From 83f9fd3adff42a97e83e491632a624bffcf6c0a6 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Michael=20Vo=C5=99=C3=AD=C5=A1ek?= <mvorisek@mvorisek.cz>
Date: Tue, 13 Apr 2021 17:30:56 +0200
Subject: [PATCH] Fix/Rewrite .gitignore regex builder

---
 src/Symfony/Component/Finder/Gitignore.php    | 132 +++-----
 .../Component/Finder/Tests/GitignoreTest.php  | 297 ++++++++++++++----
 2 files changed, 287 insertions(+), 142 deletions(-)

diff --git a/src/Symfony/Component/Finder/Gitignore.php b/src/Symfony/Component/Finder/Gitignore.php
index dfe0a0a26d..5bd6829491 100644
--- a/src/Symfony/Component/Finder/Gitignore.php
+++ b/src/Symfony/Component/Finder/Gitignore.php
@@ -14,6 +14,7 @@ namespace Symfony\Component\Finder;
 /**
  * Gitignore matches against text.
  *
+ * @author Michael Voříšek <vorismi3@fel.cvut.cz>
  * @author Ahmed Abdou <mail@ahmd.io>
  */
 class Gitignore
@@ -21,113 +22,66 @@ class Gitignore
     /**
      * Returns a regexp which is the equivalent of the gitignore pattern.
      *
-     * @return string The regexp
+     * Format specification: https://git-scm.com/docs/gitignore#_pattern_format
      */
     public static function toRegex(string $gitignoreFileContent): string
     {
-        $gitignoreFileContent = preg_replace('/^[^\\\r\n]*#.*/m', '', $gitignoreFileContent);
-        $gitignoreLines = preg_split('/\r\n|\r|\n/', $gitignoreFileContent);
+        $gitignoreFileContent = preg_replace('~(?<!\\\\)#[^\n\r]*~', '', $gitignoreFileContent);
+        $gitignoreLines = preg_split('~\r\n?|\n~', $gitignoreFileContent);
 
-        $positives = [];
-        $negatives = [];
+        $res = self::lineToRegex('');
         foreach ($gitignoreLines as $i => $line) {
-            $line = trim($line);
-            if ('' === $line) {
-                continue;
+            $line = preg_replace('~(?<!\\\\)[ \t]+$~', '', $line);
+
+            if ('!' === substr($line, 0, 1)) {
+                $line = substr($line, 1);
+                $isNegative = true;
+            } else {
+                $isNegative = false;
             }
 
-            if (1 === preg_match('/^!/', $line)) {
-                $positives[$i] = null;
-                $negatives[$i] = self::getRegexFromGitignore(preg_replace('/^!(.*)/', '${1}', $line), true);
-
-                continue;
+            if ('' !== $line) {
+                if ($isNegative) {
+                    $res = '(?!'.self::lineToRegex($line).'$)'.$res;
+                } else {
+                    $res = '(?:'.$res.'|'.self::lineToRegex($line).')';
+                }
             }
-            $negatives[$i] = null;
-            $positives[$i] = self::getRegexFromGitignore($line);
         }
 
-        $index = 0;
-        $patterns = [];
-        foreach ($positives as $pattern) {
-            if (null === $pattern) {
-                continue;
-            }
-
-            $negativesAfter = array_filter(\array_slice($negatives, ++$index));
-            if ([] !== $negativesAfter) {
-                $pattern .= sprintf('(?<!%s)', implode('|', $negativesAfter));
-            }
-
-            $patterns[] = $pattern;
-        }
-
-        return sprintf('/^((%s))$/', implode(')|(', $patterns));
+        return '~^(?:'.$res.')~s';
     }
 
-    private static function getRegexFromGitignore(string $gitignorePattern, bool $negative = false): string
+    private static function lineToRegex(string $gitignoreLine): string
     {
-        $regex = '';
-        $isRelativePath = false;
-        // If there is a separator at the beginning or middle (or both) of the pattern, then the pattern is relative to the directory level of the particular .gitignore file itself
-        $slashPosition = strpos($gitignorePattern, '/');
-        if (false !== $slashPosition && \strlen($gitignorePattern) - 1 !== $slashPosition) {
-            if (0 === $slashPosition) {
-                $gitignorePattern = substr($gitignorePattern, 1);
-            }
-
-            $isRelativePath = true;
-            $regex .= '^';
+        if ('' === $gitignoreLine) {
+            return '$f'; // always false
         }
 
