Tobias Oberstein <[email protected]> added the comment: CPython 2.7.6 tests for ftime() availability
https://gist.github.com/oberstet/7516529#file-gistfile1-txt-L327 and then only includes if HAVE_FTIME http://hg.python.org/cpython/file/cc27c0aba18b/Modules/timemodule.c#l30 On above build log, ftime() is undetected - I'm not sure how that works, since ftime() is still there on FreeBSD (as is the header sys/timeb.h), but it's deprecated .. and Clang will produce a warning. As far as I can see from the source here http://hg.python.org/cpython/file/cc27c0aba18b/Modules/timemodule.c#l905 CPy will prefer gettimeofday to ftime, even if both are available. I find that C style .. strange. They have 2 blocks of C, where the first returns, so the second is unreachable code if both HAVE_GETTIMEOFDAY and HAVE_FTIME are defined. Well. In no case does CPy 2 use clock_gettime() - which is a pity. It seems, this is only available in CPy 3 via http://www.python.org/dev/peps/pep-0418/ So if PyPy is to model CPy 2, we might use gettimeofday() when avail, otherwise fallback to ftime(). ________________________________________ PyPy bug tracker <[email protected]> <https://bugs.pypy.org/issue1640> ________________________________________ _______________________________________________ pypy-issue mailing list [email protected] https://mail.python.org/mailman/listinfo/pypy-issue
