Andrew Ross wrote: >Hi Arjen, > >We had this discussion when I first implemented the date / time >functions. It is a trade off between using the widely available system >routines (with their limitations) and supporting our own set of date / >time formatting routines with an associated large overhead on the >developers. Of course in the long term 64-bit will solve many of the >problems. > >In the intermediate term there are often solutions. If you are plotting >over many years and just formatting the year, then you may well not need >to use the date / time formatting at all. > >If you are working over a shorter range of dates / times and don't need >the year, then it doesn't actually matter what year you choose. Example >29 illustrates this. (Just be careful about leap years). Also, you can >always format the year as a two digit year (%y, %g), and so the exact >century doesn't matter. > >The only easyish way around this I could think of which still retained >second accurancy would be to allow you to specify the epoch (i.e. the >date / time which corresponds to t = 0). We'd still need our own >implementation of strftime to actually use this though. > > Yes, I remember this discussion :). I had not realised the workarounds you mention, but they seem promising enough. I will add your answer to the SF request (so that it is immediately accessible to the OP).
Regards, Arjen ------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ _______________________________________________ Plplot-devel mailing list Plplot-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/plplot-devel
