On Tue, 30 Oct 2018 at 19:42, Paul Eggert <eggert@cs.ucla.edu> wrote:
On 10/30/18 8:52 AM, Carter, Tom via tz wrote:
Please advise when will this be fixed?
As announced in <https://mm.icann.org/pipermail/tz-announce/2018-October/000052.html> a rearguard-format tarball <https://web.cs.ucla.edu/~eggert/tz/release/2018g/tzdata2018g-rearguard.tar.g...> is intended to be usable by programs that don't yet support the timestamps in question. Please give that tarball a try and let us know of any problems with it.
Although the timestamp format was changed in 2007 this change was evidently overlooked in TZUpdater. Sorry, I don't know when TZUpdater itself will be fixed to support these timestamps; TZUpdater is a separate project maintained by a different group.
I made a very specific request to put 01:00 back in the main file for 15 months. A timeframe that would allow fixed/updated tools to be rolled out. Had my request and advice been followed, these developers and many others would not be frustrated today. FWIW, I've done my part - there are new releases of both time-zone parsers that I directly control (Joda-Time and ThreeTen-Backport), and I have argued for changes to the JDK. But changes to the JDK or TZUpdater will not be quick. I say again, the right course of action is to revert the change and use 01:00 in the main file until the JDK teams have had a chance to fix their tools. Stephen