path: root/src/pkg/net/textproto/reader.go

// Copyright 2010 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.

package textproto

import (
	"bufio"
	"bytes"
	"io"
	"io/ioutil"
	"strconv"
	"strings"
)

// BUG(rsc): To let callers manage exposure to denial of service
// attacks, Reader should allow them to set and reset a limit on
// the number of bytes read from the connection.

// A Reader implements convenience methods for reading requests
// or responses from a text protocol network connection.
type Reader struct {
	R   *bufio.Reader
	dot *dotReader
	buf []byte // a re-usable buffer for readContinuedLineSlice
}

// NewReader returns a new Reader reading from r.
func NewReader(r *bufio.Reader) *Reader {
	return &Reader{R: r}
}

// ReadLine reads a single line from r,
// eliding the final \n or \r\n from the returned string.
func (r *Reader) ReadLine() (string, error) {
	line, err := r.readLineSlice()
	return string(line), err
}

// ReadLineBytes is like ReadLine but returns a []byte instead of a string.
func (r *Reader) ReadLineBytes() ([]byte, error) {
	line, err := r.readLineSlice()
	if line != nil {
		buf := make([]byte, len(line))
		copy(buf, line)
		line = buf
	}
	return line, err
}

func (r *Reader) readLineSlice() ([]byte, error) {
	r.closeDot()
	var line []byte
	for {
		l, more, err := r.R.ReadLine()
		if err != nil {
			return nil, err
		}
		// Avoid the copy if the first call produced a full line.
		if line == nil && !more {
			return l, nil
		}
		line = append(line, l...)
		if !more {
			break
		}
	}
	return line, nil
}

// ReadContinuedLine reads a possibly continued line from r,
// eliding the final trailing ASCII white space.
// Lines after the first are considered continuations if they
// begin with a space or tab character.  In the returned data,
// continuation lines are separated from the previous line
// only by a single space: the newline and leading white space
// are removed.
//
// For example, consider this input:
//
//	Line 1
//	  continued...
//	Line 2
//
// The first call to ReadContinuedLine will return "Line 1 continued..."
// and the second will return "Line 2".
//
// A line consisting of only white space is never continued.
//
func (r *Reader) ReadContinuedLine() (string, error) {
	line, err := r.readContinuedLineSlice()
	return string(line), err
}

// trim returns s with leading and trailing spaces and tabs removed.
// It does not assume Unicode or UTF-8.
func trim(s []byte) []byte {
	i := 0
	for i < len(s) && (s[i] == ' ' || s[i] == '\t') {
		i++
	}
	n := len(s)
	for n > i && (s[n-1] == ' ' || s[n-1] == '\t') {
		n--
	}
	return s[i:n]
}

// ReadContinuedLineBytes is like ReadContinuedLine but
// returns a []byte instead of a string.
func (r *Reader) ReadContinuedLineBytes() ([]byte, error) {
	line, err := r.readContinuedLineSlice()
	if line != nil {
		buf := make([]byte, len(line))
		copy(buf, line)
		line = buf
	}
	return line, err
}

func (r *Reader) readContinuedLineSlice() ([]byte, error) {
	// Read the first line.
	line, err := r.readLineSlice()
	if err != nil {
		return nil, err
	}
	if len(line) == 0 { // blank line - no continuation
		return line, nil
	}

	// Optimistically assume that we have started to buffer the next line
	// and it starts with an ASCII letter (the next header key), so we can
	// avoid copying that buffered data around in memory and skipping over
	// non-existent whitespace.
	if r.R.Buffered() > 1 {
		peek, err := r.R.Peek(1)
		if err == nil && isASCIILetter(peek[0]) {
			return trim(line), nil
		}
	}

	// ReadByte or the next readLineSlice will flush the read buffer;
	// copy the slice into buf.
	r.buf = append(r.buf[:0], trim(line)...)

