1. 06 Jan, 2009 1 commit
  2. 05 Jan, 2009 2 commits
  3. 19 Jun, 2007 1 commit
  4. 18 Jun, 2007 1 commit
  5. 02 Mar, 2006 1 commit
  6. 01 Mar, 2006 1 commit
  7. 12 Jul, 2005 1 commit
  8. 29 Apr, 2005 1 commit
  9. 27 Apr, 2005 1 commit
  10. 11 Jun, 2004 1 commit
  11. 04 Jun, 2004 1 commit
  12. 10 Apr, 2004 1 commit
  13. 16 Mar, 2004 1 commit
  14. 05 Mar, 2004 1 commit
  15. 03 Mar, 2004 1 commit
  16. 11 Apr, 2003 1 commit
  17. 29 Nov, 2002 1 commit
  18. 16 Oct, 2002 1 commit
  19. 09 Sep, 2002 1 commit
  20. 01 Jul, 2002 1 commit
  21. 03 Jun, 2002 2 commits
  22. 20 Feb, 2002 1 commit
  23. 09 Jan, 2002 1 commit
  24. 30 Nov, 2001 1 commit
  25. 31 Aug, 2001 1 commit
  26. 08 Aug, 2001 1 commit
  27. 12 Jul, 2001 1 commit
  28. 27 Jun, 2001 1 commit
  29. 10 Jun, 2001 1 commit
  30. 08 Jun, 2001 1 commit
  31. 04 Jun, 2001 1 commit
  32. 29 May, 2001 1 commit
  33. 18 May, 2001 1 commit
  34. 28 Apr, 2001 1 commit
  35. 25 Apr, 2001 1 commit
    • David Lawrence's avatar
      Make greatest_version() return an isc_result_t, which it should have · 2b5f436a
      David Lawrence authored
      been doing all along since there was the possibility isc_dir_open() would
      fail and overloading the return value of "0" was bad practice.
      
      If isc_dir_open() fails in greatest_version(), make sure the path separator
      is restored to the destination pathname if it was NULified before returning
      the error.
      
      The underlying problem of isc_log_open() not having the bad result of
      isc_dir_open() reported anywhere will be slightly mediate by a change
      to bin/named/logconf.c to test the directory.  lib/isc/log.c can't really
      handle it because it has no idea to where such an error should be reported,
      and it only shows up during isc_log_open() as part of the isc_log_write()
      family of calls, which are all void.
      2b5f436a
  36. 12 Apr, 2001 1 commit
  37. 28 Mar, 2001 2 commits