Date: Tue, 8 Aug 2000 09:40:48 -0400 From: "Olson, Arthur David (NCI)" <olsona@dc37a.nci.nih.gov> Message-ID: <4E0F9DDCA724D411B2120008C75DB71DD0F3F7@nihexchange2.nih.gov> | Hmmm...well then how about finessing the issue? That certainly avoids the problem for now, and I see that Paul has included that in his most recent update. Keeping the South Aus zone that way certainly makes sense (I will avoid comment on just why it is that Australia causes more problems than anywhere else with this kind of thing...) But the tzcode sources still need to be fixed one way or another, this change has just deferred the problem to later to bite us again when there is a need in the future (just maybe US Presidential time will appear on the radar again....) and by then all this discussion is likely to have been forgotten. What's currently in zic as the default is just broken. Some kind of decision needs to be made. Given that we now need no yearistype data at all, perhaps the solution is to simply drop the default of "yearistype" if no -y is given, and require explicit usage of the -y arg if a yearistype script is required to compile a zone file (no implicit default script name). That's a pretty simple change, I enclose a patch below that achieves that (though maybe this is the kind of error that should cause zic to abort rather than just report the error and continue, and I am by no means confident that the text of the error message is right). Probably the current yearistype.sh ought have some comments added to indicate that it is a sample of a script that might be used, and has been used in the past, but is no longer needed for anything. kre