Optimize Cursor: skip strrpos on tab-free documents

src/Parser/Cursor.php

@@ -59,7 +59,7 @@ class Cursor
     /**
      * @param string $line The line being parsed (ASCII or UTF-8)
      */
-    public function __construct(string $line)
+    public function __construct(string $line, bool $lineCouldHaveTabs = true)
     {
         if (! \mb_check_encoding($line, 'UTF-8')) {
             throw new UnexpectedEncodingException('Unexpected encoding - UTF-8 or ASCII was expected');
@@ -68,7 +68,9 @@ public function __construct(string $line)
         $this->line            = $line;
         $this->length          = \mb_strlen($line, 'UTF-8') ?: 0;
         $this->isMultibyte     = $this->length !== \strlen($line);
-        $this->lastTabPosition = $this->isMultibyte ? \mb_strrpos($line, "\t", 0, 'UTF-8') : \strrpos($line, "\t");
+        $this->lastTabPosition = $lineCouldHaveTabs
+            ? ($this->isMultibyte ? \mb_strrpos($line, "\t", 0, 'UTF-8') : \strrpos($line, "\t"))
+            : false;
     }
 
     /**

src/Parser/MarkdownParser.php

@@ -53,6 +53,9 @@ final class MarkdownParser implements MarkdownParserInterface
     /** @psalm-readonly-allow-private-mutation */
     private Cursor $cursor;
 
+    /** @psalm-readonly-allow-private-mutation */
+    private bool $documentHasTabs = false;
+
     /**
      * @var array<int, BlockContinueParserInterface>
      *
@@ -96,6 +99,8 @@ public function parse(string $input): Document
         $this->environment->dispatch($preParsedEvent);
         $markdownInput = $preParsedEvent->getMarkdown();
 
+        $this->documentHasTabs = \str_contains($markdownInput->getContent(), "\t");
+
         foreach ($markdownInput->getLines() as $lineNumber => $line) {
             $this->lineNumber = $lineNumber;
             $this->parseLine($line);
@@ -119,7 +124,7 @@ private function parseLine(string $line): void
         // replace NUL characters for security
         $line = \str_replace("\0", "\u{FFFD}", $line);
 
-        $this->cursor = new Cursor($line);
+        $this->cursor = new Cursor($line, $this->documentHasTabs);
 
         $matches = $this->parseBlockContinuation();
         if ($matches === null) {

tests/benchmark/benchmark.php

@@ -368,7 +368,7 @@
     }
 };
 
-if (extension_loaded('xdebug')) {
+if (extension_loaded('xdebug') && getenv('XDEBUG_MODE') !== 'off') {
     fwrite(STDERR, 'The xdebug extension is loaded, this can significantly skew benchmarks. Disable it for accurate results. For xdebug 3, prefix your command with "XDEBUG_MODE=off"' . PHP_EOL . PHP_EOL);
 }
 

