// Copyright 2009 The Go Authors. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
// This package implements buffered I/O. It wraps an io.Reader or io.Writer
// object, creating another object (Reader or Writer) that also implements
// the interface but provides buffering and some help for textual I/O.
package bufio
import (
"bytes";
"io";
"os";
"strconv";
"utf8";
)
const (
defaultBufSize = 4096;
)
// Errors introduced by this package.
type Error struct {
os.ErrorString;
}
var (
ErrInvalidUnreadByte os.Error = &Error{"bufio: invalid use of UnreadByte"};
ErrBufferFull os.Error = &Error{"bufio: buffer full"};
errInternal os.Error = &Error{"bufio: internal error"};
)
// BufSizeError is the error representing an invalid buffer size.
type BufSizeError int
func (b BufSizeError) String() string {
return "bufio: bad buffer size " + strconv.Itoa(int(b))
}
// Buffered input.
// Reader implements buffering for an io.Reader object.
type Reader struct {
buf []byte;
rd io.Reader;
r, w int;
err os.Error;
lastbyte int;
}
// NewReaderSize creates a new Reader whose buffer has the specified size,
// which must be greater than zero. If the argument io.Reader is already a
// Reader with large enough size, it returns the underlying Reader.
// It returns the Reader and any error.
func NewReaderSize(rd io.Reader, size int) (*Reader, os.Error) {
if size <= 0 {
return nil, BufSizeError(size)
}
// Is it already a Reader?
b, ok := rd.(*Reader);
if ok && len(b.buf) >= size {
return b, nil
}
b = new(Reader);
b.buf = make([]byte, size);
b.rd = rd;
b.lastbyte = -1;
return b, nil;
}
// NewReader returns a new Reader whose buffer has the default size.
func NewReader(rd io.Reader) *Reader {
b, err := NewReaderSize(rd, defaultBufSize);
if err != nil {
// cannot happen - defaultBufSize is a valid size
panic("bufio: NewReader: ", err.String())
}
return b;
}
// fill reads a new chunk into the buffer.
func (b *Reader) fill() {
// Slide existing data to beginning.
if b.w > b.r {
copy(b.buf[0:b.w-b.r], b.buf[b.r:b.w]);
b.w -= b.r;
} else {
b.w = 0
}
b.r = 0;
// Read new data.
n, e := b.rd.Read(b.buf[b.w:]);
b.w += n;
if e != nil {
b.err = e
}
}
// Read reads data into p.
// It returns the number of bytes read into p.
// If nn < len(p), also returns an error explaining
// why the read is short. At EOF, the count will be
// zero and err will be os.EOF.
func (b *Reader) Read(p []byte) (nn int, err os.Error) {
nn = 0;
for len(p) > 0 {
n := len(p);
if b.w == b.r {
if b.err != nil {
return nn, b.err
}
if len(p) >= len(b.buf) {
// Large read, empty buffer.
// Read directly into p to avoid copy.
n, b.err = b.rd.Read(p);
if n > 0 {
b.lastbyte = int(p[n-1])
}
p = p[n:];
nn += n;
continue;
}
b.fill();
continue;
}
if n > b.w-b.r {
n = b.w - b.r
}
copy(p[0:n], b.buf[b.r:b.r+n]);
p = p[n:];
b.r += n;
b.lastbyte = int(b.buf[b.r-1]);
nn += n;
}
return nn, nil;
}
// ReadByte reads and returns a single byte.
// If no byte is available, returns an error.
func (b *Reader) ReadByte() (c byte, err os.Error) {
for b.w == b.r {
if b.err != nil {
return 0, b.err
}
b.fill();
}
c = b.buf[b.r];
b.r++;
b.lastbyte = int(c);
return c, nil;
}
// UnreadByte unreads the last byte. Only the most recently read byte can be unread.
func (b *Reader) UnreadByte() os.Error {
if b.r == b.w && b.lastbyte >= 0 {
b.w = 1;
b.r = 0;
b.buf[0] = byte(b.lastbyte);
b.lastbyte = -1;
return nil;
}
if b.r <= 0 {
return ErrInvalidUnreadByte
}
b.r--;
b.lastbyte = -1;
return nil;
}
// ReadRune reads a single UTF-8 encoded Unicode character and returns the
// rune and its size in bytes.
