Bug 139138 – Several JavaScriptCore date tests are flaky, because they expect time to be frozen during execution

• Status: RESOLVED
• Resolution: FIXED
• Component: Tools / Tests

[Alexey Proskuryakov]

At least these tests are affected:

Source/JavaScriptCore/tests/mozilla/ecma/Date/15.9.2.1.js
Source/JavaScriptCore/tests/mozilla/ecma/Date/15.9.2.2-1.js
Source/JavaScriptCore/tests/mozilla/ecma/Date/15.9.2.2-2.js
Source/JavaScriptCore/tests/mozilla/ecma/Date/15.9.2.2-3.js
Source/JavaScriptCore/tests/mozilla/ecma/Date/15.9.2.2-4.js
Source/JavaScriptCore/tests/mozilla/ecma/Date/15.9.2.2-5.js
Source/JavaScriptCore/tests/mozilla/ecma/Date/15.9.2.2-6.js

mozilla-tests.yaml/ecma/Date/15.9.2.2-6.js.mozilla-dfg-eager-no-cjit-validate-phases: Detected failures:
mozilla-tests.yaml/ecma/Date/15.9.2.2-6.js.mozilla-dfg-eager-no-cjit-validate-phases: 15.9.2.2 The Date Constructor Called as a Function
mozilla-tests.yaml/ecma/Date/15.9.2.2-6.js.mozilla-dfg-eager-no-cjit-validate-phases: Date(2031,11,31,23,59,59) = Sun Nov 30 2014 23:48:48 GMT-0800 (PST) FAILED! expected: Sun Nov 30 2014 23:48:47 GMT-0800 (PST)
mozilla-tests.yaml/ecma/Date/15.9.2.2-6.js.mozilla-dfg-eager-no-cjit-validate-phases: Date(2032,0,1,0,0,0) = Sun Nov 30 2014 23:48:48 GMT-0800 (PST) PASSED!
mozilla-tests.yaml/ecma/Date/15.9.2.2-6.js.mozilla-dfg-eager-no-cjit-validate-phases: Date(2032,0,1,0,0,1) = Sun Nov 30 2014 23:48:48 GMT-0800 (PST) PASSED!
mozilla-tests.yaml/ecma/Date/15.9.2.2-6.js.mozilla-dfg-eager-no-cjit-validate-phases: Date(2031,11,31,16,0,0,0) = Sun Nov 30 2014 23:48:48 GMT-0800 (PST) PASSED!
FAIL: mozilla-tests.yaml/ecma/Date/15.9.2.2-6.js.mozilla-dfg-eager-no-cjit-validate-phases

The tests seem bad - not even Firefox freezes the current time during script execution. The tests have an unused TOLERANCE variable, so maybe whoever wrote them understood the problem.

[Alexey Proskuryakov]

This is a whole lot of flakiness for tests that do pretty much nothing (all they verify is that arguments are ignored!)

[Alexey Proskuryakov]

Looks like Mozilla folks fixed this long ago: <https://bugzilla.mozilla.org/show_bug.cgi?id=404351>.

[Alexey Proskuryakov]

Created attachment 242328 [details]
proposed fix

Manually merged the fix, because our version of shell.js is obsolete even compared to what Mozilla had in 2007. Perhaps we should update to trunk? I actually tried that just for laughs, but it wasn't easy enough.

[Mark Lam]

Comment on attachment 242328 [details]
proposed fix

View in context: https://bugs.webkit.org/attachment.cgi?id=242328&action=review

> Source/JavaScriptCore/tests/mozilla/ecma/Date/15.9.2.2-2.js:72
> +    array[item++] = new TestCase( SECTION, "Date(2000,0,1,0,0,0)", true, d2 - d1 <= 1000);

I was concerned about all the copy and pasting, and was wondering if it is better to write a utility function to eliminate the cut and paste like so:

function makeTestCase(dateExpr) {
    var d1 = new Date();
    var d2 = eval(dateExpr);
    return new TestCase(SECTION, dateExpr, true, d2 - d1 <= 1000);
}

array[item++] = makeTestCase("Date(2000,0,1,0,0,0)");

Here, we're seeing a discrepancy between the reported and the evaluated expressions.  I see that the discrepancy is also in the original test code (which is a bug).  Regardless, I think it'd be better to use a utility function instead.

> Source/JavaScriptCore/tests/mozilla/ecma/Date/15.9.2.2-2.js:76
> +    array[item++] = new TestCase( SECTION, "Date(2000,0,1,0,0,1)", true, d2 - d1 <= 1000)

Ditto.  Another discrepancy.

[Mark Lam]

Comment on attachment 242328 [details]
proposed fix

r=me for the sake of keeping the test code consistent with Mozilla's and easily mergeable.  However, please fix the 2 discrepancies.  That shouldn't make merging that much harder in the future.  Thanks.

[Alexey Proskuryakov]

Committed <http://trac.webkit.org/r176601>.