Source file src/bytes/buffer.go

Documentation: bytes

     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 bytes
     6  
     7  // Simple byte buffer for marshaling data.
     8  
     9  import (
    10  	"errors"
    11  	"io"
    12  	"unicode/utf8"
    13  )
    14  
    15  // smallBufferSize is an initial allocation minimal capacity.
    16  const smallBufferSize = 64
    17  
    18  // A Buffer is a variable-sized buffer of bytes with Read and Write methods.
    19  // The zero value for Buffer is an empty buffer ready to use.
    20  type Buffer struct {
    21  	buf      []byte // contents are the bytes buf[off : len(buf)]
    22  	off      int    // read at &buf[off], write at &buf[len(buf)]
    23  	lastRead readOp // last read operation, so that Unread* can work correctly.
    24  
    25  	// FIXME: it would be advisable to align Buffer to cachelines to avoid false
    26  	// sharing.
    27  }
    28  
    29  // The readOp constants describe the last action performed on
    30  // the buffer, so that UnreadRune and UnreadByte can check for
    31  // invalid usage. opReadRuneX constants are chosen such that
    32  // converted to int they correspond to the rune size that was read.
    33  type readOp int8
    34  
    35  // Don't use iota for these, as the values need to correspond with the
    36  // names and comments, which is easier to see when being explicit.
    37  const (
    38  	opRead      readOp = -1 // Any other read operation.
    39  	opInvalid   readOp = 0  // Non-read operation.
    40  	opReadRune1 readOp = 1  // Read rune of size 1.
    41  	opReadRune2 readOp = 2  // Read rune of size 2.
    42  	opReadRune3 readOp = 3  // Read rune of size 3.
    43  	opReadRune4 readOp = 4  // Read rune of size 4.
    44  )
    45  
    46  // ErrTooLarge is passed to panic if memory cannot be allocated to store data in a buffer.
    47  var ErrTooLarge = errors.New("bytes.Buffer: too large")
    48  var errNegativeRead = errors.New("bytes.Buffer: reader returned negative count from Read")
    49  
    50  const maxInt = int(^uint(0) >> 1)
    51  
    52  // Bytes returns a slice of length b.Len() holding the unread portion of the buffer.
    53  // The slice is valid for use only until the next buffer modification (that is,
    54  // only until the next call to a method like Read, Write, Reset, or Truncate).
    55  // The slice aliases the buffer content at least until the next buffer modification,
    56  // so immediate changes to the slice will affect the result of future reads.
    57  func (b *Buffer) Bytes() []byte { return b.buf[b.off:] }
    58  
    59  // String returns the contents of the unread portion of the buffer
    60  // as a string. If the Buffer is a nil pointer, it returns "<nil>".
    61  //
    62  // To build strings more efficiently, see the strings.Builder type.
    63  func (b *Buffer) String() string {
    64  	if b == nil {
    65  		// Special case, useful in debugging.
    66  		return "<nil>"
    67  	}
    68  	return string(b.buf[b.off:])
    69  }
    70  
    71  // empty reports whether the unread portion of the buffer is empty.
    72  func (b *Buffer) empty() bool { return len(b.buf) <= b.off }
    73  
    74  // Len returns the number of bytes of the unread portion of the buffer;
    75  // b.Len() == len(b.Bytes()).
    76  func (b *Buffer) Len() int { return len(b.buf) - b.off }
    77  
    78  // Cap returns the capacity of the buffer's underlying byte slice, that is, the
    79  // total space allocated for the buffer's data.
    80  func (b *Buffer) Cap() int { return cap(b.buf) }
    81  
    82  // Truncate discards all but the first n unread bytes from the buffer
    83  // but continues to use the same allocated storage.
    84  // It panics if n is negative or greater than the length of the buffer.
