Documentation inaccuracy in sysutils/floattostrf.html
The documentation for FloatToStrF says:
ffFixed = Fixed point format. Value is converted to a string using fixed point notation. The result is composed of all digits of the integer part of Value, preceded by a minus sign if Value is negative. Following the integer part is DecimalSeparator and then the fractional part of Value, rounded off to Digits numbers. If the number is too large then the result will be in scientific notation.
This text suggests that the DecimalSeparator will always be included. However, it is omitted when using Digits=0.