On Wed, 26 Mar 2025 20:31:07 GMT, Justin Lu <j...@openjdk.org> wrote:
> Please review this PR which clarifies the behavior for integer and fraction > limits in NumberFormat and implementing classes. An associated CSR is filed. > > There have been a few bugs submitted which indicate a misconception that > these limits impact parsing. The actual behavior is that these limits only > affect formatting. The specification is vague regarding this, and can be > explicitly updated to eliminate confusion. As the implementing classes are > updated to use `inheritDoc`, some shuffling around in the method specs are > included in this change as well. > > Alternatively I considered making this change as implementation specific to > DecimalFormat and CompactNumberFormat only. (i.e. leave flexibility for other > NumberFormat subclasses to define their own behavior on whether the limits > affect parsing.) I am open to this option as well, but initially decided > against it as > 1) Unlike formatting, it seems like a rare use case that you would want to > suppress the range of digits of accepted during parsing. > `setParseIntegerOnly()` already provides functionality to toggle between > integer and fraction parsing. > 2) The limits affecting formatting only has been the long-standing behavior > for all the subclasses of NumberFormat provided by the OpenJDK reference > implementation. This pull request has now been integrated. Changeset: e149bd3a Author: Justin Lu <j...@openjdk.org> URL: https://git.openjdk.org/jdk/commit/e149bd3a0c387698fe7b2ff6c373944ded773b8b Stats: 99 lines in 3 files changed: 9 ins; 20 del; 70 mod 8352755: Misconceptions about j.text.DecimalFormat digits during parsing Reviewed-by: naoto ------------- PR: https://git.openjdk.org/jdk/pull/24265