    85  func (b *Buffer) Truncate(n int) {
    86  	if n == 0 {
    87  		b.Reset()
    88  		return
    89  	}
    90  	b.lastRead = opInvalid
    91  	if n < 0 || n > b.Len() {
    92  		panic("bytes.Buffer: truncation out of range")
    93  	}
    94  	b.buf = b.buf[:b.off+n]
    95  }
    96  
    97  // Reset resets the buffer to be empty,
    98  // but it retains the underlying storage for use by future writes.
    99  // Reset is the same as Truncate(0).
   100  func (b *Buffer) Reset() {
   101  	b.buf = b.buf[:0]
   102  	b.off = 0
   103  	b.lastRead = opInvalid
   104  }
   105  
   106  // tryGrowByReslice is a inlineable version of grow for the fast-case where the
   107  // internal buffer only needs to be resliced.
   108  // It returns the index where bytes should be written and whether it succeeded.
   109  func (b *Buffer) tryGrowByReslice(n int) (int, bool) {
   110  	if l := len(b.buf); n <= cap(b.buf)-l {
   111  		b.buf = b.buf[:l+n]
   112  		return l, true
   113  	}
   114  	return 0, false
   115  }
   116  
   117  // grow grows the buffer to guarantee space for n more bytes.
   118  // It returns the index where bytes should be written.
   119  // If the buffer can't grow it will panic with ErrTooLarge.
   120  func (b *Buffer) grow(n int) int {
   121  	m := b.Len()
   122  	// If buffer is empty, reset to recover space.
   123  	if m == 0 && b.off != 0 {
   124  		b.Reset()
   125  	}
   126  	// Try to grow by means of a reslice.
   127  	if i, ok := b.tryGrowByReslice(n); ok {
   128  		return i
   129  	}
   130  	if b.buf == nil && n <= smallBufferSize {
   131  		b.buf = make([]byte, n, smallBufferSize)
   132  		return 0
   133  	}
   134  	c := cap(b.buf)
   135  	if n <= c/2-m {
   136  		// We can slide things down instead of allocating a new
   137  		// slice. We only need m+n <= c to slide, but
   138  		// we instead let capacity get twice as large so we
   139  		// don't spend all our time copying.
   140  		copy(b.buf, b.buf[b.off:])
   141  	} else if c > maxInt-c-n {
   142  		panic(ErrTooLarge)
   143  	} else {
   144  		// Not enough space anywhere, we need to allocate.
   145  		buf := makeSlice(2*c + n)
   146  		copy(buf, b.buf[b.off:])
   147  		b.buf = buf
   148  	}
   149  	// Restore b.off and len(b.buf).
   150  	b.off = 0
   151  	b.buf = b.buf[:m+n]
   152  	return m
   153  }
   154  
   155  // Grow grows the buffer's capacity, if necessary, to guarantee space for
   156  // another n bytes. After Grow(n), at least n bytes can be written to the
   157  // buffer without another allocation.
   158  // If n is negative, Grow will panic.
   159  // If the buffer can't grow it will panic with ErrTooLarge.
   160  func (b *Buffer) Grow(n int) {
   161  	if n < 0 {
   162  		panic("bytes.Buffer.Grow: negative count")
   163  	}
   164  	m := b.grow(n)
   165  	b.buf = b.buf[:m]
   166  }
   167  
   168  // Write appends the contents of p to the buffer, growing the buffer as
   169  // needed. The return value n is the length of p; err is always nil. If the
   170  // buffer becomes too large, Write will panic with ErrTooLarge.
   171  func (b *Buffer) Write(p []byte) (n int, err error) {
   172  	b.lastRead = opInvalid
   173  	m, ok := b.tryGrowByReslice(len(p))
   174  	if !ok {
   175  		m = b.grow(len(p))
   176  	}
   177  	return copy(b.buf[m:], p), nil
   178  }
   179  
   180  // WriteString appends the contents of s to the buffer, growing the buffer as
   181  // needed. The return value n is the length of s; err is always nil. If the
   182  // buffer becomes too large, WriteString will panic with ErrTooLarge.
