On 23-2-2019 13:17, Lester Caine wrote:
On 23/02/2019 11:39, Mark Rotteveel wrote:
That ignores the reality of drivers that are aware of the existence of
a feature but haven't yet come around to implementing it (or don't
want to implement it).
Currently I have no intention of touching FB4 simply because it will
conflict with proper reliable management of timezone data. In order to
ensure my historic material is not corrupted by the limits imposed by
the FB4 hacks to ignore second accuracy in TZ offsets. It's bad enough
that the raw data is already corrupted by ignoring pre-1970 rules and
serving up the truncated data without ANY indication. Then taking that
data and limiting it to minute accuracy is simply wrong. I see the
current offering as being something that will have to be completely
reworked at some point to be of any use for historic data sets, but then
the complete TZ system has to be fixed first! We take the time to build
accurate rules sets for events as recent as the 2nd world war, and then
they get thrown out because it's 'too much trouble to manage them'!
Currently many other parts of the world are going through the same
exercise of accurately documenting pre-1970 data, but there is no way
currently of distributing this material :(
I'm not sure why you replied this to my post. It doesn't seem relevant
in the context of my thread.
Mark
--
Mark Rotteveel
Firebird-Devel mailing list, web interface at
https://lists.sourceforge.net/lists/listinfo/firebird-devel