tests/benchmark/benchmark_parse.php

@@ -0,0 +1,128 @@
+#!/usr/bin/env php
+<?php
+
+/*
+ * This file is part of the league/commonmark package.
+ *
+ * (c) Colin O'Dell <colinodell@gmail.com>
+ *
+ * For the full copyright and license information, please view the LICENSE
+ * file that was distributed with this source code.
+ */
+
+/**
+ * Warm benchmark: measures only parsing, not converter initialization.
+ *
+ * Usage:
+ *   php tests/benchmark/benchmark_parse.php [--iterations N] [--warmup N] [--md file]
+ *
+ * Defaults: 100 iterations, 10 warmup rounds, all built-in corpus files.
+ */
+require_once __DIR__ . '/../../vendor/autoload.php';
+
+use League\CommonMark\CommonMarkConverter;
+
+if (extension_loaded('xdebug') && getenv('XDEBUG_MODE') !== 'off') {
+    fwrite(STDERR, 'Xdebug is active — results will be skewed. Use XDEBUG_MODE=off.' . PHP_EOL . PHP_EOL);
+}
+
+// --- Argument parsing ---
+
+$iterations = 100;
+$warmup     = 10;
+$files      = [];
+
+for ($i = 1; $i < $argc; $i++) {
+    switch ($argv[$i]) {
+        case '--iterations':
+            $iterations = (int) ($argv[++$i] ?? $iterations);
+            break;
+        case '--warmup':
+            $warmup = (int) ($argv[++$i] ?? $warmup);
+            break;
+        case '--md':
+            $files[] = $argv[++$i] ?? '';
+            break;
+        default:
+            fwrite(STDERR, "Unknown option: {$argv[$i]}" . PHP_EOL);
+            fwrite(STDERR, 'Usage: benchmark_parse.php [--iterations N] [--warmup N] [--md file]' . PHP_EOL);
+            exit(1);
+    }
+}
+
+if ($files === []) {
+    $files = [
+        __DIR__ . '/sample.md',
+    ];
+}
+
+// --- Benchmark runner ---
+
+/**
+ * @return array{min: float, median: float, p95: float, mean: float}
+ */
+function bench(callable $fn, int $warmup, int $iterations): array
+{
+    for ($i = 0; $i < $warmup; $i++) {
+        $fn();
+    }
+
+    $times = [];
+    for ($i = 0; $i < $iterations; $i++) {
+        $t0      = \hrtime(true);
+        $fn();
+        $times[] = (\hrtime(true) - $t0) / 1e6; // ms
+    }
+
+    \sort($times);
+
+    $count  = \count($times);
+    $mean   = \array_sum($times) / $count;
+    $median = $count % 2 === 0
+        ? ($times[$count / 2 - 1] + $times[$count / 2]) / 2
+        : $times[(int) ($count / 2)];
+    $p95    = $times[(int) \ceil($count * 0.95) - 1];
+
+    return ['min' => $times[0], 'median' => $median, 'p95' => $p95, 'mean' => $mean];
+}
+
+// --- Run ---
+
+printf(
+    'Warm benchmark — %d iterations (+%d warmup)%s%s',
+    $iterations,
+    $warmup,
+    PHP_EOL,
+    PHP_EOL
+);
+printf("%-52s %8s %8s %8s %8s%s", 'File', 'min', 'median', 'p95', 'mean', PHP_EOL);
+printf("%s%s", \str_repeat('-', 88), PHP_EOL);
+
+$converter = new CommonMarkConverter();
+
+foreach ($files as $file) {
+    if (! \is_file($file)) {
+        fwrite(STDERR, "File not found: {$file}" . PHP_EOL);
+        continue;
+    }
+
+    $markdown = \file_get_contents($file);
+    $size     = \round(\strlen($markdown) / 1024, 1);
+    $label    = \basename($file) . " ({$size} KB)";
+
+    $stats = bench(
+        static fn () => $converter->convert($markdown),
+        $warmup,
+        $iterations
+    );
+
+    printf(
+        "%-52s %7.2f ms %7.2f ms %7.2f ms %7.2f ms%s",
+        $label,
+        $stats['min'],
+        $stats['median'],
+        $stats['p95'],
+        $stats['mean'],
+        PHP_EOL
+    );
+}

tests/unit/Parser/CursorTest.php

@@ -32,6 +32,22 @@ public function testConstructorWithInvalidUTF8(): void
         new Cursor(\hex2bin('A5A5A5'));
     }
 
+    public function testConstructorWithLineCouldHaveTabsFalse(): void
+    {
+        // When the caller guarantees no tabs are present, advanceBy must still
+        // track position and column correctly via the tab-free fast path.
+        $cursor = new Cursor('Hello world', false);
+        $this->assertSame('Hello world', $cursor->getLine());
+
+        $cursor->advanceBy(5);
+        $this->assertSame(5, $cursor->getPosition());
+        $this->assertSame(5, $cursor->getColumn());
+
+        $cursor->advanceBy(3);
+        $this->assertSame(8, $cursor->getPosition());
+        $this->assertSame(8, $cursor->getColumn());
+    }
+
     /**
      * @dataProvider dataForTestingNextNonSpaceMethods
      */