...
Run Format

Source file src/archive/tar/strconv.go

Documentation: archive/tar

  // Copyright 2016 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 tar
  
  import (
  	"bytes"
  	"fmt"
  	"strconv"
  	"strings"
  	"time"
  )
  
  // hasNUL reports whether the NUL character exists within s.
  func hasNUL(s string) bool {
  	return strings.IndexByte(s, 0) >= 0
  }
  
  // isASCII reports whether the input is an ASCII C-style string.
  func isASCII(s string) bool {
  	for _, c := range s {
  		if c >= 0x80 || c == 0x00 {
  			return false
  		}
  	}
  	return true
  }
  
  // toASCII converts the input to an ASCII C-style string.
  // This a best effort conversion, so invalid characters are dropped.
  func toASCII(s string) string {
  	if isASCII(s) {
  		return s
  	}
  	b := make([]byte, 0, len(s))
  	for _, c := range s {
  		if c < 0x80 && c != 0x00 {
  			b = append(b, byte(c))
  		}
  	}
  	return string(b)
  }
  
  type parser struct {
  	err error // Last error seen
  }
  
  type formatter struct {
  	err error // Last error seen
  }
  
  // parseString parses bytes as a NUL-terminated C-style string.
  // If a NUL byte is not found then the whole slice is returned as a string.
  func (*parser) parseString(b []byte) string {
  	if i := bytes.IndexByte(b, 0); i >= 0 {
  		return string(b[:i])
  	}
  	return string(b)
  }
  
  // formatString copies s into b, NUL-terminating if possible.
  func (f *formatter) formatString(b []byte, s string) {
  	if len(s) > len(b) {
  		f.err = ErrFieldTooLong
  	}
  	copy(b, s)
  	if len(s) < len(b) {
  		b[len(s)] = 0
  	}
  
  	// Some buggy readers treat regular files with a trailing slash
  	// in the V7 path field as a directory even though the full path
  	// recorded elsewhere (e.g., via PAX record) contains no trailing slash.
  	if len(s) > len(b) && b[len(b)-1] == '/' {
  		n := len(strings.TrimRight(s[:len(b)], "/"))
  		b[n] = 0 // Replace trailing slash with NUL terminator
  	}
  }
  
  // fitsInBase256 reports whether x can be encoded into n bytes using base-256
  // encoding. Unlike octal encoding, base-256 encoding does not require that the
  // string ends with a NUL character. Thus, all n bytes are available for output.
  //
  // If operating in binary mode, this assumes strict GNU binary mode; which means
  // that the first byte can only be either 0x80 or 0xff. Thus, the first byte is
  // equivalent to the sign bit in two's complement form.
  func fitsInBase256(n int, x int64) bool {
  	binBits := uint(n-1) * 8
  	return n >= 9 || (x >= -1<<binBits && x < 1<<binBits)
  }
  
  // parseNumeric parses the input as being encoded in either base-256 or octal.
  // This function may return negative numbers.
  // If parsing fails or an integer overflow occurs, err will be set.
  func (p *parser) parseNumeric(b []byte) int64 {
  	// Check for base-256 (binary) format first.
  	// If the first bit is set, then all following bits constitute a two's
  	// complement encoded number in big-endian byte order.
  	if len(b) > 0 && b[0]&0x80 != 0 {
  		// Handling negative numbers relies on the following identity:
  		//	-a-1 == ^a
  		//
  		// If the number is negative, we use an inversion mask to invert the
  		// data bytes and treat the value as an unsigned number.
  		var inv byte // 0x00 if positive or zero, 0xff if negative
  		if b[0]&0x40 != 0 {
  			inv = 0xff
  		}
  
  		var x uint64
  		for i, c := range b {
  			c ^= inv // Inverts c only if inv is 0xff, otherwise does nothing
  			if i == 0 {
  				c &= 0x7f // Ignore signal bit in first byte
  			}
  			if (x >> 56) > 0 {
  				p.err = ErrHeader // Integer overflow
  				return 0
  			}
  			x = x<<8 | uint64(c)
  		}
  		if (x >> 63) > 0 {
  			p.err = ErrHeader // Integer overflow
  			return 0
  		}
  		if inv == 0xff {
  			return ^int64(x)
  		}
  		return int64(x)
  	}
  
  	// Normal case is base-8 (octal) format.
  	return p.parseOctal(b)
  }
  
  // formatNumeric encodes x into b using base-8 (octal) encoding if possible.
  // Otherwise it will attempt to use base-256 (binary) encoding.
  func (f *formatter) formatNumeric(b []byte, x int64) {
  	if fitsInOctal(len(b), x) {
  		f.formatOctal(b, x)
  		return
  	}
  
  	if fitsInBase256(len(b), x) {
  		for i := len(b) - 1; i >= 0; i-- {
  			b[i] = byte(x)
  			x >>= 8
  		}
  		b[0] |= 0x80 // Highest bit indicates binary format
  		return
  	}
  
  	f.formatOctal(b, 0) // Last resort, just write zero
  	f.err = ErrFieldTooLong
  }
  
  func (p *parser) parseOctal(b []byte) int64 {
  	// Because unused fields are filled with NULs, we need
  	// to skip leading NULs. Fields may also be padded with
  	// spaces or NULs.
  	// So we remove leading and trailing NULs and spaces to
  	// be sure.
  	b = bytes.Trim(b, " \x00")
  
  	if len(b) == 0 {
  		return 0
  	}
  	x, perr := strconv.ParseUint(p.parseString(b), 8, 64)
  	if perr != nil {
  		p.err = ErrHeader
  	}
  	return int64(x)
  }
  
  func (f *formatter) formatOctal(b []byte, x int64) {
  	if !fitsInOctal(len(b), x) {
  		x = 0 // Last resort, just write zero
  		f.err = ErrFieldTooLong
  	}
  
  	s := strconv.FormatInt(x, 8)
  	// Add leading zeros, but leave room for a NUL.
  	if n := len(b) - len(s) - 1; n > 0 {
  		s = strings.Repeat("0", n) + s
  	}
  	f.formatString(b, s)
  }
  