-        if ('/' === $gitignorePattern[\strlen($gitignorePattern) - 1]) {
-            $gitignorePattern = substr($gitignorePattern, 0, -1);
+        $slashPos = strpos($gitignoreLine, '/');
+        if (false !== $slashPos && \strlen($gitignoreLine) - 1 !== $slashPos) {
+            if (0 === $slashPos) {
+                $gitignoreLine = substr($gitignoreLine, 1);
+            }
+            $isAbsolute = true;
+        } else {
+            $isAbsolute = false;
         }
 
-        $iMax = \strlen($gitignorePattern);
-        for ($i = 0; $i < $iMax; ++$i) {
-            $tripleChars = substr($gitignorePattern, $i, 3);
-            if ('**/' === $tripleChars || '/**' === $tripleChars) {
-                $regex .= '.*';
-                $i += 2;
-                continue;
-            }
+        $parts = array_map(function (string $v): string {
+            $v = preg_quote(str_replace('\\', '', $v), '~');
+            $v = preg_replace_callback('~\\\\\[([^\[\]]*)\\\\\]~', function (array $matches): string {
+                return '['.str_replace('\\-', '-', $matches[1]).']';
+            }, $v);
+            $v = preg_replace('~\\\\\*\\\\\*~', '[^/]+(?:/[^/]+)*', $v);
+            $v = preg_replace('~\\\\\*~', '[^/]*', $v);
+            $v = preg_replace('~\\\\\?~', '[^/]', $v);
 
-            $doubleChars = substr($gitignorePattern, $i, 2);
-            if ('**' === $doubleChars) {
-                $regex .= '.*';
-                ++$i;
-                continue;
-            }
-            if ('*/' === $doubleChars) {
-                $regex .= '[^\/]*\/?[^\/]*';
-                ++$i;
-                continue;
-            }
+            return $v;
+        }, explode('/', $gitignoreLine));
 
-            $c = $gitignorePattern[$i];
-            switch ($c) {
-                case '*':
-                    $regex .= $isRelativePath ? '[^\/]*' : '[^\/]*\/?[^\/]*';
-                    break;
-                case '/':
-                case '.':
-                case ':':
-                case '(':
-                case ')':
-                case '{':
-                case '}':
-                    $regex .= '\\'.$c;
-                    break;
-                default:
-                    $regex .= $c;
-            }
-        }
-
-        if ($negative) {
-            // a lookbehind assertion has to be a fixed width (it can not have nested '|' statements)
-            return sprintf('%s$|%s\/$', $regex, $regex);
-        }
-
-        return '(?>'.$regex.'($|\/.*))';
+        return ($isAbsolute ? '' : '(?:[^/]+/)*')
+            .implode('/', $parts)
+            .('' !== end($parts) ? '(?:$|/)' : '');
     }
 }
diff --git a/src/Symfony/Component/Finder/Tests/GitignoreTest.php b/src/Symfony/Component/Finder/Tests/GitignoreTest.php
index f772bdb63f..01d9fc5f36 100644
--- a/src/Symfony/Component/Finder/Tests/GitignoreTest.php
+++ b/src/Symfony/Component/Finder/Tests/GitignoreTest.php
@@ -13,136 +13,327 @@ namespace Symfony\Component\Finder\Tests;
 use PHPUnit\Framework\TestCase;
 use Symfony\Component\Finder\Gitignore;
 
