1. 21 Dec, 2017 1 commit
  2. 18 Dec, 2017 1 commit
    • Steve Lhomme's avatar
      video_output: reinit the vout if the crop changes · 44b083a0
      Steve Lhomme authored
      If you play a file that has the same decoder size but different display size
      consecutively, the decoder picture pool needs to be updated so that the
      i_visible_lines of each picture plane matches the new picture format.
      
      The easiest way is to reinit the vout. Another way would be to update the
      pictures in each picture pool dynamically. It cannot just be handled on the
      vout side as some decoders and filters rely on i_visible/offset to copy
      only the required data (for example lavc_CopyPicture).
      
      Fixes #19268
      Signed-off-by: Thomas Guillem's avatarThomas Guillem <thomas@gllm.fr>
      (cherry picked from commit 17efaa77)
      Signed-off-by: Thomas Guillem's avatarThomas Guillem <thomas@gllm.fr>
      44b083a0
  3. 27 Nov, 2017 6 commits
  4. 15 Nov, 2017 1 commit
  5. 07 Nov, 2017 1 commit
  6. 06 Nov, 2017 1 commit
  7. 25 Oct, 2017 1 commit
  8. 21 Oct, 2017 1 commit
  9. 12 Oct, 2017 2 commits
  10. 29 Sep, 2017 1 commit
    • Thomas Guillem's avatar
      vout: also flush vd render filters · d94db1b1
      Thomas Guillem authored
      This fixes a video freeze after a seek with VDPAU HW rendering. Indeed, the
      VLC_CODEC_VDPAU_OUTPUT video converter need to be flushed to clear pictures
      history.
      
      Fixes #18423
      d94db1b1
  11. 28 Sep, 2017 1 commit
  12. 27 Sep, 2017 1 commit
  13. 24 Sep, 2017 3 commits
  14. 04 Sep, 2017 1 commit
  15. 02 Sep, 2017 10 commits
  16. 23 Aug, 2017 1 commit
  17. 21 Aug, 2017 1 commit
  18. 10 Aug, 2017 1 commit
  19. 06 Aug, 2017 5 commits
    • Rémi Denis-Courmont's avatar
      vout: rename and invert has_hide_mouse · fae9d3db
      Rémi Denis-Courmont authored
      The name was meant to indicate that the plugin hid the pointer
      autonomously. But it was ostensibly confused as implying that the
      plugin had a mouse hiding control (basically the exact opposite).
      In fact, it really meant that the mouse hiding control would NOT be
      used.
      
      This tries to provide a slightly less confusing name, and accordingly
      inverts the meaning. This is also realigns with the convention that
      false should be the default setting (which was broken in the grand
      grand parent of this changeset).
      fae9d3db
    • Rémi Denis-Courmont's avatar
      vout: make has_hide_mouse true by default · 2341a17f
      Rémi Denis-Courmont authored
      Most video output displays either defer mouse hiding to the window
      provider plugin, or do not handle the mouse events at all.
      So that ought to be the default.
      
      Also kepe the default in those display plugins that don´t handle
      cursor hiding (notably X11 ones).
      2341a17f
    • Rémi Denis-Courmont's avatar
      display: always compute mouse hide timeout · 758adf9e
      Rémi Denis-Courmont authored
      Cursor management should eventually be moved from display to the window.
      (The video output core still needs to take care of the window to video
      translation, and the mouse event filtering, as it already does.)
      And cursor hiding should be left entirely within the window plugins.
      
      But in the mean time, compute the correct timestamps regardless of the
      cursor hiding support in display plugin, so that it also works if the
      cursor is hidden by the window plugin.
      758adf9e
    • Rémi Denis-Courmont's avatar
    • Rémi Denis-Courmont's avatar
      window: report all double clicks · 8bf2942e
      Rémi Denis-Courmont authored
      Because we can.
      8bf2942e