>All this will become moot anyway as soon as tzcode2000f/tzdata2000f are
>imported into NetBSD (or you just fetch and use those) - yearistype is
>no longer needed for anything in that data set.

indeed!  i just installed these on a machine of mine.

>But to comment on the comment that "ordinary mortals" (those who don't
>build snapshots) don't (or perhaps even shouldn't) be running zic in
>the first place.   That's nonsense - if that were true then zic wouldn't
>need to exist in the first place, the rules could just be compiled into
>libc (the way they used to be in the old days).  Snapshot builders can
>produce a new libc as easily as anything else (well, almost...)  The whole
>point of the external timezone database is to allow users to update their
>timezone data when the rules change.

please...i never said that "ordinary mortals" would not use it.  i
only said that its only purpose was to aid in the creation of the time
zone data files.

i understand that users should feel free to update their time zone
databases as they see fit.  or even make their own (although i've
never bothered with that).  just because you *can* do something,
doesn't mean you should...or that you will.  i do it all the time.
but...

(a) how many *other* people do?
(b) out of how many people who use the resulting files?
(c) out of how many people who could update them if they wanted to?

i think the ratio of a:b or a:c will be *VERY* small.  *that* was my
point.  i always thought that yearistype was not installed for general
use because it would only be used by about 1/10000 people.  probably
less.  certainly it *is* being used by people who do make build
frequently, but they don't notice it.  aren't there programs in the
kerberos source that also fit that profile?

-- 
|-----< "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."