   183  func (b *Buffer) WriteString(s string) (n int, err error) {
   184  	b.lastRead = opInvalid
   185  	m, ok := b.tryGrowByReslice(len(s))
   186  	if !ok {
   187  		m = b.grow(len(s))
   188  	}
   189  	return copy(b.buf[m:], s), nil
   190  }
   191  
   192  // MinRead is the minimum slice size passed to a Read call by
   193  // Buffer.ReadFrom. As long as the Buffer has at least MinRead bytes beyond
   194  // what is required to hold the contents of r, ReadFrom will not grow the
   195  // underlying buffer.
   196  const MinRead = 512
   197  
   198  // ReadFrom reads data from r until EOF and appends it to the buffer, growing
   199  // the buffer as needed. The return value n is the number of bytes read. Any
   200  // error except io.EOF encountered during the read is also returned. If the
   201  // buffer becomes too large, ReadFrom will panic with ErrTooLarge.
   202  func (b *Buffer) ReadFrom(r io.Reader) (n int64, err error) {
   203  	b.lastRead = opInvalid
   204  	for {
   205  		i := b.grow(MinRead)
   206  		b.buf = b.buf[:i]
   207  		m, e := r.Read(b.buf[i:cap(b.buf)])
   208  		if m < 0 {
   209  			panic(errNegativeRead)
   210  		}
   211  
   212  		b.buf = b.buf[:i+m]
   213  		n += int64(m)
   214  		if e == io.EOF {
   215  			return n, nil // e is EOF, so return nil explicitly
   216  		}
   217  		if e != nil {
   218  			return n, e
   219  		}
   220  	}
   221  }
   222  
   223  // makeSlice allocates a slice of size n. If the allocation fails, it panics
   224  // with ErrTooLarge.
   225  func makeSlice(n int) []byte {
   226  	// If the make fails, give a known error.
   227  	defer func() {
   228  		if recover() != nil {
   229  			panic(ErrTooLarge)
   230  		}
   231  	}()
   232  	return make([]byte, n)
   233  }
   234  
   235  // WriteTo writes data to w until the buffer is drained or an error occurs.
   236  // The return value n is the number of bytes written; it always fits into an
   237  // int, but it is int64 to match the io.WriterTo interface. Any error
   238  // encountered during the write is also returned.
   239  func (b *Buffer) WriteTo(w io.Writer) (n int64, err error) {
   240  	b.lastRead = opInvalid
   241  	if nBytes := b.Len(); nBytes > 0 {
   242  		m, e := w.Write(b.buf[b.off:])
   243  		if m > nBytes {
   244  			panic("bytes.Buffer.WriteTo: invalid Write count")
   245  		}
   246  		b.off += m
   247  		n = int64(m)
   248  		if e != nil {
   249  			return n, e
   250  		}
   251  		// all bytes should have been written, by definition of
   252  		// Write method in io.Writer
   253  		if m != nBytes {
   254  			return n, io.ErrShortWrite
   255  		}
   256  	}
   257  	// Buffer is now empty; reset.
   258  	b.Reset()
   259  	return n, nil
   260  }
   261  
   262  // WriteByte appends the byte c to the buffer, growing the buffer as needed.
   263  // The returned error is always nil, but is included to match bufio.Writer's
   264  // WriteByte. If the buffer becomes too large, WriteByte will panic with
   265  // ErrTooLarge.
   266  func (b *Buffer) WriteByte(c byte) error {
   267  	b.lastRead = opInvalid
   268  	m, ok := b.tryGrowByReslice(1)
   269  	if !ok {
   270  		m = b.grow(1)
   271  	}
   272  	b.buf[m] = c
   273  	return nil
   274  }
   275  
   276  // WriteRune appends the UTF-8 encoding of Unicode code point r to the
   277  // buffer, returning its length and an error, which is always nil but is
   278  // included to match bufio.Writer's WriteRune. The buffer is grown as needed;
   279  // if it becomes too large, WriteRune will panic with ErrTooLarge.
