Subject: Re: Australian Timezone Correction - NetBSD-1.4.2
To: Robert Elz <kre@munnari.OZ.AU>
From: Andrew Brown <atatat@atatdot.net>
List: port-sparc
Date: 08/13/2000 00:51:33
>  | yep.  the sort of thing it does should really be *in* zic.  i agree.
>
>No, that's not the right fix.   There's no way to predict what kind
>of crazy rule some legislature somewhere might dream up.  Attempting
>to put enough generality into zic to deal with it would probably be
>impossible, and certainly be unreasonable.

i only said that because the meat of what yearistype.sh does now is:

        2-even) 
                case $1 in
                        *[24680])                       exit 0 ;;
                        *)                              exit 1 ;;
                esac ;;
        2-nonpres|2-nonuspres)
                case $1 in
                        *[02468][048]|*[13579][26])     exit 1 ;;
                        *)                              exit 0 ;;
                esac ;;
        2-odd)  
                case $1 in
                        *[13579])                       exit 0 ;;
                        *)                              exit 1 ;;
                esac ;;
        2-uspres)
                case $1 in
                        *[02468][048]|*[13579][26])     exit 0 ;;
                        *)                              exit 1 ;;
                esac ;;

and that seems easy enough to code in c.

>  | indeed!  since zic is the only program that uses it, the functionality
>  | that it provides ought to be *in* zic.
>
>No, that's no different than saying that the tzdata* files should be built
>into zic.   yearistype is part of the data set.   It is something that
>any random site needs to be able to change if they have a need to
>implement some strange timezone rule in a hurry (and waiting for a new
>tz data set, let alone code update, is not always an option).

it's very different.  i would never even begin to consider thinking
about suggesting the possibility of putting the tzdata files in zic.
that would be silly.  true...year is type does come from the same tar
ball as the data files, but it seems more like a massive kluge than
anything else.

if the time zone data files were that flexible, i ought to be able to
make one for local adjusted solar time (and have date suck my
longitude from somewhere) or for a judaic calendar.

at least...i'd like to see those things.

>What needs to be fixed is the way zic invokes yearistype when it is
>needed, not the existance of it in the first place.

make it stuff . in the front of the path it uses?  and/or use the
prefix (up to the last /) of the input files it's reading?

-- 
|-----< "CODE WARRIOR" >-----|
codewarrior@daemon.org             * "ah!  i see you have the internet
twofsonet@graffiti.com (Andrew Brown)                that goes *ping*!"
andrew@crossbar.com       * "information is power -- share the wealth."