match.go

     1  // Copyright 2010 The Go Authors. All rights reserved.
     2  // Use of this source code is governed by a BSD-style
     3  // license that can be found in the LICENSE file.
     4  
     5  package filepath
     6  
     7  import (
     8  	"errors"
     9  	"internal/filepathlite"
    10  	"os"
    11  	"runtime"
    12  	"slices"
    13  	"strings"
    14  	"unicode/utf8"
    15  )
    16  
    17  // ErrBadPattern indicates a pattern was malformed.
    18  var ErrBadPattern = errors.New("syntax error in pattern")
    19  
    20  // Match reports whether name matches the shell file name pattern.
    21  // The pattern syntax is:
    22  //
    23  //	pattern:
    24  //		{ term }
    25  //	term:
    26  //		'*'         matches any sequence of non-Separator characters
    27  //		'?'         matches any single non-Separator character
    28  //		'[' [ '^' ] { character-range } ']'
    29  //		            character class (must be non-empty)
    30  //		c           matches character c (c != '*', '?', '\\', '[')
    31  //		'\\' c      matches character c
    32  //
    33  //	character-range:
    34  //		c           matches character c (c != '\\', '-', ']')
    35  //		'\\' c      matches character c
    36  //		lo '-' hi   matches character c for lo <= c <= hi
    37  //
    38  // Match requires pattern to match all of name, not just a substring.
    39  // The only possible returned error is [ErrBadPattern], when pattern
    40  // is malformed.
    41  //
    42  // On Windows, escaping is disabled. Instead, '\\' is treated as
    43  // path separator.
    44  func Match(pattern, name string) (matched bool, err error) {
    45  Pattern:
    46  	for len(pattern) > 0 {
    47  		var star bool
    48  		var chunk string
    49  		star, chunk, pattern = scanChunk(pattern)
    50  		if star && chunk == "" {
    51  			// Trailing * matches rest of string unless it has a /.
    52  			return !strings.Contains(name, string(Separator)), nil
    53  		}
    54  		// Look for match at current position.
    55  		t, ok, err := matchChunk(chunk, name)
    56  		// if we're the last chunk, make sure we've exhausted the name
    57  		// otherwise we'll give a false result even if we could still match
    58  		// using the star
    59  		if ok && (len(t) == 0 || len(pattern) > 0) {
    60  			name = t
    61  			continue
    62  		}
    63  		if err != nil {
    64  			return false, err
    65  		}
    66  		if star {
    67  			// Look for match skipping i+1 bytes.
    68  			// Cannot skip /.
    69  			for i := 0; i < len(name) && name[i] != Separator; i++ {
    70  				t, ok, err := matchChunk(chunk, name[i+1:])
    71  				if ok {
    72  					// if we're the last chunk, make sure we exhausted the name
    73  					if len(pattern) == 0 && len(t) > 0 {
    74  						continue
    75  					}
    76  					name = t
    77  					continue Pattern
    78  				}
    79  				if err != nil {
    80  					return false, err
    81  				}
    82  			}
    83  		}
    84  		return false, nil
    85  	}
    86  	return len(name) == 0, nil
    87  }
    88  
    89  // scanChunk gets the next segment of pattern, which is a non-star string
    90  // possibly preceded by a star.
    91  func scanChunk(pattern string) (star bool, chunk, rest string) {
    92  	for len(pattern) > 0 && pattern[0] == '*' {
    93  		pattern = pattern[1:]
    94  		star = true
    95  	}
    96  	inrange := false
    97  	for i := 0; i < len(pattern); i++ {
    98  		switch pattern[i] {
    99  		case '\\':
   100  			// error check handled in matchChunk: bad pattern.
   101  			if runtime.GOOS != "windows" && i+1 < len(pattern) {
   102  				i++
   103  			}
   104  		case '[':
   105  			inrange = true
   106  		case ']':
   107  			inrange = false
   108  		case '*':
   109  			if !inrange {
   110  				return star, pattern[:i], pattern[i:]
   111  			}
   112  		}
   113  	}
   114  	return star, pattern, ""
   115  }
   116  
   117  // matchChunk checks whether chunk matches the beginning of s.
