bugfix: floating point representaion without fractional part

[jgerhards]

closes #278

[ploxiln]

It may be appropriate to move the which-format-string logic into a couple separate lines. Otherwise, looks good to me.

[jgerhards]

I could change it if the maintainer wishes me to do so, however I don't deem it as necessary.

[hawicz]

This seems like it might be a bit fragile. I would think that checking the generated output for a '.' character would be more reliable, but if this works then let's go with it.

[jgerhards]

as specified in man snprintf, this is guaranteed to work, see last sentence:

 g, G   The  double argument is converted in style f or e (or F or E for
              G conversions).  The precision specifies the number of  signifi‐
              cant  digits.   If the precision is missing, 6 digits are given;
              if the precision is zero, it is treated as 1.  Style e  is  used
              if  the  exponent from its conversion is less than -4 or greater
              than or equal to the precision.  Trailing zeros are removed from
              the  fractional part of the result; a decimal point appears only
              if it is followed by at least one digit.