  // fitsInOctal reports whether the integer x fits in a field n-bytes long
  // using octal encoding with the appropriate NUL terminator.
  func fitsInOctal(n int, x int64) bool {
  	octBits := uint(n-1) * 3
  	return x >= 0 && (n >= 22 || x < 1<<octBits)
  }
  
  // parsePAXTime takes a string of the form %d.%d as described in the PAX
  // specification. Note that this implementation allows for negative timestamps,
  // which is allowed for by the PAX specification, but not always portable.
  func parsePAXTime(s string) (time.Time, error) {
  	const maxNanoSecondDigits = 9
  
  	// Split string into seconds and sub-seconds parts.
  	ss, sn := s, ""
  	if pos := strings.IndexByte(s, '.'); pos >= 0 {
  		ss, sn = s[:pos], s[pos+1:]
  	}
  
  	// Parse the seconds.
  	secs, err := strconv.ParseInt(ss, 10, 64)
  	if err != nil {
  		return time.Time{}, ErrHeader
  	}
  	if len(sn) == 0 {
  		return time.Unix(secs, 0), nil // No sub-second values
  	}
  
  	// Parse the nanoseconds.
  	if strings.Trim(sn, "0123456789") != "" {
  		return time.Time{}, ErrHeader
  	}
  	if len(sn) < maxNanoSecondDigits {
  		sn += strings.Repeat("0", maxNanoSecondDigits-len(sn)) // Right pad
  	} else {
  		sn = sn[:maxNanoSecondDigits] // Right truncate
  	}
  	nsecs, _ := strconv.ParseInt(sn, 10, 64) // Must succeed
  	if len(ss) > 0 && ss[0] == '-' {
  		return time.Unix(secs, -1*nsecs), nil // Negative correction
  	}
  	return time.Unix(secs, nsecs), nil
  }
  
  // formatPAXTime converts ts into a time of the form %d.%d as described in the
  // PAX specification. This function is capable of negative timestamps.
  func formatPAXTime(ts time.Time) (s string) {
  	secs, nsecs := ts.Unix(), ts.Nanosecond()
  	if nsecs == 0 {
  		return strconv.FormatInt(secs, 10)
  	}
  
  	// If seconds is negative, then perform correction.
  	sign := ""
  	if secs < 0 {
  		sign = "-"             // Remember sign
  		secs = -(secs + 1)     // Add a second to secs
  		nsecs = -(nsecs - 1E9) // Take that second away from nsecs
  	}
  	return strings.TrimRight(fmt.Sprintf("%s%d.%09d", sign, secs, nsecs), "0")
  }
  
  // parsePAXRecord parses the input PAX record string into a key-value pair.
  // If parsing is successful, it will slice off the currently read record and
  // return the remainder as r.
  func parsePAXRecord(s string) (k, v, r string, err error) {
  	// The size field ends at the first space.
  	sp := strings.IndexByte(s, ' ')
  	if sp == -1 {
  		return "", "", s, ErrHeader
  	}
  
  	// Parse the first token as a decimal integer.
  	n, perr := strconv.ParseInt(s[:sp], 10, 0) // Intentionally parse as native int
  	if perr != nil || n < 5 || int64(len(s)) < n {
  		return "", "", s, ErrHeader
  	}
  
  	// Extract everything between the space and the final newline.
  	rec, nl, rem := s[sp+1:n-1], s[n-1:n], s[n:]
  	if nl != "\n" {
  		return "", "", s, ErrHeader
  	}
  
  	// The first equals separates the key from the value.
  	eq := strings.IndexByte(rec, '=')
  	if eq == -1 {
  		return "", "", s, ErrHeader
  	}
  	k, v = rec[:eq], rec[eq+1:]
  
  	if !validPAXRecord(k, v) {
  		return "", "", s, ErrHeader
  	}
  	return k, v, rem, nil
  }
  
  // formatPAXRecord formats a single PAX record, prefixing it with the
  // appropriate length.
  func formatPAXRecord(k, v string) (string, error) {
  	if !validPAXRecord(k, v) {
  		return "", ErrHeader
  	}
  
  	const padding = 3 // Extra padding for ' ', '=', and '\n'
  	size := len(k) + len(v) + padding
  	size += len(strconv.Itoa(size))
  	record := strconv.Itoa(size) + " " + k + "=" + v + "\n"
  
  	// Final adjustment if adding size field increased the record size.
  	if len(record) != size {
  		size = len(record)
  		record = strconv.Itoa(size) + " " + k + "=" + v + "\n"
  	}
  	return record, nil
  }
  
  // validPAXRecord reports whether the key-value pair is valid where each
  // record is formatted as:
  //	"%d %s=%s\n" % (size, key, value)
  //
  // Keys and values should be UTF-8, but the number of bad writers out there
  // forces us to be a more liberal.
  // Thus, we only reject all keys with NUL, and only reject NULs in values
  // for the PAX version of the USTAR string fields.
  // The key must not contain an '=' character.
  func validPAXRecord(k, v string) bool {
  	if k == "" || strings.IndexByte(k, '=') >= 0 {
  		return false
  	}
  	switch k {
  	case paxPath, paxLinkpath, paxUname, paxGname:
  		return !hasNUL(v)
  	default:
  		return !hasNUL(k)
  	}
  }
  

View as plain text