+/**
+ * @author Michael Voříšek <vorismi3@fel.cvut.cz>
+ */
 class GitignoreTest extends TestCase
 {
     /**
      * @dataProvider provider
+     * @dataProvider providerExtended
      */
-    public function testCases(string $patterns, array $matchingCases, array $nonMatchingCases)
+    public function testToRegex(array $gitignoreLines, array $matchingCases, array $nonMatchingCases)
     {
+        $patterns = implode("\n", $gitignoreLines);
+
         $regex = Gitignore::toRegex($patterns);
+        $this->assertSame($regex, Gitignore::toRegex(implode("\r\n", $gitignoreLines)));
+        $this->assertSame($regex, Gitignore::toRegex(implode("\r", $gitignoreLines)));
 
         foreach ($matchingCases as $matchingCase) {
-            $this->assertMatchesRegularExpression($regex, $matchingCase, sprintf('Failed asserting path [%s] matches gitignore patterns [%s] using regex [%s]', $matchingCase, $patterns, $regex));
+            $this->assertMatchesRegularExpression(
+                $regex,
+                $matchingCase,
+                sprintf(
+                    "Failed asserting path:\n%s\nmatches gitignore patterns:\n%s",
+                    preg_replace('~^~m', '    ', $matchingCase),
+                    preg_replace('~^~m', '    ', $patterns)
+                )
+            );
         }
 
         foreach ($nonMatchingCases as $nonMatchingCase) {
-            $this->assertDoesNotMatchRegularExpression($regex, $nonMatchingCase, sprintf('Failed asserting path [%s] not matching gitignore patterns [%s] using regex [%s]', $nonMatchingCase, $patterns, $regex));
+            $this->assertDoesNotMatchRegularExpression(
+                $regex,
+                $nonMatchingCase,
+                sprintf("Failed asserting path:\n%s\nNOT matching gitignore patterns:\n%s",
+                    preg_replace('~^~m', '    ', $nonMatchingCase),
+                    preg_replace('~^~m', '    ', $patterns)
+                )
+            );
         }
     }
 