   118  // If so, it returns the remainder of s (after the match).
   119  // Chunk is all single-character operators: literals, char classes, and ?.
   120  func matchChunk(chunk, s string) (rest string, ok bool, err error) {
   121  	// failed records whether the match has failed.
   122  	// After the match fails, the loop continues on processing chunk,
   123  	// checking that the pattern is well-formed but no longer reading s.
   124  	failed := false
   125  	for len(chunk) > 0 {
   126  		failed = failed || len(s) == 0
   127  		switch chunk[0] {
   128  		case '[':
   129  			// character class
   130  			var r rune
   131  			if !failed {
   132  				var n int
   133  				r, n = utf8.DecodeRuneInString(s)
   134  				s = s[n:]
   135  			}
   136  			chunk = chunk[1:]
   137  			// possibly negated
   138  			negated := false
   139  			if len(chunk) > 0 && chunk[0] == '^' {
   140  				negated = true
   141  				chunk = chunk[1:]
   142  			}
   143  			// parse all ranges
   144  			match := false
   145  			nrange := 0
   146  			for {
   147  				if len(chunk) > 0 && chunk[0] == ']' && nrange > 0 {
   148  					chunk = chunk[1:]
   149  					break
   150  				}
   151  				var lo, hi rune
   152  				if lo, chunk, err = getEsc(chunk); err != nil {
   153  					return "", false, err
   154  				}
   155  				hi = lo
   156  				if chunk[0] == '-' {
   157  					if hi, chunk, err = getEsc(chunk[1:]); err != nil {
   158  						return "", false, err
   159  					}
   160  				}
   161  				match = match || lo <= r && r <= hi
   162  				nrange++
   163  			}
   164  			failed = failed || match == negated
   165  
   166  		case '?':
   167  			if !failed {
   168  				failed = s[0] == Separator
   169  				_, n := utf8.DecodeRuneInString(s)
   170  				s = s[n:]
   171  			}
   172  			chunk = chunk[1:]
   173  
   174  		case '\\':
   175  			if runtime.GOOS != "windows" {
   176  				chunk = chunk[1:]
   177  				if len(chunk) == 0 {
   178  					return "", false, ErrBadPattern
   179  				}
   180  			}
   181  			fallthrough
   182  
   183  		default:
   184  			if !failed {
   185  				failed = chunk[0] != s[0]
   186  				s = s[1:]
   187  			}
   188  			chunk = chunk[1:]
   189  		}
   190  	}
   191  	if failed {
   192  		return "", false, nil
   193  	}
   194  	return s, true, nil
   195  }
   196  
   197  // getEsc gets a possibly-escaped character from chunk, for a character class.
   198  func getEsc(chunk string) (r rune, nchunk string, err error) {
   199  	if len(chunk) == 0 || chunk[0] == '-' || chunk[0] == ']' {
   200  		err = ErrBadPattern
   201  		return
   202  	}
   203  	if chunk[0] == '\\' && runtime.GOOS != "windows" {
   204  		chunk = chunk[1:]
   205  		if len(chunk) == 0 {
   206  			err = ErrBadPattern
   207  			return
   208  		}
   209  	}
   210  	r, n := utf8.DecodeRuneInString(chunk)
   211  	if r == utf8.RuneError && n == 1 {
   212  		err = ErrBadPattern
   213  	}
   214  	nchunk = chunk[n:]
   215  	if len(nchunk) == 0 {
   216  		err = ErrBadPattern
   217  	}
   218  	return
   219  }
   220  
   221  // Glob returns the names of all files matching pattern or nil
   222  // if there is no matching file. The syntax of patterns is the same
   223  // as in [Match]. The pattern may describe hierarchical names such as
   224  // /usr/*/bin/ed (assuming the [Separator] is '/').
   225  //
   226  // Glob ignores file system errors such as I/O errors reading directories.
   227  // The only possible returned error is [ErrBadPattern], when pattern
   228  // is malformed.
   229  func Glob(pattern string) (matches []string, err error) {
   230  	return globWithLimit(pattern, 0)
   231  }
   232  
   233  func globWithLimit(pattern string, depth int) (matches []string, err error) {
   234  	// This limit is used prevent stack exhaustion issues. See CVE-2022-30632.
