Find Buffer search candidates with memchr

packages/compiler/test/ts7/baselines/order-parity/tests/corpus/7249-buffer-index-of-candidates.ts.json

@@ -0,0 +1,6 @@
+{
+ "order": [
+  "<repo>/tests/corpus/7249-buffer-index-of-candidates.ts"
+ ],
+ "diags": []
+}

packages/runtime/src/scr_bytes.c

@@ -3271,16 +3271,24 @@ double scr_bytes_index_of(const ScrBytes *b, const ScrBytes *needle, double off,
   }
   if (nlen == 0) return o > (double)len ? (double)len : o; /* '' matches at the clamp */
   if (nlen > len) return -1;
+  /* Candidates are the positions holding the needle's first byte; only
+   * those compare the rest. */
+  const uint8_t first = needle->data[0];
   if (fwd) {
     size_t start = o > (double)len ? len : (size_t)o;
-    for (size_t i = start; i + nlen <= len; i++) {
-      if (memcmp(b->data + i, needle->data, nlen) == 0) return (double)i;
+    const uint8_t *at = b->data + start, *end = b->data + (len - nlen) + 1;
+    while (at < end) {
+      at = memchr(at, first, (size_t)(end - at));
+      if (!at) return -1;
+      if (nlen == 1 || memcmp(at + 1, needle->data + 1, nlen - 1) == 0) return (double)(at - b->data);
+      at++;
     }
     return -1;
   }
   size_t start = o > (double)(len - nlen) ? len - nlen : (size_t)o;
   for (size_t i = start;; i--) {
-    if (memcmp(b->data + i, needle->data, nlen) == 0) return (double)i;
+    if (b->data[i] == first && (nlen == 1 || memcmp(b->data + i + 1, needle->data + 1, nlen - 1) == 0))
+      return (double)i;
     if (i == 0) break;
   }
   return -1;

tests/corpus/7249-buffer-index-of-candidates.ts

@@ -0,0 +1,27 @@
+// Buffer indexOf, lastIndexOf and includes with needles whose first byte
+// recurs: a candidate whose later bytes differ is skipped, matches at the
+// first and last positions are found, byteOffset clamps and counts from the
+// end, and the search reads only the buffer's own window of its memory.
+const text = Buffer.from("aab aaab aaaab ab\nab");
+console.log(text.indexOf("ab"), text.indexOf("aab"), text.indexOf("aaab"), text.indexOf("aaaab"), text.indexOf("aaaaab"));
+console.log(text.lastIndexOf("ab"), text.lastIndexOf("aab"), text.lastIndexOf("aaab"), text.lastIndexOf("b\na"));
+console.log(text.indexOf("ab", 2), text.indexOf("ab", -3), text.indexOf("ab", 100), text.indexOf("ab", -100));
+console.log(text.lastIndexOf("ab", 5), text.lastIndexOf("ab", -4), text.lastIndexOf("ab", 0), text.lastIndexOf("aa", 1));
+console.log(text.indexOf(10), text.indexOf(97, 3), text.lastIndexOf(98), text.lastIndexOf(97, -2), text.includes("b a"));
+console.log(text.indexOf(""), text.indexOf("", 4), text.indexOf("", 100), text.lastIndexOf(""), text.indexOf("x"));
+console.log(text.indexOf(Buffer.from("a\nab")), text.indexOf("ab\nab"), text.indexOf("ab\nabc"), text.lastIndexOf("aab", 4));
+
+// A window into a larger buffer never sees the bytes outside it.
+const whole = Buffer.from("xxneedlexx needle needlexx");
+const window = whole.subarray(3, 22);
+console.log(window.toString(), window.indexOf("needle"), window.lastIndexOf("needle"), window.indexOf("xx"), window.includes("xne"));
+console.log(window.indexOf("eedle", 1), window.lastIndexOf("eed"), window.indexOf("needlex"), window.lastIndexOf("n", 0));
+
+// A long search with many first-byte candidates.
+const long = Buffer.alloc(20000, "ab");
+long.write("abc", 19990);
+console.log(long.indexOf("abc"), long.lastIndexOf("aba"), long.indexOf("abc", 19991), long.lastIndexOf("abc"), long.indexOf("bb"));
+const lines: number[] = [];
+const fasta = Buffer.from(">ONE\nACGT\nTTGA\n>TWO\nGGCC\n");
+for (let at = fasta.indexOf(10); at !== -1; at = fasta.indexOf(10, at + 1)) lines.push(at);
+console.log(lines.join(","), fasta.indexOf(62, 1), fasta.lastIndexOf(62));

tests/harness/corpus-costs/7249-buffer-index-of-candidates.ts.json

@@ -0,0 +1 @@
+{ "plain": 2870, "san": 3656 }