On Wed, 10 Sep 2025 22:20:22 GMT, Justin Lu <[email protected]> wrote:
> This PR expands the DateFormat benchmark with parsing cases (in addition to > the existing formatting ones) such that performance regressions are easier to > identify for parsing. Also adding in a time instance and respective > measurement tests. The patterns used correspond to FULL Locale.ENGLISH > patterns (w/ the NBSP replaced for time pattern for clarity sake). > > Further context in comment on JBS issue. Looks good. I'd suggest renaming the test case file, since this is now exclusively testing `SimpleDateFormat`, thus the test case name can be more specific. test/micro/org/openjdk/bench/java/text/DateFormatterBench.java line 61: > 59: > 60: private final String datePattern = "EEEE, MMMM d, y"; > 61: private final String timePattern = "h:mm:ss a zzzz"; Nit: These can be static fields (then the field names need to be all capital letters) ------------- PR Review: https://git.openjdk.org/jdk/pull/27204#pullrequestreview-3208139452 PR Review Comment: https://git.openjdk.org/jdk/pull/27204#discussion_r2338137214
