Merge pull request ruby#887 from byroot/improve-consecutive-digits

parser.c: Optimize `json_parse_digits`

Author: byroot

ext/json/ext/parser/parser.c

@@ -612,12 +612,14 @@ json_eat_whitespace(JSON_ParserState *state)
                 while (rest(state) > 8) {
                     uint64_t chunk;
                     memcpy(&chunk, state->cursor, sizeof(uint64_t));
-                    size_t consecutive_spaces = trailing_zeros64(chunk ^ 0x2020202020202020) / CHAR_BIT;
+                    if (chunk == 0x2020202020202020) {
+                        state->cursor += 8;
+                        continue;
+                    }
 
+                    uint32_t consecutive_spaces = trailing_zeros64(chunk ^ 0x2020202020202020) / CHAR_BIT;
                     state->cursor += consecutive_spaces;
-                    if (consecutive_spaces != 8) {
-                        break;
-                    }
+                    break;
                 }
 #endif
                 break;
@@ -1101,13 +1103,21 @@ static inline int json_parse_digits(JSON_ParserState *state, uint64_t *accumulat
             continue;
         }
 
-        if ((match & 0xFFFFFFFF) == 0x33333333) { // 4 consecutive digits
+        uint32_t consecutive_digits = trailing_zeros64(match ^ 0x3333333333333333) / CHAR_BIT;
+
+        if (consecutive_digits >= 4) {
             *accumulator = (*accumulator * 10000) + decode_4digits_unrolled((uint32_t)next_8bytes);
             state->cursor += 4;
-            break;
+            consecutive_digits -= 4;
+        }
+
+        while (consecutive_digits) {
+            *accumulator = *accumulator * 10 + (*state->cursor - '0');
+            consecutive_digits--;
+            state->cursor++;
         }
 
-        break;
+        return (int)(state->cursor - start);
     }
 #endif
 

ext/json/ext/simd/simd.h

@@ -1,3 +1,7 @@
+#ifdef JSON_DEBUG
+#include <assert.h>
+#endif
+
 typedef enum {
     SIMD_NONE,
     SIMD_NEON,
@@ -18,6 +22,10 @@ typedef enum {
 
 static inline uint32_t trailing_zeros64(uint64_t input)
 {
+#ifdef JSON_DEBUG
+    assert(input > 0); // __builtin_ctz(0) is undefined behavior
+#endif
+
 #if HAVE_BUILTIN_CTZLL
     return __builtin_ctzll(input);
 #else
@@ -33,6 +41,10 @@ static inline uint32_t trailing_zeros64(uint64_t input)
 
 static inline int trailing_zeros(int input)
 {
+#ifdef JSON_DEBUG
+    assert(input > 0); // __builtin_ctz(0) is undefined behavior
+#endif
+
 #if HAVE_BUILTIN_CTZLL
     return __builtin_ctz(input);
 #else

test/json/json_parser_test.rb

@@ -804,6 +804,10 @@ def test_parse_leading_slash
     end
   end
 
+  def test_parse_whitespace_after_newline
+    assert_equal [], JSON.parse("[\n#{' ' * (8 + 8 + 4 + 3)}]")
+  end
+
   private
 
   def assert_equal_float(expected, actual, delta = 1e-2)