On Wed, 10 Feb 2021 02:24:57 GMT, Nir Lisker <nlis...@openjdk.org> wrote:
>> Fixes a mutability issue for `LocalDateTimeStringConverter` (and >> `LocalDateStringConverter`) where the chronology can change during the >> lifetime of the instance and cause an inconsistent state. The following >> changes were made: >> >> * The input is now formatted and parsed directly with the formatter and >> parser (which are configured with a chronology) without the chronology field >> value of the class. >> * The chronology field value does not change anymore when there is an >> exception due to inability to format the date. >> * Logging on failed formatting was removed as it did not seem useful. The >> formatter will throw the same exception that the chronology field does >> anyway, so there is not much use to catching the exception before hand. >> >> Added a test that fails without this patch. The test creates a converter >> with an explicit `Chronology` and `null` parser and formatter. The default >> formatter that is created with the given chronology formats a legal date >> (with respect to the chronology) to a string, which the parser should be >> able to convert back to a date. However, by forcing an exception in the >> formatting process using an illegal date, the chronology changes, and now >> when the parser is used it is configured with the new chronology and it >> can't parse the string correctly. > > Nir Lisker has updated the pull request incrementally with one additional > commit since the last revision: > > Removed printing and narrowed the exception looks ok to me ------------- Marked as reviewed by pbansal (Committer). PR: https://git.openjdk.java.net/jfx/pull/393