Tests w/ Decimals More Than 8 Places Are Not Necessarily Invalid

[cmoesel]

The DecimalTenthStep, DecimalPosTenthStep, and DecimalNegTenthStep tests all provided literal decimals with more than 8 decimal places as input. They are all marked as invalid - meaning that engines should return an error. But the spec explicitly says that engines may support more precise decimals here:

Note that implementations must support at least the 28 digits of precision and 8 digits of scale, but may support more precise values.

As such, I don't think there is a single correct expected output; it depends on the engine. Perhaps these tests shouldn't be included at all?

[bryantaustin13]

The same applies to three more tests in ValueLiteralsAndSelectors.xml, which are also marked invalid="true":

• Decimal10Pow28: 10000000000000000000000000000.00000000
• DecimalPos10Pow28: +10000000000000000000000000000.00000000
• DecimalNeg10Pow28: -10000000000000000000000000000.00000000

Each has 29 integer digits, so they exceed the 28 digits of precision where the TenthStep tests exceed the 8 digits of scale. The sentence quoted above covers both ("must support at least the 28 digits of precision and 8 digits of scale, but may support more precise values"). An engine that accepts them is conformant: the Java engine returns 1e+28 and -1e+28. So whatever is decided for the TenthStep tests should probably apply to these three as well.