1. 22 Aug, 2008 1 commit
  2. 20 Aug, 2008 1 commit
  3. 17 Aug, 2008 3 commits
  4. 14 Aug, 2008 1 commit
  5. 12 Aug, 2008 2 commits
  6. 01 Aug, 2008 1 commit
  7. 22 Jul, 2008 2 commits
  8. 21 Jul, 2008 1 commit
  9. 19 Jul, 2008 2 commits
  10. 16 Jul, 2008 1 commit
  11. 15 Jul, 2008 2 commits
  12. 13 Jul, 2008 1 commit
  13. 05 Jul, 2008 1 commit
  14. 17 Jun, 2008 2 commits
  15. 15 Jun, 2008 1 commit
  16. 14 Jun, 2008 2 commits
  17. 01 Jun, 2008 4 commits
  18. 31 May, 2008 1 commit
  19. 23 Apr, 2008 1 commit
  20. 16 Apr, 2008 1 commit
  21. 14 Apr, 2008 1 commit
  22. 07 Apr, 2008 1 commit
  23. 11 Mar, 2008 1 commit
    • Rafaël Carré's avatar
      Do not use playlist_*AddInput() when the playlist has already been destroyed... · 0e86a35f
      Rafaël Carré authored
      Do not use playlist_*AddInput() when the playlist has already been destroyed and is dying (fix a crash when exiting while directory access is still reading) Try to return early if one of those function fail, because it either mean we don't have anymore memory available (I'm still looking for infinite Turing machines), or we are exiting (the playlist is dying). Use object / thread helpers for simplicity
      0e86a35f
  24. 03 Mar, 2008 1 commit
  25. 21 Feb, 2008 1 commit
  26. 20 Feb, 2008 1 commit
  27. 23 Jan, 2008 1 commit
  28. 01 Jan, 2008 1 commit
  29. 30 Dec, 2007 1 commit
    • Rafaël Carré's avatar
      fix #1410. · 7b9e1799
      Rafaël Carré authored
      You can not change a playlist_item_t to be a node, when you previously registered events on it, using input_item_t sub items adding event, because the pointer will be destroyed.
      If you want to use nodes anyway, you have to explicitely create them.
      Note: we were lucky that xspf did work so far, because we were EXCLUSIVELY using invalid pointers, except for the 1st item being added to each playlist_item_t (it did happen before the conversion obviously).
      7b9e1799