Add support for xsimd::widen on WebAssembly

Related to #1179

Author: serge-sans-paille

include/xsimd/arch/xsimd_wasm.hpp

@@ -1701,6 +1701,24 @@ namespace xsimd
             return wasm_f64x2_trunc(self);
         }
 
+        // widen
+        template <class A>
+        XSIMD_INLINE std::array<batch<double, A>, 2> widen(batch<float, A> const& x, requires_arch<wasm>) noexcept
+        {
+            return { batch<double, A>(wasm_f64x2_promote_low_f32x4(x)),
+                     batch<double, A>(wasm_f64x2_promote_low_f32x4(wasm_i32x4_shuffle(x, x, 2, 3, 0, 1))) };
+        }
+        template <class A>
+        XSIMD_INLINE std::array<batch<widen_t<uint8_t>, A>, 2> widen(batch<uint8_t, A> const& x, requires_arch<wasm>) noexcept
+        {
+            return { batch<widen_t<uint8_t>, A>(wasm_u16x8_extend_low_u8x16(x)), batch<widen_t<uint8_t>, A>(wasm_u16x8_extend_high_u8x16(x)) };
+        }
+        template <class A>
+        XSIMD_INLINE std::array<batch<widen_t<int8_t>, A>, 2> widen(batch<int8_t, A> const& x, requires_arch<wasm>) noexcept
+        {
+            return { batch<widen_t<int8_t>, A>(wasm_i16x8_extend_low_i8x16(x)), batch<widen_t<int8_t>, A>(wasm_i16x8_extend_high_i8x16(x)) };
+        }
+
         // zip_hi
         template <class A>
         XSIMD_INLINE batch<float, A> zip_hi(batch<float, A> const& self, batch<float, A> const& other, requires_arch<wasm>) noexcept