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  	"os"
    10  	"runtime"
    11  	"sort"
    12  	"strings"
    13  	"unicode/utf8"
    14  )
    15  
    16  // ErrBadPattern indicates a pattern was malformed.
    17  var ErrBadPattern = errors.New("syntax error in pattern")
    18  
    19  // Match reports whether name matches the shell file name pattern.
    20  // The pattern syntax is:
    21  //
    22  //	pattern:
    23  //		{ term }
    24  //	term:
    25  //		'*'         matches any sequence of non-Separator characters
    26  //		'?'         matches any single non-Separator character
    27  //		'[' [ '^' ] { character-range } ']'
    28  //		            character class (must be non-empty)
    29  //		c           matches character c (c != '*', '?', '\\', '[')
    30  //		'\\' c      matches character c
    31  //
    32  //	character-range:
    33  //		c           matches character c (c != '\\', '-', ']')
    34  //		'\\' c      matches character c
    35  //		lo '-' hi   matches character c for lo <= c <= hi
    36  //
    37  // Match requires pattern to match all of name, not just a substring.
    38  // The only possible returned error is ErrBadPattern, when pattern
    39  // is malformed.
    40  //
    41  // On Windows, escaping is disabled. Instead, '\\' is treated as
    42  // path separator.
    43  //
    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  	var i int
    98  Scan:
    99  	for i = 0; i < len(pattern); i++ {
   100  		switch pattern[i] {
   101  		case '\\':
   102  			if runtime.GOOS != "windows" {
   103  				// error check handled in matchChunk: bad pattern.
   104  				if i+1 < len(pattern) {
   105  					i++
   106  				}
   107  			}
   108  		case '[':
   109  			inrange = true
   110  		case ']':
   111  			inrange = false
   112  		case '*':
   113  			if !inrange {
   114  				break Scan
   115  			}
   116  		}
   117  	}
   118  	return star, pattern[0:i], pattern[i:]
   119  }
   120  
   121  // matchChunk checks whether chunk matches the beginning of s.
   122  // If so, it returns the remainder of s (after the match).
   123  // Chunk is all single-character operators: literals, char classes, and ?.
   124  func matchChunk(chunk, s string) (rest string, ok bool, err error) {
   125  	// failed records whether the match has failed.
   126  	// After the match fails, the loop continues on processing chunk,
   127  	// checking that the pattern is well-formed but no longer reading s.
   128  	failed := false
   129  	for len(chunk) > 0 {
   130  		if !failed && len(s) == 0 {
   131  			failed = true
   132  		}
   133  		switch chunk[0] {
   134  		case '[':
   135  			// character class
   136  			var r rune
   137  			if !failed {
   138  				var n int
   139  				r, n = utf8.DecodeRuneInString(s)
   140  				s = s[n:]
   141  			}
   142  			chunk = chunk[1:]
   143  			// possibly negated
   144  			negated := false
   145  			if len(chunk) > 0 && chunk[0] == '^' {
   146  				negated = true
   147  				chunk = chunk[1:]
   148  			}
   149  			// parse all ranges
   150  			match := false
   151  			nrange := 0
   152  			for {
   153  				if len(chunk) > 0 && chunk[0] == ']' && nrange > 0 {
   154  					chunk = chunk[1:]
   155  					break
   156  				}
   157  				var lo, hi rune
   158  				if lo, chunk, err = getEsc(chunk); err != nil {
   159  					return "", false, err
   160  				}
   161  				hi = lo
   162  				if chunk[0] == '-' {
   163  					if hi, chunk, err = getEsc(chunk[1:]); err != nil {
   164  						return "", false, err
   165  					}
   166  				}
   167  				if lo <= r && r <= hi {
   168  					match = true
   169  				}
   170  				nrange++
   171  			}
   172  			if match == negated {
   173  				failed = true
   174  			}
   175  
   176  		case '?':
   177  			if !failed {
   178  				if s[0] == Separator {
   179  					failed = true
   180  				}
   181  				_, n := utf8.DecodeRuneInString(s)
   182  				s = s[n:]
   183  			}
   184  			chunk = chunk[1:]
   185  
   186  		case '\\':
   187  			if runtime.GOOS != "windows" {
   188  				chunk = chunk[1:]
   189  				if len(chunk) == 0 {
   190  					return "", false, ErrBadPattern
   191  				}
   192  			}
   193  			fallthrough
   194  
   195  		default:
   196  			if !failed {
   197  				if chunk[0] != s[0] {
   198  					failed = true
   199  				}
   200  				s = s[1:]
   201  			}
   202  			chunk = chunk[1:]
   203  		}
   204  	}
   205  	if failed {
   206  		return "", false, nil
   207  	}
   208  	return s, true, nil
   209  }
   210  
   211  // getEsc gets a possibly-escaped character from chunk, for a character class.
