http://build.webkit.org/results/Apple%20MountainLion%20Release%20WK1%20(Tests)/r150833%20(10397)/fast/js/date-DST-time-cusps-pretty-diff.html - PASS Passed all tests (or skipped all tests if your timezone isn't PST/PDT) + FAIL 2/4 tests were failed: (new Date(1982, 2, 14, 2, 10)).getHours() should be 1. Was 3., (new Date(1982, 2, 14, 2)).getHours() should be 1. Was 3.
Marked failing in http://trac.webkit.org/changeset/150848. Please fix by the morning or I'll roll out the offending patch.
This is actually a progression! Per the description in the test case: "For times that shouldn't happen we currently go back an hour, but in reality we would like to go forward an hour. This has been filed as a radar: 4777813" Now it does! I'll update the expected results.
Created attachment 203116 [details] Update results
Comment on attachment 203116 [details] Update results View in context: https://bugs.webkit.org/attachment.cgi?id=203116&action=review Looks convincing. > LayoutTests/ChangeLog:5 > + <rdar://problem/4777813> Safari incorrectly handles times that shouldn't exist Also <https://bugs.webkit.org/show_bug.cgi?id=11157>, I guess. > LayoutTests/ChangeLog:14 > + - Preffer the latter option for non-existant time (due to clock rolling forwards). Typo: Preffer. Mistake: existant. Looks suspicious to me: forwards. > LayoutTests/fast/js/script-tests/date-DST-time-cusps.js:6 > "For times that happen twice the behavior of all major browsers seems to be to pick the second occurance, i.e. Standard Time not Daylight Time" Mistake: occurrance.
Fixed in r151000