	// Read continuation lines.
	for r.skipSpace() > 0 {
		line, err := r.readLineSlice()
		if err != nil {
			break
		}
		r.buf = append(r.buf, ' ')
		r.buf = append(r.buf, line...)
	}
	return r.buf, nil
}

// skipSpace skips R over all spaces and returns the number of bytes skipped.
func (r *Reader) skipSpace() int {
	n := 0
	for {
		c, err := r.R.ReadByte()
		if err != nil {
			// Bufio will keep err until next read.
			break
		}
		if c != ' ' && c != '\t' {
			r.R.UnreadByte()
			break
		}
		n++
	}
	return n
}

func (r *Reader) readCodeLine(expectCode int) (code int, continued bool, message string, err error) {
	line, err := r.ReadLine()
	if err != nil {
		return
	}
	return parseCodeLine(line, expectCode)
}

func parseCodeLine(line string, expectCode int) (code int, continued bool, message string, err error) {
	if len(line) < 4 || line[3] != ' ' && line[3] != '-' {
		err = ProtocolError("short response: " + line)
		return
	}
	continued = line[3] == '-'
	code, err = strconv.Atoi(line[0:3])
	if err != nil || code < 100 {
		err = ProtocolError("invalid response code: " + line)
		return
	}
	message = line[4:]
	if 1 <= expectCode && expectCode < 10 && code/100 != expectCode ||
		10 <= expectCode && expectCode < 100 && code/10 != expectCode ||
		100 <= expectCode && expectCode < 1000 && code != expectCode {
		err = &Error{code, message}
	}
	return
}

// ReadCodeLine reads a response code line of the form
//	code message
// where code is a 3-digit status code and the message
// extends to the rest of the line.  An example of such a line is:
//	220 plan9.bell-labs.com ESMTP
//
// If the prefix of the status does not match the digits in expectCode,
// ReadCodeLine returns with err set to &Error{code, message}.
// For example, if expectCode is 31, an error will be returned if
// the status is not in the range [310,319].
//
// If the response is multi-line, ReadCodeLine returns an error.
//
// An expectCode <= 0 disables the check of the status code.
//
func (r *Reader) ReadCodeLine(expectCode int) (code int, message string, err error) {
	code, continued, message, err := r.readCodeLine(expectCode)
	if err == nil && continued {
		err = ProtocolError("unexpected multi-line response: " + message)
	}
	return
}

// ReadResponse reads a multi-line response of the form:
//
//	code-message line 1
//	code-message line 2
//	...
//	code message line n
//
// where code is a 3-digit status code. The first line starts with the
// code and a hyphen. The response is terminated by a line that starts
// with the same code followed by a space. Each line in message is
// separated by a newline (\n).
//
// See page 36 of RFC 959 (http://www.ietf.org/rfc/rfc959.txt) for
// details.
//
// If the prefix of the status does not match the digits in expectCode,
// ReadResponse returns with err set to &Error{code, message}.
// For example, if expectCode is 31, an error will be returned if
// the status is not in the range [310,319].
//
// An expectCode <= 0 disables the check of the status code.
//
func (r *Reader) ReadResponse(expectCode int) (code int, message string, err error) {
	code, continued, message, err := r.readCodeLine(expectCode)
	for err == nil && continued {
		line, err := r.ReadLine()
		if err != nil {
			return 0, "", err
		}

		var code2 int
		var moreMessage string
		code2, continued, moreMessage, err = parseCodeLine(line, expectCode)
		if err != nil || code2 != code {
			message += "\n" + strings.TrimRight(line, "\r\n")
			continued = true
			continue
		}
		message += "\n" + moreMessage
	}
	return
}

// DotReader returns a new Reader that satisfies Reads using the
// decoded text of a dot-encoded block read from r.
// The returned Reader is only valid until the next call
// to a method on r.
//
// Dot encoding is a common framing used for data blocks
// in text protocols such as SMTP.  The data consists of a sequence
// of lines, each of which ends in "\r\n".  The sequence itself
// ends at a line containing just a dot: ".\r\n".  Lines beginning
// with a dot are escaped with an additional dot to avoid
// looking like the end of the sequence.
//
// The decoded form returned by the Reader's Read method
// rewrites the "\r\n" line endings into the simpler "\n",
// removes leading dot escapes if present, and stops with error io.EOF
// after consuming (and discarding) the end-of-sequence line.
func (r *Reader) DotReader() io.Reader {
	r.closeDot()
	r.dot = &dotReader{r: r}
	return r.dot
}

