Fix Formatter.floatformat and Object.to_format (#1602)

* fix: change float format specifier in Object.to_format

Fix the float format specifier in Object.to_format. If there is a float
stored in a Object such as 3.14, it would be printed out as 3 because
the format specifier is %d but should be %g.

* fix: print nan in floatformat

Fix floatformat to print 'nan' if float is nan. Currently,
io::printn(float.nan) will produce 'inf' instead of 'nan'.

lib/std/collections/object.c3

@@ -67,7 +67,7 @@ fn usz! Object.to_format(&self, Formatter* formatter) @dynamic
 				case UNSIGNED_INT:
 					return formatter.printf("%d", (uint128)self.i)!;
 				case FLOAT:
-					return formatter.printf("%d", self.f)!;
+					return formatter.printf("%g", self.f)!;
 				case ENUM:
 					return formatter.printf("%d", self.i)!;
 				default:

lib/std/io/formatter_private.c3

@@ -214,7 +214,7 @@ fn usz! Formatter.floatformat(&self, FloatFormatting formatting, double y) @priv
 		// Add padding
 		if (!self.flags.left) len += self.pad(' ', self.width, 3 + pl)!;
 		String s = self.flags.uppercase ? "INF" : "inf";
-		if (y != y) s = self.flags.uppercase ? "NAN" : "nan";
+		if (math::is_nan(y)) s = self.flags.uppercase ? "NAN" : "nan";
 		len += s.len;
 		if (pl) len += self.out(is_neg ? '-' : '+')!;
 		len += self.out_chars(s)!;