Guys, please don't waste too much time on this issue! The tracker still
has more than 1,700 open issues for to debate on. While I enjoy the
fact, that you three are enthusiastic, I strongly feel the urge to
re-route your men power. This bug isn't important enough to waste your
precious time on it.

My opinion as junior core developer is:

sysconf(_SC_CLK_TCK) is the winner and it should be used instead of HZ
when available. A default value should not be used because it will lead
to wrong data. Wrong results are worse than no results.

Since calls to sysconf seem to cost several CPU cycles "clk_tck" should
be cached somehow. I prefer a local static variable in the function but
a static var on file level is fine, too.

The compilation of Python must not fail. When neither HZ nor sysconf is
available but HAVE_TIMES is defined then the function must not be
included. Either you skip the function plus undef HAVE_TIMES or you add
some code to configure.in that does it earlier. The C89 standard doesn't
define #warn so that not an option, too. But configure is allowed fail
for a broken system.

Christian