Fix build when if constexpr is unavailable

Author: onalante-ebay

include/xsimd/arch/xsimd_neon.hpp

@@ -3281,70 +3281,76 @@ namespace xsimd
         /********
          * mask *
          ********/
-        template <class A, class T>
-        XSIMD_INLINE uint64_t mask(batch_bool<T, A> const& self, requires_arch<neon>) noexcept
+        namespace detail
         {
 #ifdef XSIMD_LITTLE_ENDIAN
             static constexpr bool do_swap = false;
 #else
             static constexpr bool do_swap = true;
 #endif
+        }
+    
+        template <class A, class T, detail::enable_sized_t<T, 1> = 0>
+        XSIMD_INLINE uint64_t mask(batch_bool<T, A> const& self, requires_arch<neon>) noexcept
+        {
+            uint8x16_t inner = self;
+            XSIMD_IF_CONSTEXPR(detail::do_swap)
+            {
+                inner = vrev16q_u8(inner);
+            }
 
-            XSIMD_IF_CONSTEXPR(sizeof(T) == 1)
+            uint16x8_t pairs = vreinterpretq_u16_u8(inner);
+            uint8x8_t narrowed = vshrn_n_u16(pairs, 4);
+            XSIMD_IF_CONSTEXPR(detail::do_swap)
             {
-                uint8x16_t inner = self;
-                XSIMD_IF_CONSTEXPR(do_swap)
-                {
-                    inner = vrev16q_u8(inner);
-                }
+                narrowed = vrev64_u8(narrowed);
+            }
 
-                uint16x8_t pairs = vreinterpretq_u16_u8(inner);
-                uint8x8_t narrowed = vshrn_n_u16(pairs, 4);
-                XSIMD_IF_CONSTEXPR(do_swap)
-                {
-                    narrowed = vrev64_u8(narrowed);
-                }
+            uint64_t mask = vget_lane_u64(vreinterpret_u64_u8(narrowed), 0);
+            mask &= 0x1111111111111111;
+            mask = mask | mask >> 3;
+            mask = (mask | mask >> 6) & 0x000F000F000F000F;
+            mask = (mask | mask >> 12) & 0x000000FF000000FF;
+            return (mask | mask >> 24) & 0xFFFF;
+        }
 
-                uint64_t mask = vget_lane_u64(vreinterpret_u64_u8(narrowed), 0);
-                mask &= 0x1111111111111111;
-                mask = mask | mask >> 3;
-                mask = (mask | mask >> 6) & 0x000F000F000F000F;
-                mask = (mask | mask >> 12) & 0x000000FF000000FF;
-                return (mask | mask >> 24) & 0xFFFF;
-            }
-            else XSIMD_IF_CONSTEXPR(sizeof(T) == 2)
+        template <class A, class T, detail::enable_sized_t<T, 2> = 0>
+        XSIMD_INLINE uint64_t mask(batch_bool<T, A> const& self, requires_arch<neon>) noexcept
+        {
+            uint8x8_t narrowed = vmovn_u16(self);
+            XSIMD_IF_CONSTEXPR(detail::do_swap)
             {
-                uint8x8_t narrowed = vmovn_u16(self);
-                XSIMD_IF_CONSTEXPR(do_swap)
-                {
-                    narrowed = vrev64_u8(narrowed);
-                }
-
-                uint64_t mask = vget_lane_u64(vreinterpret_u64_u8(narrowed), 0);
-                mask &= 0x0101010101010101;
-                mask = mask | mask >> 7;
-                mask = mask | mask >> 14;
-                return (mask | mask >> 28) & 0xFF;
+                narrowed = vrev64_u8(narrowed);
             }
-            else XSIMD_IF_CONSTEXPR(sizeof(T) == 4)
-            {
-                uint16x4_t narrowed = vmovn_u32(self);
-                XSIMD_IF_CONSTEXPR(do_swap)
-                {
-                    narrowed = vrev64_u16(narrowed);
-                }
 
-                uint64_t mask = vget_lane_u64(vreinterpret_u64_u16(narrowed), 0);
-                mask &= 0x0001000100010001;
-                mask = mask | mask >> 15;
-                return (mask | mask >> 30) & 0xF;
-            }
-            else
+            uint64_t mask = vget_lane_u64(vreinterpret_u64_u8(narrowed), 0);
+            mask &= 0x0101010101010101;
+            mask = mask | mask >> 7;
+            mask = mask | mask >> 14;
+            return (mask | mask >> 28) & 0xFF;
+        }
+
+        template <class A, class T, detail::enable_sized_t<T, 4> = 0>
+        XSIMD_INLINE uint64_t mask(batch_bool<T, A> const& self, requires_arch<neon>) noexcept
+        {
+            uint16x4_t narrowed = vmovn_u32(self);
+            XSIMD_IF_CONSTEXPR(detail::do_swap)
             {
-                uint64_t mask_lo = vgetq_lane_u64(self, 0);
-                uint64_t mask_hi = vgetq_lane_u64(self, 1);
-                return ((mask_lo >> 63) | (mask_hi << 1)) & 0x3;
+                narrowed = vrev64_u16(narrowed);
             }
+
+            uint64_t mask = vget_lane_u64(vreinterpret_u64_u16(narrowed), 0);
+            mask &= 0x0001000100010001;
+            mask = mask | mask >> 15;
+            return (mask | mask >> 30) & 0xF;
+        }
+
+        template <class A, class T, detail::enable_sized_t<T, 8> = 0>
+        XSIMD_INLINE uint64_t mask(batch_bool<T, A> const& self, requires_arch<neon>) noexcept
+        {
+            uint64_t mask_lo = vgetq_lane_u64(self, 0);
+            uint64_t mask_hi = vgetq_lane_u64(self, 1);
+            return ((mask_lo >> 63) | (mask_hi << 1)) & 0x3;
         }
     }