Incorrect rounding of numbers in Flight Data display in flight

[Virindi-AC]

In the "Flight Data" display, sometimes the angle of attack number reads as ".10", when it is not actually .1. That is, the number will go from eg. 7.7, 7.8, 7.9, 7.10, 8.0, 8.1. In the tests it does smoothly run through those values so it is obvious that it is not actually "7.1" in that example.

1.2.2 Win64, FAR f38f1f36784a43a471ac583ab12ad15112de4eb7

Screenshot: http://www.virindi.net/junk/far_decimal.png

[Virindi-AC]

Hold on, my checked out code doesn't match what is showing on github. Jeez, this might be my bad.

[Virindi-AC]

Nope, this report is correct. Seems to be a problem with the stringbuilder extensions.

Ran a simple test, the following code: StringBuilder sb = new StringBuilder(); sb.Concat(7.99f,1); Console.WriteLine(sb.ToString());

produces the result "7.10"

[Virindi-AC]

The logic of this library code to cat a number to a stringbuilder is completely broken; it fails in many cases.

Some examples of incorrect output of Concat( this StringBuilder string_builder, float float_val, uint decimal_places, uint pad_amount, char pad_char ):

• -0.78, 2 digits -> "0.78" (negative sign missing)
• 0.006, 2 digits -> "0.60" (leading zeroes in decimal portion missing)
• -1, 2 digits -> "1.0" (trailing zero missing)
• NaN, 2 digits -> "-2147483648.0"
• PositiveInfinity, 2 digits -> "-2147483648.0"
• NegativeInfinity, 2 digits -> "-2147483648.0"

as well as numerous possible ways that rounding can be wrong due to floating point inaccuracy.

[ferram4]

Merged the PR, so this can close.