   280  func (b *Buffer) WriteRune(r rune) (n int, err error) {
   281  	if r < utf8.RuneSelf {
   282  		b.WriteByte(byte(r))
   283  		return 1, nil
   284  	}
   285  	b.lastRead = opInvalid
   286  	m, ok := b.tryGrowByReslice(utf8.UTFMax)
   287  	if !ok {
   288  		m = b.grow(utf8.UTFMax)
   289  	}
   290  	n = utf8.EncodeRune(b.buf[m:m+utf8.UTFMax], r)
   291  	b.buf = b.buf[:m+n]
   292  	return n, nil
   293  }
   294  
   295  // Read reads the next len(p) bytes from the buffer or until the buffer
   296  // is drained. The return value n is the number of bytes read. If the
   297  // buffer has no data to return, err is io.EOF (unless len(p) is zero);
   298  // otherwise it is nil.
   299  func (b *Buffer) Read(p []byte) (n int, err error) {
   300  	b.lastRead = opInvalid
   301  	if b.empty() {
   302  		// Buffer is empty, reset to recover space.
   303  		b.Reset()
   304  		if len(p) == 0 {
   305  			return 0, nil
   306  		}
   307  		return 0, io.EOF
   308  	}
   309  	n = copy(p, b.buf[b.off:])
   310  	b.off += n
   311  	if n > 0 {
   312  		b.lastRead = opRead
   313  	}
   314  	return n, nil
   315  }
   316  
   317  // Next returns a slice containing the next n bytes from the buffer,
   318  // advancing the buffer as if the bytes had been returned by Read.
   319  // If there are fewer than n bytes in the buffer, Next returns the entire buffer.
   320  // The slice is only valid until the next call to a read or write method.
   321  func (b *Buffer) Next(n int) []byte {
   322  	b.lastRead = opInvalid
   323  	m := b.Len()
   324  	if n > m {
   325  		n = m
   326  	}
   327  	data := b.buf[b.off : b.off+n]
   328  	b.off += n
   329  	if n > 0 {
   330  		b.lastRead = opRead
   331  	}
   332  	return data
   333  }
   334  
   335  // ReadByte reads and returns the next byte from the buffer.
   336  // If no byte is available, it returns error io.EOF.
   337  func (b *Buffer) ReadByte() (byte, error) {
   338  	if b.empty() {
   339  		// Buffer is empty, reset to recover space.
   340  		b.Reset()
   341  		return 0, io.EOF
   342  	}
   343  	c := b.buf[b.off]
   344  	b.off++
   345  	b.lastRead = opRead
   346  	return c, nil
   347  }
   348  
   349  // ReadRune reads and returns the next UTF-8-encoded
   350  // Unicode code point from the buffer.
   351  // If no bytes are available, the error returned is io.EOF.
   352  // If the bytes are an erroneous UTF-8 encoding, it
   353  // consumes one byte and returns U+FFFD, 1.
   354  func (b *Buffer) ReadRune() (r rune, size int, err error) {
   355  	if b.empty() {
   356  		// Buffer is empty, reset to recover space.
   357  		b.Reset()
   358  		return 0, 0, io.EOF
   359  	}
   360  	c := b.buf[b.off]
   361  	if c < utf8.RuneSelf {
   362  		b.off++
   363  		b.lastRead = opReadRune1
   364  		return rune(c), 1, nil
   365  	}
   366  	r, n := utf8.DecodeRune(b.buf[b.off:])
   367  	b.off += n
   368  	b.lastRead = readOp(n)
   369  	return r, n, nil
   370  }
   371  
   372  // UnreadRune unreads the last rune returned by ReadRune.
   373  // If the most recent read or write operation on the buffer was
   374  // not a successful ReadRune, UnreadRune returns an error.  (In this regard
   375  // it is stricter than UnreadByte, which will unread the last byte
   376  // from any read operation.)