-    /**
-     * @return array return is array of
-     *               [
-     *               [
-     *               '', // Git-ignore Pattern
-     *               [], // array of file paths matching
-     *               [], // array of file paths not matching
-     *               ],
-     *               ]
-     */
     public function provider(): array
     {
-        return [
+        $cases = [
             [
-                '
-                    *
-                    !/bin
-                    !/bin/bash
-                ',
+                [''],
+                [],
+                ['a', 'a/b', 'a/b/c', 'aa', 'm.txt', '.txt'],
+            ],
+            [
+                ['a', 'X'],
+                ['a', 'a/b', 'a/b/c', 'X', 'b/a', 'b/c/a', 'a/X', 'a/X/y', 'b/a/X/y'],
+                ['A', 'x', 'aa', 'm.txt', '.txt', 'aa/b', 'b/aa'],
+            ],
+            [
+                ['/a', 'x', 'd/'],
+                ['a', 'a/b', 'a/b/c', 'x', 'a/x', 'a/x/y', 'b/a/x/y', 'd/', 'd/u', 'e/d/', 'e/d/u'],
+                ['b/a', 'b/c/a', 'aa', 'm.txt', '.txt', 'aa/b', 'b/aa', 'e/d'],
+            ],
+            [
+                ['a/', 'x'],
+                ['a/b', 'a/b/c', 'x', 'a/x', 'a/x/y', 'b/a/x/y'],
+                ['a', 'b/a', 'b/c/a', 'aa', 'm.txt', '.txt', 'aa/b', 'b/aa'],
+            ],
+            [
+                ['*'],
+                ['a', 'a/b', 'a/b/c', 'aa', 'm.txt', '.txt'],
+                [],
+            ],
+            [
+                ['/*'],
+                ['a', 'a/b', 'a/b/c', 'aa', 'm.txt', '.txt'],
+                [],
+            ],
+            [
+                ['/a', 'm/*'],
+                ['a', 'a/b', 'a/b/c', 'm/'],
+                ['aa', 'm', 'b/m', 'b/m/'],
+            ],
+            [
+                ['a', '!x'],
+                ['a', 'a/b', 'a/b/c', 'b/a', 'b/c/a'],
+                ['x', 'aa', 'm.txt', '.txt', 'aa/b', 'b/aa'],
+            ],
+            [
+                ['a', '!a/', 'b', '!b/b'],
+                ['a', 'a/x', 'x/a', 'x/a/x', 'b', 'b'],
+                ['a/', 'x/a/', 'bb', 'b/b', 'bb'],
+            ],
+            [
+                ['[a-c]', 'x[C-E][][o]', 'g-h'],
+                ['a', 'b', 'c', 'xDo', 'g-h'],
+                ['A', 'xdo', 'u', 'g', 'h'],
+            ],
+            [
+                ['a?', '*/??b?'],
+                ['ax', 'x/xxbx'],
+                ['a', 'axy', 'xxax', 'x/xxax', 'x/y/xxax'],
+            ],
+            [
+                [' ', ' \ ', '  \  ', '/a ', '/b/c \ '],
+                ['  ', '   ', 'x/  ', 'x/   ', 'a', 'a/x', 'b/c  '],
+                [' ', '    ', 'x/ ', 'x/    ', 'a ', 'b/c   '],
+            ],
+            [
+                ['#', ' #', '/ #', '  #', '/  #', '  \ #', '   \  #', 'a #', 'a  #', 'a  \ #', 'a   \  #'],
+                ['   ', '    ', 'a', 'a   ', 'a    '],
+                [' ', '  ', 'a ', 'a  '],
+            ],
+            [
+                ["\t", "\t\\\t", " \t\\\t ", "\t#", "a\t#", "a\t\t#", "a \t#", "a\t\t\\\t#", "a \t\t\\\t\t#"],
+                ["\t\t", " \t\t", 'a', "a\t\t\t", "a \t\t\t"],
+                ["\t", "\t\t ", " \t\t ", "a\t", 'a ', "a \t", "a\t\t"],
+            ],
+            [
+                [' a', 'b ', '\ ', 'c\ '],
+                [' a', 'b', ' ', 'c '],
+                ['a', 'b ', 'c'],
+            ],
+            [
+                ['#a', '\#b', '\#/'],
+                ['#b', '#/'],
+                ['#a', 'a', 'b'],
+            ],
+            [
+                ['*', '!!', '!!*x', '\!!b'],
+                ['a', '!!', '!!b'],
+                ['!', '!x', '!xx'],
+            ],
+            [
+                [
+                    '*',
+                    '!/bin',
+                    '!/bin/bash',
+                ],
                 ['bin/cat', 'abc/bin/cat'],
                 ['bin/bash'],
             ],
             [
-                'fi#le.txt',
+                ['fi#le.txt'],
                 [],
                 ['#file.txt'],
             ],
             [
-                '
-                /bin/
-                /usr/local/
-                !/bin/bash
-                !/usr/local/bin/bash
-                ',
+                [
+                    '/bin/',
+                    '/usr/local/',
+                    '!/bin/bash',
+                    '!/usr/local/bin/bash',
+                ],
                 ['bin/cat'],
                 ['bin/bash'],
             ],
             [
-                '*.py[co]',
+                ['*.py[co]'],
                 ['file.pyc', 'file.pyc'],
                 ['filexpyc', 'file.pycx', 'file.py'],
             ],
             [
-                'dir1/**/dir2/',
+                ['dir1/**/dir2/'],
                 ['dir1/dirA/dir2/', 'dir1/dirA/dirB/dir2/'],
                 [],
             ],
             [
-                'dir1/*/dir2/',
+                ['dir1/*/dir2/'],
                 ['dir1/dirA/dir2/'],
                 ['dir1/dirA/dirB/dir2/'],
             ],
             [
-                '/*.php',
+                ['/*.php'],
                 ['file.php'],
                 ['app/file.php'],
             ],
             [
-                '\#file.txt',
+                ['\#file.txt'],
                 ['#file.txt'],
