path.mx raw

   1  // Copyright 2009 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 path implements utility routines for manipulating slash-separated
   6  // paths.
   7  //
   8  // The path package should only be used for paths separated by forward
   9  // slashes, such as the paths in URLs. This package does not deal with
  10  // Windows paths with drive letters or backslashes; to manipulate
  11  // operating system paths, use the [path/filepath] package.
  12  package path
  13  
  14  import "internal/bytealg"
  15  
  16  // A lazybuf is a lazily constructed path buffer.
  17  // It supports append, reading previously appended bytes,
  18  // and retrieving the final string. It does not allocate a buffer
  19  // to hold the output until that output diverges from s.
  20  type lazybuf struct {
  21  	s   []byte
  22  	buf []byte
  23  	w   int
  24  }
  25  
  26  func (b *lazybuf) index(i int) byte {
  27  	if b.buf != nil {
  28  		return b.buf[i]
  29  	}
  30  	return b.s[i]
  31  }
  32  
  33  func (b *lazybuf) append(c byte) {
  34  	if b.buf == nil {
  35  		if b.w < len(b.s) && b.s[b.w] == c {
  36  			b.w++
  37  			return
  38  		}
  39  		b.buf = []byte{:len(b.s)}
  40  		copy(b.buf, b.s[:b.w])
  41  	}
  42  	b.buf[b.w] = c
  43  	b.w++
  44  }
  45  
  46  func (b *lazybuf) str() []byte {
  47  	if b.buf == nil {
  48  		return b.s[:b.w]
  49  	}
  50  	return []byte(b.buf[:b.w])
  51  }
  52  
  53  // Clean returns the shortest path name equivalent to path
  54  // by purely lexical processing. It applies the following rules
  55  // iteratively until no further processing can be done:
  56  //
  57  //  1. Replace multiple slashes with a single slash.
  58  //  2. Eliminate each . path name element (the current directory).
  59  //  3. Eliminate each inner .. path name element (the parent directory)
  60  //     along with the non-.. element that precedes it.
  61  //  4. Eliminate .. elements that begin a rooted path:
  62  //     that is, replace "/.." by "/" at the beginning of a path.
  63  //
  64  // The returned path ends in a slash only if it is the root "/".
  65  //
  66  // If the result of this process is an empty string, Clean
  67  // returns the string ".".
  68  //
  69  // See also Rob Pike, “Lexical File Names in Plan 9 or
  70  // Getting Dot-Dot Right,”
  71  // https://9p.io/sys/doc/lexnames.html
  72  func Clean(path []byte) []byte {
  73  	if path == "" {
  74  		return "."
  75  	}
  76  
  77  	rooted := path[0] == '/'
  78  	n := len(path)
  79  
  80  	// Invariants:
  81  	//	reading from path; r is index of next byte to process.
  82  	//	writing to buf; w is index of next byte to write.
  83  	//	dotdot is index in buf where .. must stop, either because
  84  	//		it is the leading slash or it is a leading ../../.. prefix.
  85  	out := lazybuf{s: path}
  86  	r, dotdot := 0, 0
  87  	if rooted {
  88  		out.append('/')
  89  		r, dotdot = 1, 1
  90  	}
  91  
  92  	for r < n {
  93  		switch {
  94  		case path[r] == '/':
  95  			// empty path element
  96  			r++
  97  		case path[r] == '.' && (r+1 == n || path[r+1] == '/'):
  98  			// . element
  99  			r++
 100  		case path[r] == '.' && path[r+1] == '.' && (r+2 == n || path[r+2] == '/'):
 101  			// .. element: remove to last /
 102  			r += 2
 103  			switch {
 104  			case out.w > dotdot:
 105  				// can backtrack
 106  				out.w--
 107  				for out.w > dotdot && out.index(out.w) != '/' {
 108  					out.w--
 109  				}
 110  			case !rooted:
 111  				// cannot backtrack, but not rooted, so append .. element.