   212  func getEsc(chunk string) (r rune, nchunk string, err error) {
   213  	if len(chunk) == 0 || chunk[0] == '-' || chunk[0] == ']' {
   214  		err = ErrBadPattern
   215  		return
   216  	}
   217  	if chunk[0] == '\\' && runtime.GOOS != "windows" {
   218  		chunk = chunk[1:]
   219  		if len(chunk) == 0 {
   220  			err = ErrBadPattern
   221  			return
   222  		}
   223  	}
   224  	r, n := utf8.DecodeRuneInString(chunk)
   225  	if r == utf8.RuneError && n == 1 {
   226  		err = ErrBadPattern
   227  	}
   228  	nchunk = chunk[n:]
   229  	if len(nchunk) == 0 {
   230  		err = ErrBadPattern
   231  	}
   232  	return
   233  }
   234  
   235  // Glob returns the names of all files matching pattern or nil
   236  // if there is no matching file. The syntax of patterns is the same
   237  // as in Match. The pattern may describe hierarchical names such as
   238  // /usr/*/bin/ed (assuming the Separator is '/').
   239  //
   240  // Glob ignores file system errors such as I/O errors reading directories.
   241  // The only possible returned error is ErrBadPattern, when pattern
   242  // is malformed.
   243  func Glob(pattern string) (matches []string, err error) {
   244  	// Check pattern is well-formed.
   245  	if _, err := Match(pattern, ""); err != nil {
   246  		return nil, err
   247  	}
   248  	if !hasMeta(pattern) {
   249  		if _, err = os.Lstat(pattern); err != nil {
   250  			return nil, nil
   251  		}
   252  		return []string{pattern}, nil
   253  	}
   254  
   255  	dir, file := Split(pattern)
   256  	volumeLen := 0
   257  	if runtime.GOOS == "windows" {
   258  		volumeLen, dir = cleanGlobPathWindows(dir)
   259  	} else {
   260  		dir = cleanGlobPath(dir)
   261  	}
   262  
   263  	if !hasMeta(dir[volumeLen:]) {
   264  		return glob(dir, file, nil)
   265  	}
   266  
   267  	// Prevent infinite recursion. See issue 15879.
   268  	if dir == pattern {
   269  		return nil, ErrBadPattern
   270  	}
   271  
   272  	var m []string
   273  	m, err = Glob(dir)
   274  	if err != nil {
   275  		return
   276  	}
   277  	for _, d := range m {
   278  		matches, err = glob(d, file, matches)
   279  		if err != nil {
   280  			return
   281  		}
   282  	}
   283  	return
   284  }
   285  
   286  // cleanGlobPath prepares path for glob matching.
   287  func cleanGlobPath(path string) string {
   288  	switch path {
   289  	case "":
   290  		return "."
   291  	case string(Separator):
   292  		// do nothing to the path
   293  		return path
   294  	default:
   295  		return path[0 : len(path)-1] // chop off trailing separator
   296  	}
   297  }
   298  
   299  // cleanGlobPathWindows is windows version of cleanGlobPath.
   300  func cleanGlobPathWindows(path string) (prefixLen int, cleaned string) {
   301  	vollen := volumeNameLen(path)
   302  	switch {
   303  	case path == "":
   304  		return 0, "."
   305  	case vollen+1 == len(path) && os.IsPathSeparator(path[len(path)-1]): // /, \, C:\ and C:/
   306  		// do nothing to the path
   307  		return vollen + 1, path
   308  	case vollen == len(path) && len(path) == 2: // C:
   309  		return vollen, path + "." // convert C: into C:.
   310  	default:
   311  		if vollen >= len(path) {
   312  			vollen = len(path) - 1
   313  		}
   314  		return vollen, path[0 : len(path)-1] // chop off trailing separator
   315  	}
   316  }
   317  
   318  // glob searches for files matching pattern in the directory dir
   319  // and appends them to matches. If the directory cannot be
   320  // opened, it returns the existing matches. New matches are
   321  // added in lexicographical order.
   322  func glob(dir, pattern string, matches []string) (m []string, e error) {
   323  	m = matches
   324  	fi, err := os.Stat(dir)
   325  	if err != nil {
   326  		return // ignore I/O error
   327  	}
   328  	if !fi.IsDir() {
   329  		return // ignore I/O error
   330  	}
   331  	d, err := os.Open(dir)
   332  	if err != nil {
   333  		return // ignore I/O error
   334  	}
   335  	defer d.Close()
   336  
   337  	names, _ := d.Readdirnames(-1)
   338  	sort.Strings(names)
   339  
   340  	for _, n := range names {
   341  		matched, err := Match(pattern, n)
   342  		if err != nil {
   343  			return m, err
   344  		}
   345  		if matched {
   346  			m = append(m, Join(dir, n))
   347  		}
   348  	}
   349  	return
   350  }
   351  
   352  // hasMeta reports whether path contains any of the magic characters
   353  // recognized by Match.
   354  func hasMeta(path string) bool {
   355  	magicChars := `*?[`
   356  	if runtime.GOOS != "windows" {
   357  		magicChars = `*?[\`
   358  	}
   359  	return strings.ContainsAny(path, magicChars)
   360  }
   361
View as plain text