The GNU library's float<->string routines are based on David Gay's. 
Therefore you can compare those to Gay's originals to see how much
effort was required to make them "mostly" portable, and can look at the
history of those to get some feel for the maintenance burden.

There's no difficulty in doing this correctly except for denormals.  The
pain is almost entirely in doing it correctly /fast/ -- and that is
indeed hard (which accounts for the extreme length and complexity of
Gay's code).