vbatts/tar-split
1
0
Fork 0
forked from mirrors/tar-split
Code Releases Activity

archive/tar: properly format GNU base-256 encoding

Browse source 
Motivation:
* Previous implementation silently failed when an integer overflow
occurred. Now, we report an ErrFieldTooLong.
* Previous implementation did not encode in two's complement format and was
unable to encode negative numbers.

The relevant GNU specification says:
<<<
GNU format uses two's-complement base-256 notation to store values
that do not fit into standard ustar range.
>>>

Fixes #12436

Change-Id: I09c20602eabf8ae3a7e0db35b79440a64bfaf807
Reviewed-on: https://go-review.googlesource.com/17425
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
This commit is contained in:
Joe Tsai 2015-12-02 15:48:06 -08:00 • committed by Vincent Batts
parent be9ac88117
commit ce5aac17f9
2 changed files with 105 additions and 4 deletions
Download patch file Download diff file Expand all files Collapse all files

26
archive/tar/writer.go
View file

@ -94,13 +94,31 @@ func (f *formatter) formatOctal(b []byte, x int64) {
f.formatString(b, s)
}
// 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 {
var binBits = uint(n-1) * 8
return n >= 9 || (x >= -1<<binBits && x < 1<<binBits)
}
// Write x into b, as binary (GNUtar/star extension).
func (f *formatter) formatNumeric(b []byte, x int64) {
for i := len(b) - 1; x > 0 && i >= 0; i-- {
b[i] = byte(x)
x >>= 8
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
}
b[0] |= 0x80 // highest bit indicates binary format
f.formatOctal(b, 0) // Last resort, just write zero
f.err = ErrFieldTooLong
}
var (