A whole bunch of packages in Gentoo has problems with the difference between TAI (time relative to cesium clocks) and UTC (time relative to astronomical obversations). For an explanation of this difference and it's effects, please see http://www.thedjbway.org/clockspeed/leapsecs.html and http://cr.yp.to/proto/utctai.html. The problems I've encountered in Gentoo so far were: - OpenNTPd not setting the clock correctly (as compared with my radio controlled wrist watch) - test suite failures due to hardcoded, not leap second-adjusted timestamps I'm creating this tracker bug in order to group those - otherwise apparently unrelated - bugs together.
Bug #109472 is/was somewhat related. It's about updating the leap seconds table contained in glibc.
Wikipedia has a good description of TAI [1], also noting that UTC in fact is derived from TAI, so you might see why it's important to get the leap second correction right (instead of just ignoring it). [1] http://en.wikipedia.org/wiki/Atomic_time
Fix dependancy sense (was inverted).
err, what has curl got to do with this bug? the cookie timestamp was expired, not off by a leap second.
See bug #111555 comment #23 for an explanation of why it's listed on this bug.
so what exactly is the solution to all this?
(In reply to comment #6) > so what exactly is the solution to all this? +1
The fact we dont have a solution shall not prevent talking about the issue. That's why BTS exist: list problemes that do not have solution yet. Please add "TRACKER" to "Keywords" field.
Closing due to depending bugs being fixed.