   377  func (b *Buffer) UnreadRune() error {
   378  	if b.lastRead <= opInvalid {
   379  		return errors.New("bytes.Buffer: UnreadRune: previous operation was not a successful ReadRune")
   380  	}
   381  	if b.off >= int(b.lastRead) {
   382  		b.off -= int(b.lastRead)
   383  	}
   384  	b.lastRead = opInvalid
   385  	return nil
   386  }
   387  
   388  // UnreadByte unreads the last byte returned by the most recent successful
   389  // read operation that read at least one byte. If a write has happened since
   390  // the last read, if the last read returned an error, or if the read read zero
   391  // bytes, UnreadByte returns an error.
   392  func (b *Buffer) UnreadByte() error {
   393  	if b.lastRead == opInvalid {
   394  		return errors.New("bytes.Buffer: UnreadByte: previous operation was not a successful read")
   395  	}
   396  	b.lastRead = opInvalid
   397  	if b.off > 0 {
   398  		b.off--
   399  	}
   400  	return nil
   401  }
   402  
   403  // ReadBytes reads until the first occurrence of delim in the input,
   404  // returning a slice containing the data up to and including the delimiter.
   405  // If ReadBytes encounters an error before finding a delimiter,
   406  // it returns the data read before the error and the error itself (often io.EOF).
   407  // ReadBytes returns err != nil if and only if the returned data does not end in
   408  // delim.
   409  func (b *Buffer) ReadBytes(delim byte) (line []byte, err error) {
   410  	slice, err := b.readSlice(delim)
   411  	// return a copy of slice. The buffer's backing array may
   412  	// be overwritten by later calls.
   413  	line = append(line, slice...)
   414  	return line, err
   415  }
   416  
   417  // readSlice is like ReadBytes but returns a reference to internal buffer data.
   418  func (b *Buffer) readSlice(delim byte) (line []byte, err error) {
   419  	i := IndexByte(b.buf[b.off:], delim)
   420  	end := b.off + i + 1
   421  	if i < 0 {
   422  		end = len(b.buf)
   423  		err = io.EOF
   424  	}
   425  	line = b.buf[b.off:end]
   426  	b.off = end
   427  	b.lastRead = opRead
   428  	return line, err
   429  }
   430  
   431  // ReadString reads until the first occurrence of delim in the input,
   432  // returning a string containing the data up to and including the delimiter.
   433  // If ReadString encounters an error before finding a delimiter,
   434  // it returns the data read before the error and the error itself (often io.EOF).
   435  // ReadString returns err != nil if and only if the returned data does not end
   436  // in delim.
   437  func (b *Buffer) ReadString(delim byte) (line string, err error) {
   438  	slice, err := b.readSlice(delim)
   439  	return string(slice), err
   440  }
   441  
   442  // NewBuffer creates and initializes a new Buffer using buf as its
   443  // initial contents. The new Buffer takes ownership of buf, and the
   444  // caller should not use buf after this call. NewBuffer is intended to
   445  // prepare a Buffer to read existing data. It can also be used to set
   446  // the initial size of the internal buffer for writing. To do that,
   447  // buf should have the desired capacity but a length of zero.
   448  //
   449  // In most cases, new(Buffer) (or just declaring a Buffer variable) is
   450  // sufficient to initialize a Buffer.
   451  func NewBuffer(buf []byte) *Buffer { return &Buffer{buf: buf} }
   452  
   453  // NewBufferString creates and initializes a new Buffer using string s as its
   454  // initial contents. It is intended to prepare a buffer to read an existing
   455  // string.
   456  //
   457  // In most cases, new(Buffer) (or just declaring a Buffer variable) is
   458  // sufficient to initialize a Buffer.
   459  func NewBufferString(s string) *Buffer {
   460  	return &Buffer{buf: []byte(s)}
   461  }
   462  

View as plain text