func (b *Reader) ReadRune() (rune int, size int, err os.Error) {
for b.r+utf8.UTFMax > b.w && !utf8.FullRune(b.buf[b.r:b.w]) && b.err == nil {
b.fill()
}
if b.r == b.w {
return 0, 0, b.err
}
rune, size = int(b.buf[b.r]), 1;
if rune >= 0x80 {
rune, size = utf8.DecodeRune(b.buf[b.r:b.w])
}
b.r += size;
b.lastbyte = int(b.buf[b.r-1]);
return rune, size, nil;
}
// Buffered returns the number of bytes that can be read from the current buffer.
func (b *Reader) Buffered() int { return b.w - b.r }
// ReadSlice reads until the first occurrence of delim in the input,
// returning a slice pointing at the bytes in the buffer.
// The bytes stop being valid at the next read call.
// If ReadSlice encounters an error before finding a delimiter,
// it returns all the data in the buffer and the error itself (often os.EOF).
// ReadSlice fails with error ErrBufferFull if the buffer fills without a delim.
// Because the data returned from ReadSlice will be overwritten
// by the next I/O operation, most clients should use
// ReadBytes or ReadString instead.
// ReadSlice returns err != nil if and only if line does not end in delim.
func (b *Reader) ReadSlice(delim byte) (line []byte, err os.Error) {
// Look in buffer.
if i := bytes.IndexByte(b.buf[b.r:b.w], delim); i >= 0 {
line1 := b.buf[b.r : b.r+i+1];
b.r += i + 1;
return line1, nil;
}
// Read more into buffer, until buffer fills or we find delim.
for {
if b.err != nil {
line := b.buf[b.r:b.w];
b.r = b.w;
return line, b.err;
}
n := b.Buffered();
b.fill();
// Search new part of buffer
if i := bytes.IndexByte(b.buf[n:b.w], delim); i >= 0 {
line := b.buf[0 : n+i+1];
b.r = n + i + 1;
return line, nil;
}
// Buffer is full?
if b.Buffered() >= len(b.buf) {
return nil, ErrBufferFull
}
}
panic("not reached");
}
// ReadBytes reads until the first occurrence of delim in the input,
// returning a string containing the data up to and including the delimiter.
// If ReadBytes encounters an error before finding a delimiter,
// it returns the data read before the error and the error itself (often os.EOF).
// ReadBytes returns err != nil if and only if line does not end in delim.
func (b *Reader) ReadBytes(delim byte) (line []byte, err os.Error) {
// Use ReadSlice to look for array,
// accumulating full buffers.
var frag []byte;
var full [][]byte;
nfull := 0;
err = nil;
for {
var e os.Error;
frag, e = b.ReadSlice(delim);
if e == nil { // got final fragment
break
}
if e != ErrBufferFull { // unexpected error
err = e;
break;
}
// Read bytes out of buffer.
buf := make([]byte, b.Buffered());
var n int;
n, e = b.Read(buf);
if e != nil {
frag = buf[0:n];
err = e;
break;
}
if n != len(buf) {
frag = buf[0:n];
err = errInternal;
break;
}
// Grow list if needed.
if full == nil {
full = make([][]byte, 16)
} else if nfull >= len(full) {
newfull := make([][]byte, len(full)*2);
for i := 0; i < len(full); i++ {
newfull[i] = full[i]
}
full = newfull;
}
// Save buffer
full[nfull] = buf;
nfull++;
}
// Allocate new buffer to hold the full pieces and the fragment.
n := 0;
for i := 0; i < nfull; i++ {
n += len(full[i])
}
n += len(frag);
// Copy full pieces and fragment in.
buf := make([]byte, n);
n = 0;
for i := 0; i < nfull; i++ {
copy(buf[n:n+len(full[i])], full[i]);
n += len(full[i]);
}
copy(buf[n:n+len(frag)], frag);
return buf, err;
}
// ReadString reads until the first occurrence of delim in the input,
// returning a string containing the data up to and including the delimiter.