                 [],
             ],
             [
-                '*.php',
+                ['*.php'],
                 ['app/file.php', 'file.php'],
                 ['file.phps', 'file.phps', 'filephps'],
             ],
             [
-                'app/cache/',
+                ['app/cache/'],
                 ['app/cache/file.txt', 'app/cache/dir1/dir2/file.txt'],
                 ['a/app/cache/file.txt'],
             ],
             [
-                '
-                #IamComment
-                /app/cache/',
+                [
+                    '#IamComment',
+                    '/app/cache/',
+                ],
                 ['app/cache/file.txt', 'app/cache/subdir/ile.txt'],
                 ['a/app/cache/file.txt', '#IamComment', 'IamComment'],
             ],
             [
-                '
-                /app/cache/
-                #LastLineIsComment',
+                [
+                    '/app/cache/',
+                    '#LastLineIsComment',
+                ],
                 ['app/cache/file.txt', 'app/cache/subdir/ile.txt'],
                 ['a/app/cache/file.txt', '#LastLineIsComment', 'LastLineIsComment'],
             ],
             [
-                '
-                /app/cache/
-                \#file.txt
-                #LastLineIsComment',
+                [
+                    '/app/cache/',
+                    '\#file.txt',
+                    '#LastLineIsComment',
+                ],
                 ['app/cache/file.txt', 'app/cache/subdir/ile.txt', '#file.txt'],
                 ['a/app/cache/file.txt', '#LastLineIsComment', 'LastLineIsComment'],
             ],
             [
-                '
-                /app/cache/
-                \#file.txt
-                #IamComment
-                another_file.txt',
+                [
+                    '/app/cache/',
+                    '\#file.txt',
+                    '#IamComment',
+                    'another_file.txt',
+                ],
                 ['app/cache/file.txt', 'app/cache/subdir/ile.txt', '#file.txt', 'another_file.txt'],
                 ['a/app/cache/file.txt', 'IamComment', '#IamComment'],
             ],
             [
-                '
-                /app/**
-                !/app/bin
-                !/app/bin/test
-                ',
+                [
+                    '/app/**',
+                    '!/app/bin',
+                    '!/app/bin/test',
+                ],
                 ['app/test/file', 'app/bin/file'],
                 ['app/bin/test'],
             ],
+            [
+                [
+                    '/app/*/img',
+                    '!/app/*/img/src',
+                ],
+                ['app/a/img', 'app/a/img/x', 'app/a/img/src/x'],
+                ['app/a/img/src', 'app/a/img/src/'],
+            ],
+            [
+                [
+                    'app/**/img',
+                    '!/app/**/img/src',
+                ],
+                ['app/a/img', 'app/a/img/x', 'app/a/img/src/x', 'app/a/b/img', 'app/a/b/img/x', 'app/a/b/img/src/x', 'app/a/b/c/img'],
+                ['app/a/img/src', 'app/a/b/img/src', 'app/a/c/b/img/src'],
+            ],
+            [
+                [
+                    '/*',
+                    '!/foo',
+                    '/foo/*',
+                    '!/foo/bar',
+                ],
+                ['bar', 'foo/ba', 'foo/barx', 'x/foo/bar'],
+                ['foo', 'foo/bar'],
+            ],
+            [
+                [
+                    '/example/**',
+                    '!/example/example.txt',
+                    '!/example/packages',
+                ],
+                ['example/test', 'example/example.txt2', 'example/packages/foo.yaml'],
+                ['example/example.txt', 'example/packages', 'example/packages/'],
+            ],
         ];
+
+        return $cases;
+    }
+
+    public function providerExtended(): array
+    {
+        $basicCases = $this->provider();
+
+        $cases = [];
+        foreach ($basicCases as $case) {
+            $cases[] = [
+                array_merge(['never'], $case[0], ['!never']),
+                $case[1],
+                $case[2],
+            ];
+
+            $cases[] = [
+                array_merge(['!*'], $case[0]),
+                $case[1],
+                $case[2],
+            ];
+
+            $cases[] = [
+                array_merge(['*', '!*'], $case[0]),
+                $case[1],
+                $case[2],
+            ];
+
+            $cases[] = [
+                array_merge(['never', '**/never2', 'never3/**'], $case[0]),
+                $case[1],
+                $case[2],
+            ];
+
+            $cases[] = [
+                array_merge(['!never', '!**/never2', '!never3/**'], $case[0]),
+                $case[1],
+                $case[2],
+            ];
+
+            $lines = [];
+            for ($i = 0; $i < 30; ++$i) {
+                foreach ($case[0] as $line) {
+                    $lines[] = $line;
+                }
+            }
+            $cases[] = [
+                array_merge(['!never', '!**/never2', '!never3/**'], $lines),
+                $case[1],
+                $case[2],
+            ];
+        }
+
+        return $cases;
     }
 }