`-infinity` mistakenly listed twice in FloatingPoint theory description

[RyanGlScott]

The descriptions for fp.to_real, fp.to_ubv, and fp.to_sbv all describe what argument values for which they return unspecified values. However, all of their descriptions mistakenly list -infinity twice:

SMT-LIB-2/Theories/FloatingPoint.smt2

Lines 437 to 438 in 4191af1

	[[fp.to_real]](x) is the real number represented by x if x is not in
	{-infinity, -infinity, not_a_number}. Otherwise, it is unspecified.

SMT-LIB-2/Theories/FloatingPoint.smt2

Lines 483 to 484 in 4191af1

	[[(_ fp.to_ubv m)]](r, x) is unspecified in all other cases (including when
	x is in {-infinity, -infinity, not_a_number}).

SMT-LIB-2/Theories/FloatingPoint.smt2

Lines 493 to 494 in 4191af1

	[[(_ fp.to_sbv m)]](r, x) is unspecified in all other cases (including when
	x is in {-infinity, -infinity, not_a_number}).

I think the intention here is to list both positive and negative infinity (i.e., +infinity, -infinity), like the description for to_fp does:

SMT-LIB-2/Theories/FloatingPoint.smt2

Line 443 in 4191af1

[[(_ to_fp eb sb)]](r, x) = x if x in {+infinity, -infinity, not_a_number}.