// If ReadString encounters an error before finding a delimiter,
// it returns the data read before the error and the error itself (often os.EOF).
// ReadString returns err != nil if and only if line does not end in delim.
func (b *Reader) ReadString(delim byte) (line string, err os.Error) {
bytes, e := b.ReadBytes(delim);
return string(bytes), e;
}
// buffered output
// Writer implements buffering for an io.Writer object.
type Writer struct {
err os.Error;
buf []byte;
n int;
wr io.Writer;
}
// NewWriterSize creates a new Writer whose buffer has the specified size,
// which must be greater than zero. If the argument io.Writer is already a
// Writer with large enough size, it returns the underlying Writer.
// It returns the Writer and any error.
func NewWriterSize(wr io.Writer, size int) (*Writer, os.Error) {
if size <= 0 {
return nil, BufSizeError(size)
}
// Is it already a Writer?
b, ok := wr.(*Writer);
if ok && len(b.buf) >= size {
return b, nil
}
b = new(Writer);
b.buf = make([]byte, size);
b.wr = wr;
return b, nil;
}
// NewWriter returns a new Writer whose buffer has the default size.
func NewWriter(wr io.Writer) *Writer {
b, err := NewWriterSize(wr, defaultBufSize);
if err != nil {
// cannot happen - defaultBufSize is valid size
panic("bufio: NewWriter: ", err.String())
}
return b;
}
// Flush writes any buffered data to the underlying io.Writer.
func (b *Writer) Flush() os.Error {
if b.err != nil {
return b.err
}
n, e := b.wr.Write(b.buf[0:b.n]);
if n < b.n && e == nil {
e = io.ErrShortWrite
}
if e != nil {
if n > 0 && n < b.n {
copy(b.buf[0:b.n-n], b.buf[n:b.n])
}
b.n -= n;
b.err = e;
return e;
}
b.n = 0;
return nil;
}
// Available returns how many bytes are unused in the buffer.
func (b *Writer) Available() int { return len(b.buf) - b.n }
// Buffered returns the number of bytes that have been written into the current buffer.
func (b *Writer) Buffered() int { return b.n }
// Write writes the contents of p into the buffer.
// It returns the number of bytes written.
// If nn < len(p), also returns an error explaining
// why the write is short.
func (b *Writer) Write(p []byte) (nn int, err os.Error) {
if b.err != nil {
return 0, b.err
}
nn = 0;
for len(p) > 0 {
n := b.Available();
if n <= 0 {
if b.Flush(); b.err != nil {
break
}
n = b.Available();
}
if b.Available() == 0 && len(p) >= len(b.buf) {
// Large write, empty buffer.
// Write directly from p to avoid copy.
n, b.err = b.wr.Write(p);
nn += n;
p = p[n:];
if b.err != nil {
break
}
continue;
}
if n > len(p) {
n = len(p)
}
copy(b.buf[b.n:b.n+n], p[0:n]);
b.n += n;
nn += n;
p = p[n:];
}
return nn, b.err;
}
// WriteByte writes a single byte.
func (b *Writer) WriteByte(c byte) os.Error {
if b.err != nil {
return b.err
}
if b.Available() <= 0 && b.Flush() != nil {
return b.err
}
b.buf[b.n] = c;
b.n++;
return nil;
}
// WriteString writes a string.
func (b *Writer) WriteString(s string) os.Error {
if b.err != nil {
return b.err
}
// Common case, worth making fast.
if b.Available() >= len(s) || len(b.buf) >= len(s) && b.Flush() == nil {
for i := 0; i < len(s); i++ { // loop over bytes, not runes.
b.buf[b.n] = s[i];
b.n++;
}
return nil;
}
for i := 0; i < len(s); i++ { // loop over bytes, not runes.
b.WriteByte(s[i])
}
return b.err;
}
// buffered input and output
// ReadWriter stores pointers to a Reader and a Writer.
// It implements io.ReadWriter.
type ReadWriter struct {
*Reader;
*Writer;
}
// NewReadWriter allocates a new ReadWriter that dispatches to r and w.
func NewReadWriter(r *Reader, w *Writer) *ReadWriter {
return &ReadWriter{r, w}
}
|