This repository is a temporary home for code to generate "validation" text files for the IANA-hosted time zone data.

For background, see http://mm.icann.org/pipermail/tz/2015-July/022349.html

If the approach proves successful, more discussion around the file format, appropriate generation tools etc will be required. Initially, all code within this repository should be seen as highly experimental, and likely to be hacked together with little thought to longevity.

Data

I'm temporarily hosting data sourced from zic/zdump (which I view as canonical) at https://nodatime.github.io/tzvalidate/. You can download a zip file of each release you're interested in, or one containing everything we've got.

Current implementations and requirements

All tools support the following command line options:

• -f: The "from" year - where to start looking for transitions. Defaults to "before any real transitions" (in a platform-specific way; often 1AD).
• -t: The "to" year - where to stop looking for transitions. Defaults to 2035.
• -z: The ID of single zone to dump, where applicable

The "initially" line is unaffected by the from/to year combination.

Noda Time

The Noda Time code is now within the Noda Time repository to enable direct handling of tz data.

Loads data from one of three sources:

• A Noda Time nzd file (local or via http/https)
• A tzdata directory
• A tzdata tar.gz file (local or via http/https)

Additional command line options:

• -s: (required) source file or directory

Reports initial offset and all transitions within bounds. Code exists in the main Noda Time repository, in order to use an assembly which isn't exported in releases.

Sample command lines (from in the C# source directory):

• dotnet run -- -s http://nodatime.org/tzdb/tzdb2015e.nzd
• dotnet run -- -s www\tzdb\tzdb2015e.nzd
• dotnet run -- -s data\tzdb\2015e
• dotnet run -- -s http://www.iana.org/time-zones/repository/releases/tzdata2015e.tar.gz
• dotnet run -- -s downloads\tzdata2015e.tar.gz
• dotnet run -- -s http://nodatime.org/tzdb/tzdb2015e.nzd -f 1900 -t 2000

TimeZoneInfo (.NET)

Reports transitions in .NET's "native" TimeZoneInfo class. Unless you're running on a system using time zone data built from the IANA sources, this is unlikely to be similar to any other output

• in particular, many zones have recurrences which effectively start in 1AD.

Code: csharp/src/NodaTime.TzValidate.TimeZoneInfoDump

Supports the regular command line options, and no additional options.

zic

C# code to report transitions stored in the output of zic.

Requires:

• A "modern" version of zic to have compiled a data source already, as per the IANA web site.

Code: csharp/src/NodaTime.TzValidate.ZicDump

Additional command line options:

• -s: (required) source file or directory; either a file for a single zone, or a directory to be searched recursively for files
• -i: (optional) whether or not to fake an initial offset for zones which don't report a transition at the big bang; values are "true" or "false" - defaults to "true"

Sample command line: dnx run -s tzdb2015e

TZS

C# code to convert the TZS file included in the complete distribution file into tzvalidate format.

Code: csharp/src/NodaTime.TzValidate.ConvertTzs

Sample command line (from in the C# source directory):

• dotnet run -- -s https://data.iana.org/time-zones/releases/tzdb-2017c.tar.lz
• dotnet run -- -s to2050.tzs -v 2017c

Java 7

Uses java.util.TimeZone, probing each day for transitions (as there's no "next transition" method).

Requires:

• A JRE update to the time zone database you want to use. See the Timezone updater tool
• Java 7 (it probably works with Java 1.4 or something ridiculous, but importantly it doesn't require Java 8)
• Apache Commons CLI - v1.3.1 or later

Code: java/org/nodatime/tzvalidate/Java7Dump

Sample command line: java -cp bin;lib/commons-cli-1.3.1.jar org.nodatime.tzvalidate.Java7Dump

Java 8

Uses java.time.* to find zones and transitions.

Requires:

• A JRE update to the time zone database you want to use. See the Timezone updater tool
• Java 8+
• Apache Commons CLI - v1.3.1 or later

Code: java/org/nodatime/tzvalidate/Java8Dump

Sample command line: java -cp bin;lib/commons-cli-1.3.1.jar org.nodatime.tzvalidate.Java8Dump

Joda Time

Uses Joda Time's built-in tz compiler to work from source data.

Requires:

• Joda Time of some version - probably nothing too recent, but I use 2.8.1
• Apache Commons CLI - v1.3.1 or later

Code: java/org/nodatime/tzvalidate/JodaDump

Additional command line options:

• -s: (required) directory containing tzdata source

Sample command line: java -cp bin;lib/joda-time-2.8.1.jar;lib/commons-cli-1.3.1.jar org.nodatime.tzvalidate.JodaDump -s ../tzdata/2015e

ICU4J

Uses the time zone database built into ICU4J.

Requires:

• ICU4J - probably nothing too recent, but I use 55.1
• An updated time zone database in the jar file; see the ICU4J Time Zone Update Utility for more details.
• Apache Commons CLI - v1.3.1 or later

Code: java/org/nodatime/tzvalidate/IcuDump

Sample command line: java -cp bin;lib/icu4j-55_1.jar;lib/commons-cli-1.3.1.jar org.nodatime.tzvalidate.IcuDump

Note: ICU4J appears not to have any way of getting at time zone abbreviations (PST, PDT etc). I've tried various approaches, but nothing gives the expected results. Improvements welcome!

Ruby

Uses the tzinfo package to get detailed data.

Requires:

• tzinfo gem
• tzinfo-data gem
• trollop gem

Code: ruby/tzdump.rb

Sample command line: ruby ruby/tzdump.rb