fix (-Inf)^-1 inconsistency (#55875)

fixes #55831

(cherry picked from commit 2364719)

Author: oscardssmith

base/math.jl

@@ -1217,7 +1217,8 @@ end
 # this method is only reliable for -2^20 < n < 2^20 (cf. #53881 #53886)
 @assume_effects :terminates_locally @noinline function pow_body(x::Float64, n::Integer)
     y = 1.0
-    xnlo = ynlo = 0.0
+    xnlo = -0.0
+    ynlo = 0.0
     n == 3 && return x*x*x # keep compatibility with literal_pow
     if n < 0
         rx = inv(x)

test/math.jl

@@ -1505,7 +1505,19 @@ end
     @test E^n == Inf
     @test E^float(n) == Inf
 
-    # #55633
+    # issue #55831
+    @testset "literal pow zero sign" begin
+        @testset "T: $T" for T ∈ (Float16, Float32, Float64, BigFloat)
+            @testset "literal `-1`" begin
+                @test -0.0 === Float64(T(-Inf)^-1)
+            end
+            @testset "`Int(-1)`" begin
+                @test -0.0 === Float64(T(-Inf)^Int(-1))
+            end
+        end
+    end
+
+    # issue #55633
     struct Issue55633_1 <: Number end
     struct Issue55633_3 <: Number end
     struct Issue55633_9 <: Number end