I've sort of been ignoring this issue since Rainer reported the same for 64-bit AIX. This is largely because I was already a bit burded out with dealing with the twists and turns of floating point. It appears that basically this is getting converted to a double value, so it doesn't end up being an exact match. I've checked in some changes for those test cases that *might* fix the problem.
Rick On Sat, Jan 24, 2009 at 2:41 PM, Mark Miesfeld <miesf...@gmail.com> wrote: > Mostly addressed to Rick, but of course anyone can chime in. > > There are 3 failures in the test suite with 64-bit ooRexx testing the > new APIs. Although, I think it is just one failure repeated in 3 > places. Even using the fuzzy assert, the expected and actual don't > match: > > When you get a chance Rick, could you take at the following output. > The expected and actual diverge after the 7th decimal place. > > 3.40282347E+38 > 3.4028234663852886E+38 > ^^ > > > [failure] [20090124 11:13:56.862000] > svn: r3829 Change date: 2008-12-30 07:10:06 -0800 > Test: TESTFLOAT01 > Class: CONVERSION.testGroup > File: C:\4.0.0\ooRexx\API\oo\CONVERSION.testGroup > Line: 871 > Failed: assertFuzzyEquals > Expected: [[3.40282347E+38], identityHash="1099510666712"] > Actual: [[3.4028234663852886E+38], identityHash="1099489226052"] > > [failure] [20090124 11:13:56.924000] > svn: r3829 Change date: 2008-12-30 07:10:06 -0800 > Test: TESTFLOAT01 > Class: FUNCTION.testGroup > File: C:\4.0.0\ooRexx\API\oo\FUNCTION.testGroup > Line: 534 > Failed: assertFuzzyEquals > Expected: [[3.40282347E+38], identityHash="1099510643432"] > Actual: [[3.4028234663852886E+38], identityHash="1099483520344"] > > [failure] [20090124 11:14:33.565000] > svn: r3829 Change date: 2008-12-30 07:10:06 -0800 > Test: TESTFLOAT01 > Class: METHOD.testGroup > File: C:\4.0.0\ooRexx\API\oo\METHOD.testGroup > Line: 658 > Failed: assertFuzzyEquals > Expected: [[3.40282347E+38], identityHash="1099510950014"] > Actual: [[3.4028234663852886E+38], identityHash="1099483793156"] > > I'm not sure what is the appropriate thing to do here. (As everyone > probably knows by now, once a number gets past 100, I start having > problems. <grin>) > > The way you set up the fuzzy assert is to use numeric digits of 16 and > a fuzz of 1. This cleared up all the other failures on 64-bit > Windows. To be truthful, I don't know if this is a true failure, or > if the test just needs to be set up differently. > > -- > Mark Miesfeld > > ------------------------------------------------------------------------------ > This SF.net email is sponsored by: > SourcForge Community > SourceForge wants to tell your story. > http://p.sf.net/sfu/sf-spreadtheword > _______________________________________________ > Oorexx-devel mailing list > Oorexx-devel@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/oorexx-devel > ------------------------------------------------------------------------------ This SF.net email is sponsored by: SourcForge Community SourceForge wants to tell your story. http://p.sf.net/sfu/sf-spreadtheword _______________________________________________ Oorexx-devel mailing list Oorexx-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/oorexx-devel
