Functional reactive programming library
#6Is NominalDiffTime really correct?
Judging by the haddock of NominalDiffTime over at http://hackage.haskell.org/package/time-1.4.2/docs/Data-Time-Clock.html#t:NominalDiffTime it seems like NominalDiffTime doesn't include leap seconds in its notion of elapsed time.
It seems to me like this would make clockSession (and thus wires using it) unreliable around leap seconds with possible off-by-one errors being introduced, although I haven't tried to test this yet. :)
- added tag bug
The correct way to get time deltas is to use the operating system's monotonic clock, which is really a monotonically increasing value that doesn't jump the way the regular system time could. In the next feature release I will switch from the time library to the clock library. It may never happen though, depending on whether or not the newer wires library has taken over at that point.
