New submission from STINNER Victor <victor.stin...@gmail.com>: floattime() raises an OSError if _PyTime_gettimeofday() returns secs=0 and usec=0. This is wrong because _PyTime_gettimeofday() cannot fail and secs=0 is valid if the current time is 0 (1970.1.1 at 12:00 UTC). Extract of _PyTime_gettimeofday() "doc":
/* Similar to POSIX gettimeofday but cannot fail. If system gettimeofday * fails or is not available, fall back to lower resolution clocks. */ PyAPI_FUNC(void) _PyTime_gettimeofday(_PyTime_timeval *tp); In practice, I suppose that nobody noticed the bug because we are in 2012, not in 1969 :-) And even if you change your system clock to 1969.11.31, you may not be able to reproduce the bug because you have to call gettimeofday() exactly at midnight to get also usec=0 ... Attached patch removes the exception. ---------- components: Library (Lib) files: floattime.patch keywords: patch messages: 156334 nosy: haypo priority: normal severity: normal status: open title: floattime() should not raise an exception versions: Python 3.3 Added file: http://bugs.python.org/file24939/floattime.patch _______________________________________ Python tracker <rep...@bugs.python.org> <http://bugs.python.org/issue14368> _______________________________________ _______________________________________________ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
