RFR: 8228469: (tz) Upgrade time-zone data to tzdata2019b
ramanand.patil at oracle.com
Tue Aug 6 04:44:02 UTC 2019
Thank you Martin and Naoto for your reviews.
Currently, we are evaluating the approach for update releases. But in my opinion, it is better to migrate to vanguard format.
From: Martin Buchholz <martinrb at google.com>
Sent: Tuesday, August 6, 2019 1:57 AM
To: Ramanand Patil <ramanand.patil at oracle.com>
Cc: core-libs-dev <core-libs-dev at openjdk.java.net>; i18n-dev <i18n-dev at openjdk.java.net>
Subject: Re: RFR: 8228469: (tz) Upgrade time-zone data to tzdata2019b
Thanks for the update and redundancy removal. Looks good to me.
What is the recommendation for older releases? Migrate to vanguard format by backporting recent changes or stay on rearguard forever?
On Mon, Aug 5, 2019 at 1:28 AM Ramanand Patil <mailto:ramanand.patil at oracle.com> wrote:
Please review the patch for tzdata2019b integration into jdk project(jdk-14).
- This patch uses "vanguard" format tzdata from IANA (instead of rearguard format), as the enhancement bug is fixed now. Many thanks to Naoto for this.
- As per the latest changes to Palestine zone rules, the hard-coded checks for "Asia/Gaza" and "Asia/Hebron" are no more needed in ZoneInfoFile.java.
Because of those checks test/jdk/sun/util/calendar/zi/TestZoneInfo310.java was failing as mentioned in the bug comment.
- test/jdk/java/util/TimeZone/TimeZoneTest.java is updated to set ZoneDescriptor.daylight value to "false" for BET zone id.
- The patch has passed all the related testing including JCK.
- As per the mail from tzdata maintainers there is a possibility that Brazil might not abolish DST.
- Since the enhancement bug is not fixed in update releases, jdk13u will still use the rearguard format tzdata. A separate review request will be sent for the same.
More information about the core-libs-dev