On Friday 26 July 2002 05:32 am, stefan@extum.com wrote:
> I was able to run the regression tests. I have got 3 tests failed
> during my run. See below.
> abstime ... FAILED
> tinterval ... FAILED
> test horology ... FAILED
> Looking and inspecting the regression.diffs I have got lots of things
> which somebody could explain me ?
> I have attached the regression.diffs . Can I make sure my PG is properly
> installed in this way, having 3 steps failed ?
These are due to Red Hat's decision to change the behavior of mktime(3), which
is used in those tests for timezone. They did this without any heads up of
any kind, and now claim it was to be 'standards compliant'. A heads up would
have allowed us to made a workaround available earlier.
The solution is not to do timestamp with timezone calculations for dates prior
to 1970, until such time this problem is fixed. Red Hat believes it is not
their problem, and we believe it's not our problem. But a workaround is
possible.
You must have run the tests under the C locale, as nothing else failed.
--
Lamar Owen
WGCR Internet Radio
1 Peter 4:11
