[14] RFR: 8212970: TZ database in "vanguard" format support

Roger Riggs Roger.Riggs at oracle.com
Thu Jul 25 14:47:53 UTC 2019

Hi Naoto,

With the composition of the tzdir path up and over to the make directory 
for the tzdir
it might be useful to do an explicit check that the directory exists.
That way if the directory structure on the build side changes,
there will be a test failure makine it obvious that the dependency has 

Looks fine.

Thanks, Roger

On 7/24/19 6:24 PM, naoto.sato at oracle.com wrote:
> Hi Joe,
> Thank you for the review.
> On 7/24/19 2:57 PM, Joe Wang wrote:
>> Hi Naoto,
>> The method findNegativeSavings method in TzdbZoneRulesProvider.java 
>> states that it "Find the minimum negative savings". While the result 
>> is correct since the rules all have the same value for SAVE, I wonder 
>> if that's ideal conceptually. Given a start LDT, shouldn't it be 
>> looking for the SAVE in the exact (narrower) date range (e.g. 1981 - 
>> 1989 vs 1981 - max)?.
> I believe it is working as such. The end year is retrieved within the 
> method (line 879) and only the minimum negative saving values within 
> the window is filtered.
>> NegativeDSTTest verifies the tzdata, that is the adjusted data after 
>> import, is that correct? I wonder a comment and a bit of details in 
>> the test summary would be helpful since there is no negative data in 
>> the test itself.
> Good point. It is confusing. I supplied summary text in the test (also 
> the similar line in TestZoneRules.java)
> Here is the updated webrev:
> http://cr.openjdk.java.net/~naoto/8212970/webrev.11/
> Naoto
>> Best,
>> Joe
>> On 7/23/19 3:15 PM, naoto.sato at oracle.com wrote:
>>> Hi,
>>> Please review the fix to the following enhancement:
>>> https://bugs.openjdk.java.net/browse/JDK-8212970
>>> The proposed changeset is located at:
>>> https://cr.openjdk.java.net/~naoto/8212970/webrev.09/
>>> This change aims to support the "vanguard" IANA time zone data 
>>> format, which uses the negative savings and transition time beyond a 
>>> day period. The change basically translates those negative savings 
>>> and transition times, such as 25:00, into the ones that the current 
>>> JDK recognizes, then produces the data file "tzdb.dat" at the build 
>>> time. At the run time, the data file is read and interpreted as 
>>> before. This way the produced tzdb.dat is compatible with the prior 
>>> JDK releases so that the TZ Updater can also distribute it as a time 
>>> zone update.
>>> I have also refactored redundant copy of ZoneRules file in the build 
>>> directory, by dynamically importing the file under src. Thus some 
>>> build related files are modified. I am hoping folks on the build-dev 
>>> can review those changes.
>>> Naoto

More information about the core-libs-dev mailing list