Single/Double/Half/Decimal need more test cases for parsing

[pgovind]

Follow up for https://github.com/dotnet/runtime/pull/47666/files#r569703812

Tests need to be added for at least:

1. Simple floating point number starting with a .. For ex: .234
2. Simple floating point number with a . as it's last digit.
3. Floating point number with a . at single/double/half/decimal.MaxDigitCount
4. Floating point number with a . after single/double/half/decimal.MaxDigitCount
5. More nuanced floating point numbers that involve trailing zeros, rounding and precision concerns.

Tagging subscribers to this area: @tannergooding, @pgovind
See info in area-owners.md if you want to be subscribed.

Issue Details

---

Follow up for https://github.com/dotnet/runtime/pull/47666/files#r569703812

Tests need to be added for at least:

1. Simple floating point number starting with a .. For ex: .234
2. Simple floating point number with a . as it's last digit.
3. Floating point number with a . at single/double/half/decimal.MaxDigitCount
4. Floating point number with a . after single/double/half/decimal.MaxDigitCount

Author:	pgovind
Assignees:	-
Labels:	area-System.Numerics, up-for-grabs
Milestone:	-

[ayousuf23]

Hey,
I haven't contributed before, but I would like to start. Do you think this is something up my alley? If so, where should I add the tests?

[danmoseley]

Welcome to the repo, @ayousuf23 ! @pgovind can point you to the right places.

[ayousuf23]

@danmoseley Thanks!

[pgovind]

Do you think this is something up my alley? If so, where should I add the tests?

Hello @ayousuf23 ! Welcome!

Test cases for Single can be added to SingleTests.cs. Double in DoubleTests.cs, Half in HalfTests.cs and Decimal in DecimalTests.cs. All of these files are in src\libraries\System.Runtime\tests\System\. I recommend adding the test cases under the Parse_Valid_TestData method in each file. Since this is your first time starting with the repo, looking at the "Workflow Instructions" will be useful and will help you ensure that the new tests pass!

Let me know if you get stuck or need help!

[ayousuf23]

Alright! Thanks for the help! @pgovind can you assign this issue to me?

[ayousuf23]

I am getting this error:
System.Tests.DoubleTests.Parse_Span_Valid(value: "-123456789012345678.", offset: 0, count: 20, style: AllowThousands | Float, provider: null, expected: -1.2345678901234566E+17) [FAIL] Assert.Equal() Failure Expected: -1.2345678901234566E+17 Actual: -1.2345678901234568E+17 Stack Trace: C:\Users\abdullah\Documents\runtime\src\libraries\System.Runtime\tests\System\DoubleTests.cs(615,0): at System.Tests.DoubleTests.Parse_Span_Valid(String value, Int32 offset, Int32 count, NumberStyles style, IFormatProvider provider, Double expected) System.Tests.DoubleTests.Parse(value: "-123456789012345678.", style: AllowThousands | Float, provider: null, expected: -1.2345678901234566E+17) [FAIL] Assert.Equal() Failure Expected: -1.2345678901234566E+17 Actual: -1.2345678901234568E+17
from this line I added:
yield return new object[] { "-123456789012345678.", defaultStyle, null, -1.2345678901234567E17 };

Is the difference between the doubles significant?

[ayousuf23]

@pgovind I got an assert failure error when trying to test a double with . at the 19 digit place. It appears the double parsed from the string is different than the double I expected (the difference is in the 17th digit).

[pgovind]

System.Tests.DoubleTests.Parse_Span_Valid(value: "-123456789012345678.", offset: 0, count: 20, style: AllowThousands | Float, provider: null, expected: -1.2345678901234566E+17) [FAIL] Assert.Equal() Failure Expected: -1.2345678901234566E+17 Actual: -1.2345678901234568E+17

I believe 1.2345678901234568E17 is the correct value here so it looks like the Expected value should be modified in the unit test?