      
      The pointer positions usually have positive coordinates, but it could
      be negative with that change.
      
      When the ximagesink handles XEvent that contains a negative pointer
      coordinate, it incorrectly generates the GstEvent that contains an
      extremely large positive pointer coordinate.
      
      This is because the negative pointer position in XEvent is incorrectly
      converted from signed to unsigned and passed as an argument to
      gst_navigation_send_mouse_event() which causes implicit conversion from
      integer to double.  So the pointer position in the received XEvent and
      generated GstEvent are completely different.
      
      This potential problem does not seem to be a real problem with unmodified
      ximagesink but there is no reason to leave it as is.  This also fixes
      xvimagesink that has the same potential problem.
      
      https://bugzilla.gnome.org/show_bug.cgi?id=791140
      6e770e0e
  3. 07 Dec, 2017 4 commits
  4. 06 Dec, 2017 3 commits
  5. 05 Dec, 2017 1 commit
  6. 01 Dec, 2017 2 commits
  7. 30 Nov, 2017 9 commits
  8. 29 Nov, 2017 1 commit
  9. 27 Nov, 2017 1 commit
  10. 26 Nov, 2017 3 commits
  11. 25 Nov, 2017 1 commit
  12. 21 Nov, 2017 1 commit
    • Edward Hervey's avatar
      videotestsrc: Avoid overflow calculation · a4788110
      Edward Hervey authored
      n_frames could end up being quite big (potentially up to G_MAXINT64). Which
      would result in overflowing 64bits when multiplying it by GST_SECOND.
      
      Instead move GST_SECOND to the num argument
      a4788110
  13. 17 Nov, 2017 1 commit
  14. 16 Nov, 2017 2 commits
    • Edward Hervey's avatar
      decodebin2: Don't spawn threads on shutdown · 38c835ec
      Edward Hervey authored
      If we are shutting down, don't spawn a cleanup thread to cleanup old
      groups and instead queue them to be cleaned up in the state change
      thread.
      
      This avoids (hopefully for good) having a race between the state change
      thread and other threads trying to deactivate elements/pads.
      38c835ec
    • Edward Hervey's avatar
      decodebin2: Avoid deactivation races · e72aa501
      Edward Hervey authored
      Deactivating pads from two threads isn't 100% MT-safe. There is a
      slim chance that the GstPadActivateFunc might be called twice with
      the same values (in this case from the cleanup thread *and* from
      the GstElement change_state function when going from PAUSED to READY).
      
      In order to avoid that, call any existing cleanup function *before*
      calling the parent change_state implementation on downwards state
      changes.
      e72aa501
  15. 15 Nov, 2017 5 commits