type dotReader struct {
	r     *Reader
	state int
}

// Read satisfies reads by decoding dot-encoded data read from d.r.
func (d *dotReader) Read(b []byte) (n int, err error) {
	// Run data through a simple state machine to
	// elide leading dots, rewrite trailing \r\n into \n,
	// and detect ending .\r\n line.
	const (
		stateBeginLine = iota // beginning of line; initial state; must be zero
		stateDot              // read . at beginning of line
		stateDotCR            // read .\r at beginning of line
		stateCR               // read \r (possibly at end of line)
		stateData             // reading data in middle of line
		stateEOF              // reached .\r\n end marker line
	)
	br := d.r.R
	for n < len(b) && d.state != stateEOF {
		var c byte
		c, err = br.ReadByte()
		if err != nil {
			if err == io.EOF {
				err = io.ErrUnexpectedEOF
			}
			break
		}
		switch d.state {
		case stateBeginLine:
			if c == '.' {
				d.state = stateDot
				continue
			}
			if c == '\r' {
				d.state = stateCR
				continue
			}
			d.state = stateData

		case stateDot:
			if c == '\r' {
				d.state = stateDotCR
				continue
			}
			if c == '\n' {
				d.state = stateEOF
				continue
			}
			d.state = stateData

		case stateDotCR:
			if c == '\n' {
				d.state = stateEOF
				continue
			}
			// Not part of .\r\n.
			// Consume leading dot and emit saved \r.
			br.UnreadByte()
			c = '\r'
			d.state = stateData

		case stateCR:
			if c == '\n' {
				d.state = stateBeginLine
				break
			}
			// Not part of \r\n.  Emit saved \r
			br.UnreadByte()
			c = '\r'
			d.state = stateData

		case stateData:
			if c == '\r' {
				d.state = stateCR
				continue
			}
			if c == '\n' {
				d.state = stateBeginLine
			}
		}
		b[n] = c
		n++
	}
	if err == nil && d.state == stateEOF {
		err = io.EOF
	}
	if err != nil && d.r.dot == d {
		d.r.dot = nil
	}
	return
}

// closeDot drains the current DotReader if any,
// making sure that it reads until the ending dot line.
func (r *Reader) closeDot() {
	if r.dot == nil {
		return
	}
	buf := make([]byte, 128)
	for r.dot != nil {
		// When Read reaches EOF or an error,
		// it will set r.dot == nil.
		r.dot.Read(buf)
	}
}

// ReadDotBytes reads a dot-encoding and returns the decoded data.
//
// See the documentation for the DotReader method for details about dot-encoding.
func (r *Reader) ReadDotBytes() ([]byte, error) {
	return ioutil.ReadAll(r.DotReader())
}

// ReadDotLines reads a dot-encoding and returns a slice
// containing the decoded lines, with the final \r\n or \n elided from each.
//
// See the documentation for the DotReader method for details about dot-encoding.
func (r *Reader) ReadDotLines() ([]string, error) {
	// We could use ReadDotBytes and then Split it,
	// but reading a line at a time avoids needing a
	// large contiguous block of memory and is simpler.
	var v []string
	var err error
	for {
		var line string
		line, err = r.ReadLine()
		if err != nil {
			if err == io.EOF {
				err = io.ErrUnexpectedEOF
			}
			break
		}

		// Dot by itself marks end; otherwise cut one dot.
		if len(line) > 0 && line[0] == '.' {
			if len(line) == 1 {
				break
			}
			line = line[1:]
		}
		v = append(v, line)
	}
	return v, err
}

// ReadMIMEHeader reads a MIME-style header from r.
// The header is a sequence of possibly continued Key: Value lines
// ending in a blank line.
// The returned map m maps CanonicalMIMEHeaderKey(key) to a
// sequence of values in the same order encountered in the input.
//
// For example, consider this input:
//
//	My-Key: Value 1
//	Long-Key: Even
//	       Longer Value
//	My-Key: Value 2
//
// Given that input, ReadMIMEHeader returns the map:
//
//	map[string][]string{
//		"My-Key": {"Value 1", "Value 2"},
//		"Long-Key": {"Even Longer Value"},
//	}
//
func (r *Reader) ReadMIMEHeader() (MIMEHeader, error) {
	m := make(MIMEHeader, 4)
	for {
		kv, err := r.readContinuedLineSlice()
		if len(kv) == 0 {
			return m, err
		}

