Steve Grubb wrote: This change I believe fixed the problem with the error test (ie. the path length = 4095 was wrongfully generating error), however it broke the following case ... this is part of the pathname with spaces test ... (it's equivalent case, without the space passes fine!): DEBUG: Expected path length = 4096: Actual is 4096 (Not counting terminating null) Starting auditd: [ OK ] watch_path_test 3 FAIL : Path length =PATH_MAX(without null)+1 did not generate error Any idea what could cause this behavior? Thanks, - Loulwa

Steve Grubb wrote: Passing Passing Passing unexpectedly (Only for path with a space) Failing (I am testing 4099) Note: the case where it passes with 4096 is only occurring when the path includes a space. The same exact test case, with path not including spaces, is behaving as expected (ie. failing). If you remember, the problem that was occurring before (ie. strlen()=4095 comes back with error from kernel), was behaving opposite to this, failing for path without spaces, but passing for path with spaces ... I don't see why the space should make a difference, shouldn't it just count as one character in the strlen(), unless I am missing something? - Loulwa