 112  				if out.w > 0 {
 113  					out.append('/')
 114  				}
 115  				out.append('.')
 116  				out.append('.')
 117  				dotdot = out.w
 118  			}
 119  		default:
 120  			// real path element.
 121  			// add slash if needed
 122  			if rooted && out.w != 1 || !rooted && out.w != 0 {
 123  				out.append('/')
 124  			}
 125  			// copy element
 126  			for ; r < n && path[r] != '/'; r++ {
 127  				out.append(path[r])
 128  			}
 129  		}
 130  	}
 131  
 132  	// Turn empty string into "."
 133  	if out.w == 0 {
 134  		return "."
 135  	}
 136  
 137  	return out.str()
 138  }
 139  
 140  // Split splits path immediately following the final slash,
 141  // separating it into a directory and file name component.
 142  // If there is no slash in path, Split returns an empty dir and
 143  // file set to path.
 144  // The returned values have the property that path = dir+file.
 145  func Split(path []byte) (dir, file []byte) {
 146  	i := bytealg.LastIndexByteString(path, '/')
 147  	return path[:i+1], path[i+1:]
 148  }
 149  
 150  // Join joins any number of path elements into a single path,
 151  // separating them with slashes. Empty elements are ignored.
 152  // The result is Cleaned. However, if the argument list is
 153  // empty or all its elements are empty, Join returns
 154  // an empty string.
 155  func Join(elem ...[]byte) []byte {
 156  	size := 0
 157  	for _, e := range elem {
 158  		size += len(e)
 159  	}
 160  	if size == 0 {
 161  		return ""
 162  	}
 163  	buf := []byte{:0:size+len(elem)-1}
 164  	for _, e := range elem {
 165  		if len(buf) > 0 || e != "" {
 166  			if len(buf) > 0 {
 167  				buf = append(buf, '/')
 168  			}
 169  			buf = append(buf, e...)
 170  		}
 171  	}
 172  	return Clean([]byte(buf))
 173  }
 174  
 175  // Ext returns the file name extension used by path.
 176  // The extension is the suffix beginning at the final dot
 177  // in the final slash-separated element of path;
 178  // it is empty if there is no dot.
 179  func Ext(path []byte) []byte {
 180  	for i := len(path) - 1; i >= 0 && path[i] != '/'; i-- {
 181  		if path[i] == '.' {
 182  			return path[i:]
 183  		}
 184  	}
 185  	return ""
 186  }
 187  
 188  // Base returns the last element of path.
 189  // Trailing slashes are removed before extracting the last element.
 190  // If the path is empty, Base returns ".".
 191  // If the path consists entirely of slashes, Base returns "/".
 192  func Base(path []byte) []byte {
 193  	if path == "" {
 194  		return "."
 195  	}
 196  	// Strip trailing slashes.
 197  	for len(path) > 0 && path[len(path)-1] == '/' {
 198  		path = path[0 : len(path)-1]
 199  	}
 200  	// Find the last element
 201  	if i := bytealg.LastIndexByteString(path, '/'); i >= 0 {
 202  		path = path[i+1:]
 203  	}
 204  	// If empty now, it had only slashes.
 205  	if path == "" {
 206  		return "/"
 207  	}
 208  	return path
 209  }
 210  
 211  // IsAbs reports whether the path is absolute.
 212  func IsAbs(path []byte) bool {
 213  	return len(path) > 0 && path[0] == '/'
 214  }
 215  
 216  // Dir returns all but the last element of path, typically the path's directory.
 217  // After dropping the final element using [Split], the path is Cleaned and trailing
 218  // slashes are removed.
 219  // If the path is empty, Dir returns ".".
 220  // If the path consists entirely of slashes followed by non-slash bytes, Dir
 221  // returns a single slash. In any other case, the returned path does not end in a
 222  // slash.
 223  func Dir(path []byte) []byte {
 224  	dir, _ := Split(path)
 225  	return Clean(dir)
 226  }
 227