Bruce,
I'm going to apply your patches, but I'm getting this:
Testcase: testSelectionBehaviour(org.apache.tools.ant.types.selectors.DateSelect orTest): FAILED expected:<TTFFTFFFTTTT> but was:<TFFFFFFFFFFT> junit.framework.AssertionFailedError: expected:<TTFFTFFFTTTT> but was:<TFFFFFFFF FFT> at junit.framework.Assert.fail(Assert.java:51) at junit.framework.Assert.failNotEquals(Assert.java:234) at junit.framework.Assert.assertEquals(Assert.java:68) at junit.framework.Assert.assertEquals(Assert.java:75) at org.apache.tools.ant.types.selectors.DateSelectorTest.testSelectionBe haviour(DateSelectorTest.java:194)
Well, at least the tests are working as they should. It passes on my platform, which is Debian Linux running JDK 1.3.1_02. The problem is platform specific.
This test is explicitly setting the number of milliseconds since January 1st, 1970 to 1006347300000L and checking for files which have exactly the same lastModified() timestamp. Now that I think of it, doesn't Windows have a 2 second granularity to its file system? I'm guessing the test failed for you on a Windows box. This failure is telling me that we need a granularity attribute on DateSelector in addition to DependSelector.
I'll whip up a patch to add this attribute and try to get it out tonight. I should probably set the default granularity to 2000 on Windows systems, 0 on everything else. If the patch is too late for Beta 2, you can just comment out the assertion on line 194 of DateSelectorTest and I'll get it all fixed up afterwards.
Thanks.
-- To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
