3381 2005-06-09 07:01:17 -0700 Date.prototype.setDate() incorrect for values >=128 2005-08-07 12:56:50 -0700 1 1 1 Unclassified WebKit JavaScriptCore 412 Mac OS X 10.4 RESOLVED FIXED P2 Major --- 3759 1 gavin ggaren oldest_to_newest 11159 0 gavin 2005-06-09 07:01:21 -0700 var d = new Date; var out = "Today: " + d + "\n"; d.setDate( d.getDate() + 280 ) out += "Future: " + d; Should result in a date 280 days in the future. This works on Firefox and IE: Today: Thu Jun 09 2005 07:59:31 GMT-0600 (MDT) Future: Thu Mar 16 2006 07:59:31 GMT-0700 (MST) but fails on the latest version of Safari (2.0.412): Today: Thu Jun 09 2005 08:00:45 GMT-0600 Future: Sun Jul 03 2005 08:00:45 GMT-0600 11659 1 gavin 2005-06-12 17:21:11 -0700 The failure seems to be with date values greater than 127. This test code: var out = ''; for ( var i=1; i<300; ++i ) { var theDate = new Date; var theValue = theDate.getDate( ) + i; theDate.setDate( theValue ); out += theValue + '::' + theDate + '\n'; } produces the following output (values from 1..120 are contiguous from the current date): 121::Thu Sep 29 2005 18:19:59 GMT-0600 122::Fri Sep 30 2005 18:19:59 GMT-0600 123::Sat Oct 01 2005 18:19:59 GMT-0600 124::Sun Oct 02 2005 18:19:59 GMT-0600 125::Mon Oct 03 2005 18:19:59 GMT-0600 126::Tue Oct 04 2005 18:19:59 GMT-0600 127::Wed Oct 05 2005 18:19:59 GMT-0600 128::Sun Jan 23 2005 18:19:59 GMT-0700 129::Mon Jan 24 2005 18:19:59 GMT-0700 130::Tue Jan 25 2005 18:19:59 GMT-0700 131::Wed Jan 26 2005 18:19:59 GMT-0700 132::Thu Jan 27 2005 18:19:59 GMT-0700 133::Fri Jan 28 2005 18:19:59 GMT-0700 15103 2 3067 ggaren 2005-07-23 11:43:27 -0700 Created attachment 3067 Patch The problem is overflow. CFGregorianDate is only an 8-bit signed value. This patch fixes setDate, but setMonth, setYear, and setFullYear still have the same problem. 15104 3 ggaren 2005-07-23 11:46:33 -0700 (In reply to comment #2) [ Should read: } In CFGregorianDate, month, day, hour, and minute are 8-bit signed values. SetMonth still has the same problem. 15173 4 3067 darin 2005-07-25 11:56:22 -0700 Comment on attachment 3067 Patch Looks good, r=me. 16117 5 cryo 2005-08-07 10:04:01 -0700 For some reason the code is thinking my current date is epoch instead of Sun Aug 7 13:02:14 EDT 2005 so the check fails because the year is wrong? This test checks for regression against: 3381 Date.prototype.setDate() incorrect for values >=128. If the test passes, all the dates below will be sequential. setDate(120): Mon Mar 30 1970 19:00:00 GMT-0500 setDate(121): Tue Mar 31 1970 19:00:00 GMT-0500 setDate(122): Wed Apr 01 1970 19:00:00 GMT-0500 setDate(123): Thu Apr 02 1970 19:00:00 GMT-0500 setDate(124): Fri Apr 03 1970 19:00:00 GMT-0500 setDate(125): Sat Apr 04 1970 19:00:00 GMT-0500 setDate(126): Sun Apr 05 1970 19:00:00 GMT-0500 setDate(127): Mon Apr 06 1970 19:00:00 GMT-0500 setDate(128): Tue Apr 07 1970 19:00:00 GMT-0500 setDate(129): Wed Apr 08 1970 19:00:00 GMT-0500 16126 6 ggaren 2005-08-07 12:56:50 -0700 The output doesn't look like a failure -- all the dates are sequential, as they should be. The test also has nothing to do with your current date. The year should be 1970, since the constructor call is new Date(0). Subsequently, setDate(120) gives the 120th day in 1970, which is Mar 30. This test also doesn't fail when I execute run-webkit-tests. Changing resolution to FIXED. 3067 2005-07-23 11:43:27 -0700 2005-07-25 11:56:22 -0700 Patch MakeDatePatch.txt text/plain 621 ggaren SW5kZXg6IGtqcy9kYXRlX29iamVjdC5jcHAKPT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PQpSQ1MgZmlsZTogL2N2cy9yb290 L0phdmFTY3JpcHRDb3JlL2tqcy9kYXRlX29iamVjdC5jcHAsdgpyZXRyaWV2aW5nIHJldmlzaW9u IDEuNDQKZGlmZiAtdSAtcjEuNDQgZGF0ZV9vYmplY3QuY3BwCi0tLSBranMvZGF0ZV9vYmplY3Qu Y3BwCTIzIEp1bCAyMDA1IDAwOjUwOjM2IC0wMDAwCTEuNDQKKysrIGtqcy9kYXRlX29iamVjdC5j cHAJMjMgSnVsIDIwMDUgMTg6NDE6MjIgLTAwMDAKQEAgLTcxMCw4ICs3MTAsOSBAQAogICAgICAg bXMgPSBhcmdzWzNdLnRvSW50MzIoZXhlYyk7CiAgICAgYnJlYWs7CiAgIGNhc2UgU2V0RGF0ZToK LSAgICB0LT50bV9tZGF5ID0gYXJnc1swXS50b0ludDMyKGV4ZWMpOwotICAgIGJyZWFrOworICAg ICAgdC0+dG1fbWRheSA9IDA7CisgICAgICBtcyArPSBhcmdzWzBdLnRvSW50MzIoZXhlYykgKiBt c1BlckRheTsKKyAgICAgIGJyZWFrOwogICBjYXNlIFNldE1vbnRoOgogICAgIHQtPnRtX21vbiA9 IGFyZ3NbMF0udG9JbnQzMihleGVjKTsKICAgICBpZiAoYXJncy5zaXplKCkgPj0gMikK