fix issue #46665, prod(::Array{BigInt}) (#46667)

The way we were counting the number of bits was assigning a negative number
to `0`, which could lead to a negative total number of bits. Better to just
exit early in this case.
Also, the estimate was slightly off because we were counting the number of leading
zeros in the least significant limb, instead of the most significant.

(cherry picked from commit f7b4ebe)

base/gmp.jl

@@ -667,8 +667,12 @@ function prod(arr::AbstractArray{BigInt})
     # to account for the rounding to limbs in MPZ.mul!
     # (BITS_PER_LIMB-1 would typically be enough, to which we add
     # 1 for the initial multiplication by init=1 in foldl)
-    nbits = GC.@preserve arr sum(arr; init=BITS_PER_LIMB) do x
-        abs(x.size) * BITS_PER_LIMB - leading_zeros(unsafe_load(x.d))
+    nbits = BITS_PER_LIMB
+    for x in arr
+        iszero(x) && return zero(BigInt)
+        xsize = abs(x.size)
+        lz = GC.@preserve x leading_zeros(unsafe_load(x.d, xsize))
+        nbits += xsize * BITS_PER_LIMB - lz
     end
     init = BigInt(; nbits)
     MPZ.set_si!(init, 1)

test/gmp.jl

@@ -227,6 +227,7 @@ let a, b
     @test 0 == sum(BigInt[]) isa BigInt
     @test prod(b) == foldl(*, b)
     @test 1 == prod(BigInt[]) isa BigInt
+    @test prod(BigInt[0, 0, 0]) == 0 # issue #46665
 end
 
 @testset "Iterated arithmetic" begin