The changeset is attached. (I don't have commit access.) On Mon, Aug 14, 2017 at 3:57 PM, Martin Buchholz <marti...@google.com> wrote:
> Reviewed! > > Please push to jdk10/jdk10/jdk. > > On Mon, Aug 14, 2017 at 1:07 PM, Liam Miller-Cushon <cus...@google.com> > wrote: > >> On Mon, Jul 31, 2017 at 11:41 AM, Liam Miller-Cushon <cus...@google.com> >> wrote: >> >>> On Fri, Jul 28, 2017 at 3:30 PM, Martin Buchholz <marti...@google.com> >>> wrote: >>> >>>> It looks like this will test only the year, not all the date fields. >>>> Shouldn't that be s/25/16/ ? >>>> Does this code handle the "true" epoch of 1980-01-01 ? >>>> >>> >>> It does now, thanks: >>> http://cr.openjdk.java.net/~cushon/8184940/webrev.02/ >>> >>> This raises the question of what it should do if a day or month is zero >>> but year/month/day are not all zero. Is 1980-0-0 special, or should >>> 1980-0-1 and 1980-1-0 also be tolerated for compatibility with JDK 8? >>> >> >> As discussed offline, there are no known instances of tools deliberately >> writing zip archives where a day or month is '0' other than 1980-0-0. So >> the current approach seems better than continuing to interpret e.g. >> 1980-1-0 as 1979-12-31 or 1980-0-1 as 1979-12-01. >> > >
