tcl-math(8.5): big-doubles? how do you do it?



The range of doubles goes up to about 1e308.
Still, expr {sqrt(10**600)} correctly gives me 1e+300.
How does it handle the too-large-for-double argument?

Just curious.
.