   235  	const pathSeparatorsLimit = 10000
   236  	if depth == pathSeparatorsLimit {
   237  		return nil, ErrBadPattern
   238  	}
   239  
   240  	// Check pattern is well-formed.
   241  	if _, err := Match(pattern, ""); err != nil {
   242  		return nil, err
   243  	}
   244  	if !hasMeta(pattern) {
   245  		if _, err = os.Lstat(pattern); err != nil {
   246  			return nil, nil
   247  		}
   248  		return []string{pattern}, nil
   249  	}
   250  
   251  	dir, file := Split(pattern)
   252  	volumeLen := 0
   253  	if runtime.GOOS == "windows" {
   254  		volumeLen, dir = cleanGlobPathWindows(dir)
   255  	} else {
   256  		dir = cleanGlobPath(dir)
   257  	}
   258  
   259  	if !hasMeta(dir[volumeLen:]) {
   260  		return glob(dir, file, nil)
   261  	}
   262  
   263  	// Prevent infinite recursion. See issue 15879.
   264  	if dir == pattern {
   265  		return nil, ErrBadPattern
   266  	}
   267  
   268  	var m []string
   269  	m, err = globWithLimit(dir, depth+1)
   270  	if err != nil {
   271  		return
   272  	}
   273  	for _, d := range m {
   274  		matches, err = glob(d, file, matches)
   275  		if err != nil {
   276  			return
   277  		}
   278  	}
   279  	return
   280  }
   281  
   282  // cleanGlobPath prepares path for glob matching.
   283  func cleanGlobPath(path string) string {
   284  	switch path {
   285  	case "":
   286  		return "."
   287  	case string(Separator):
   288  		// do nothing to the path
   289  		return path
   290  	default:
   291  		return path[0 : len(path)-1] // chop off trailing separator
   292  	}
   293  }
   294  
   295  // cleanGlobPathWindows is windows version of cleanGlobPath.
   296  func cleanGlobPathWindows(path string) (prefixLen int, cleaned string) {
   297  	vollen := filepathlite.VolumeNameLen(path)
   298  	switch {
   299  	case path == "":
   300  		return 0, "."
   301  	case vollen+1 == len(path) && os.IsPathSeparator(path[len(path)-1]): // /, \, C:\ and C:/
   302  		// do nothing to the path
   303  		return vollen + 1, path
   304  	case vollen == len(path) && len(path) == 2: // C:
   305  		return vollen, path + "." // convert C: into C:.
   306  	default:
   307  		if vollen >= len(path) {
   308  			vollen = len(path) - 1
   309  		}
   310  		return vollen, path[0 : len(path)-1] // chop off trailing separator
   311  	}
   312  }
   313  
   314  // glob searches for files matching pattern in the directory dir
   315  // and appends them to matches. If the directory cannot be
   316  // opened, it returns the existing matches. New matches are
   317  // added in lexicographical order.
   318  func glob(dir, pattern string, matches []string) (m []string, e error) {
   319  	m = matches
   320  	fi, err := os.Stat(dir)
   321  	if err != nil {
   322  		return // ignore I/O error
   323  	}
   324  	if !fi.IsDir() {
   325  		return // ignore I/O error
   326  	}
   327  	d, err := os.Open(dir)
   328  	if err != nil {
   329  		return // ignore I/O error
   330  	}
   331  	defer d.Close()
   332  
   333  	names, _ := d.Readdirnames(-1)
   334  	slices.Sort(names)
   335  
   336  	for _, n := range names {
   337  		matched, err := Match(pattern, n)
   338  		if err != nil {
   339  			return m, err
   340  		}
   341  		if matched {
   342  			m = append(m, Join(dir, n))
   343  		}
   344  	}
   345  	return
   346  }
   347  
   348  // hasMeta reports whether path contains any of the magic characters
   349  // recognized by Match.
   350  func hasMeta(path string) bool {
   351  	magicChars := `*?[`
   352  	if runtime.GOOS != "windows" {
   353  		magicChars = `*?[\`
   354  	}
   355  	return strings.ContainsAny(path, magicChars)
   356  }
   357
View as plain text