A glitch in FloatToStr

[jj2007]

Under certain conditions, FloatToStr erroneously returns 0.0xx as follows:

Code Select Expand

.data
r8n Real8 0.00999999
r8o Real8 0.0099999999
r8p Real8 0.009999999999
r8q Real8 0.00999999999999
r8r Real8 0.0099999999999999


Code Select Expand

0.009999990     r8n MasmBasic
9.999990e-003   r8n FloatToStr

0.010000000     r8o MasmBasic
0.000000e-003   r8o FloatToStr

0.010000000     r8p MasmBasic
0.000000e-003   r8p FloatToStr

0.010000000     r8q MasmBasic
0.000000e-003   r8q FloatToStr

0.010000000     r8r MasmBasic
0.000000e-003   r8r FloatToStr

FloatToStr2 is not affected by this rounding error. Example attached (for use with the MasmBasic library).