		// Key ends at first colon; should not have spaces but
		// they appear in the wild, violating specs, so we
		// remove them if present.
		i := bytes.IndexByte(kv, ':')
		if i < 0 {
			return m, ProtocolError("malformed MIME header line: " + string(kv))
		}
		endKey := i
		for endKey > 0 && kv[endKey-1] == ' ' {
			endKey--
		}
		key := canonicalMIMEHeaderKey(kv[:endKey])

		// Skip initial spaces in value.
		i++ // skip colon
		for i < len(kv) && (kv[i] == ' ' || kv[i] == '\t') {
			i++
		}
		value := string(kv[i:])

		m[key] = append(m[key], value)

		if err != nil {
			return m, err
		}
	}
	panic("unreachable")
}

// CanonicalMIMEHeaderKey returns the canonical format of the
// MIME header key s.  The canonicalization converts the first
// letter and any letter following a hyphen to upper case;
// the rest are converted to lowercase.  For example, the
// canonical key for "accept-encoding" is "Accept-Encoding".
// MIME header keys are assumed to be ASCII only.
func CanonicalMIMEHeaderKey(s string) string {
	// Quick check for canonical encoding.
	upper := true
	for i := 0; i < len(s); i++ {
		c := s[i]
		if upper && 'a' <= c && c <= 'z' {
			return canonicalMIMEHeaderKey([]byte(s))
		}
		if !upper && 'A' <= c && c <= 'Z' {
			return canonicalMIMEHeaderKey([]byte(s))
		}
		upper = c == '-'
	}
	return s
}

const toLower = 'a' - 'A'

// canonicalMIMEHeaderKey is like CanonicalMIMEHeaderKey but is
// allowed to mutate the provided byte slice before returning the
// string.
func canonicalMIMEHeaderKey(a []byte) string {
	// Look for it in commonHeaders , so that we can avoid an
	// allocation by sharing the strings among all users
	// of textproto. If we don't find it, a has been canonicalized
	// so just return string(a).
	upper := true
	lo := 0
	hi := len(commonHeaders)
	for i := 0; i < len(a); i++ {
		// Canonicalize: first letter upper case
		// and upper case after each dash.
		// (Host, User-Agent, If-Modified-Since).
		// MIME headers are ASCII only, so no Unicode issues.
		if a[i] == ' ' {
			a[i] = '-'
			upper = true
			continue
		}
		c := a[i]
		if upper && 'a' <= c && c <= 'z' {
			c -= toLower
		} else if !upper && 'A' <= c && c <= 'Z' {
			c += toLower
		}
		a[i] = c
		upper = c == '-' // for next time

		if lo < hi {
			for lo < hi && (len(commonHeaders[lo]) <= i || commonHeaders[lo][i] < c) {
				lo++
			}
			for hi > lo && commonHeaders[hi-1][i] > c {
				hi--
			}
		}
	}
	if lo < hi && len(commonHeaders[lo]) == len(a) {
		return commonHeaders[lo]
	}
	return string(a)
}

var commonHeaders = []string{
	"Accept",
	"Accept-Charset",
	"Accept-Encoding",
	"Accept-Language",
	"Accept-Ranges",
	"Cache-Control",
	"Cc",
	"Connection",
	"Content-Id",
	"Content-Language",
	"Content-Length",
	"Content-Transfer-Encoding",
	"Content-Type",
	"Date",
	"Dkim-Signature",
	"Etag",
	"Expires",
	"From",
	"Host",
	"If-Modified-Since",
	"If-None-Match",
	"In-Reply-To",
	"Last-Modified",
	"Location",
	"Message-Id",
	"Mime-Version",
	"Pragma",
	"Received",
	"Return-Path",
	"Server",
	"Set-Cookie",
	"Subject",
	"To",
	"User-Agent",
	"Via",
	"X-Forwarded-For",
	"X-Imforwards",
	"X-Powered-By",
}