@@ -4241,8 +4241,9 @@ Type Checking</h3>
42414241 * Any operation with at least one NaN argument produces NaN.
42424242
42434243 Additionally,
4244- IEEE-754 introduces the concept of "negative zero",
4245- which must be tracked within a calculation
4244+ IEEE-754 introduces the concept of a "signed zero",
4245+ which is specifically positive or negative,
4246+ and which must be tracked within a calculation
42464247 and between nested calculations:
42474248
42484249 * Negative zero
@@ -4275,13 +4276,16 @@ Type Checking</h3>
42754276 For example, ''min(0⁺, 0⁻)'' must produce 0⁻,
42764277 ''max(0⁺, 0⁻)'' must produce 0⁺,
42774278 and ''clamp(0⁺, 0⁻, 1)'' must produce 0⁺.
4279+ * Unless otherwise specified,
4280+ any other operation that produces a zero
4281+ produces positive zero (0⁺).
42784282
42794283 If a <dfn export>top-level calculation</dfn>
42804284 (a [=math function=] not nested inside of another [=math function=] )
42814285 would produce a value whose numeric part is NaN,
42824286 it instead act as though the numeric part is 0.
42834287 If a [=top-level calculation=]
4284- would produce a value whose numeric part is 0⁻,
4288+ would produce a value whose numeric part is 0⁻ or 0⁺ ,
42854289 it instead acts as though the numeric part is the standard "unsigned" zero.
42864290
42874291 <div class=example>
0 commit comments