This commit is contained in:
Mikaël Cluseau 2019-10-09 16:40:56 +11:00
parent 9dc80cd34b
commit 6e432c2a06
69 changed files with 18 additions and 17561 deletions

7
go.mod
View File

@ -1,7 +1,10 @@
module novit.nc/direktil/pkg module novit.nc/direktil/pkg
require ( require (
github.com/ulikunitz/xz v0.5.5 github.com/kr/pretty v0.1.0 // indirect
github.com/ulikunitz/xz v0.5.6
gopkg.in/check.v1 v1.0.0-20180628173108-788fd7840127 // indirect gopkg.in/check.v1 v1.0.0-20180628173108-788fd7840127 // indirect
gopkg.in/yaml.v2 v2.2.2 gopkg.in/yaml.v2 v2.2.4
) )
go 1.13

16
go.sum
View File

@ -1,8 +1,12 @@
github.com/ulikunitz/xz v0.5.4/go.mod h1:2bypXElzHzzJZwzH67Y6wb67pO62Rzfn7BSiF4ABRW8= github.com/kr/pretty v0.1.0 h1:L/CwN0zerZDmRFUapSPitk6f+Q3+0za1rQkzVuMiMFI=
github.com/ulikunitz/xz v0.5.5 h1:pFrO0lVpTBXLpYw+pnLj6TbvHuyjXMfjGeCwSqCVwok= github.com/kr/pretty v0.1.0/go.mod h1:dAy3ld7l9f0ibDNOQOHHMYYIIbhfbHSm3C4ZsoJORNo=
github.com/ulikunitz/xz v0.5.5/go.mod h1:2bypXElzHzzJZwzH67Y6wb67pO62Rzfn7BSiF4ABRW8= github.com/kr/pty v1.1.1/go.mod h1:pFQYn66WHrOpPYNljwOMqo10TkYh1fy3cYio2l3bCsQ=
github.com/kr/text v0.1.0 h1:45sCR5RtlFHMR4UwH9sdQ5TC8v0qDQCHnXt+kaKSTVE=
github.com/kr/text v0.1.0/go.mod h1:4Jbv+DJW3UT/LiOwJeYQe1efqtUx/iVham/4vfdArNI=
github.com/ulikunitz/xz v0.5.6 h1:jGHAfXawEGZQ3blwU5wnWKQJvAraT7Ftq9EXjnXYgt8=
github.com/ulikunitz/xz v0.5.6/go.mod h1:2bypXElzHzzJZwzH67Y6wb67pO62Rzfn7BSiF4ABRW8=
gopkg.in/check.v1 v0.0.0-20161208181325-20d25e280405/go.mod h1:Co6ibVJAznAaIkqp8huTwlJQCZ016jof/cbN4VW5Yz0= gopkg.in/check.v1 v0.0.0-20161208181325-20d25e280405/go.mod h1:Co6ibVJAznAaIkqp8huTwlJQCZ016jof/cbN4VW5Yz0=
gopkg.in/check.v1 v1.0.0-20180628173108-788fd7840127 h1:qIbj1fsPNlZgppZ+VLlY7N33q108Sa+fhmuc+sWQYwY=
gopkg.in/check.v1 v1.0.0-20180628173108-788fd7840127/go.mod h1:Co6ibVJAznAaIkqp8huTwlJQCZ016jof/cbN4VW5Yz0= gopkg.in/check.v1 v1.0.0-20180628173108-788fd7840127/go.mod h1:Co6ibVJAznAaIkqp8huTwlJQCZ016jof/cbN4VW5Yz0=
gopkg.in/yaml.v2 v2.2.1/go.mod h1:hI93XBmqTisBFMUTm0b8Fm+jr3Dg1NNxqwp+5A1VGuI= gopkg.in/yaml.v2 v2.2.4 h1:/eiJrUcujPVeJ3xlSWaiNi3uSVmDGBK1pDHUHAnao1I=
gopkg.in/yaml.v2 v2.2.2 h1:ZCJp+EgiOT7lHqUV2J862kp8Qj64Jo6az82+3Td9dZw= gopkg.in/yaml.v2 v2.2.4/go.mod h1:hI93XBmqTisBFMUTm0b8Fm+jr3Dg1NNxqwp+5A1VGuI=
gopkg.in/yaml.v2 v2.2.2/go.mod h1:hI93XBmqTisBFMUTm0b8Fm+jr3Dg1NNxqwp+5A1VGuI=

View File

@ -17,6 +17,7 @@ type Config struct {
type Cluster struct { type Cluster struct {
Name string Name string
Addons string Addons string
BootstrapPods string
} }
func FromBytes(data []byte) (*Config, error) { func FromBytes(data []byte) (*Config, error) {

View File

@ -1,25 +0,0 @@
# .gitignore
TODO.html
README.html
lzma/writer.txt
lzma/reader.txt
cmd/gxz/gxz
cmd/xb/xb
# test executables
*.test
# profile files
*.out
# vim swap file
.*.swp
# executables on windows
*.exe
# default compression test file
enwik8*

View File

@ -1,26 +0,0 @@
Copyright (c) 2014-2016 Ulrich Kunitz
All rights reserved.
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions are met:
* Redistributions of source code must retain the above copyright notice, this
list of conditions and the following disclaimer.
* Redistributions in binary form must reproduce the above copyright notice,
this list of conditions and the following disclaimer in the documentation
and/or other materials provided with the distribution.
* My name, Ulrich Kunitz, may not be used to endorse or promote products
derived from this software without specific prior written permission.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE
FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

View File

@ -1,73 +0,0 @@
# Package xz
This Go language package supports the reading and writing of xz
compressed streams. It includes also a gxz command for compressing and
decompressing data. The package is completely written in Go and doesn't
have any dependency on any C code.
The package is currently under development. There might be bugs and APIs
are not considered stable. At this time the package cannot compete with
the xz tool regarding compression speed and size. The algorithms there
have been developed over a long time and are highly optimized. However
there are a number of improvements planned and I'm very optimistic about
parallel compression and decompression. Stay tuned!
## Using the API
The following example program shows how to use the API.
```go
package main
import (
"bytes"
"io"
"log"
"os"
"github.com/ulikunitz/xz"
)
func main() {
const text = "The quick brown fox jumps over the lazy dog.\n"
var buf bytes.Buffer
// compress text
w, err := xz.NewWriter(&buf)
if err != nil {
log.Fatalf("xz.NewWriter error %s", err)
}
if _, err := io.WriteString(w, text); err != nil {
log.Fatalf("WriteString error %s", err)
}
if err := w.Close(); err != nil {
log.Fatalf("w.Close error %s", err)
}
// decompress buffer and write output to stdout
r, err := xz.NewReader(&buf)
if err != nil {
log.Fatalf("NewReader error %s", err)
}
if _, err = io.Copy(os.Stdout, r); err != nil {
log.Fatalf("io.Copy error %s", err)
}
}
```
## Using the gxz compression tool
The package includes a gxz command line utility for compression and
decompression.
Use following command for installation:
$ go get github.com/ulikunitz/xz/cmd/gxz
To test it call the following command.
$ gxz bigfile
After some time a much smaller file bigfile.xz will replace bigfile.
To decompress it use the following command.
$ gxz -d bigfile.xz

View File

@ -1,319 +0,0 @@
# TODO list
## Release v0.6
1. Review encoder and check for lzma improvements under xz.
2. Fix binary tree matcher.
3. Compare compression ratio with xz tool using comparable parameters
and optimize parameters
4. Do some optimizations
- rename operation action and make it a simple type of size 8
- make maxMatches, wordSize parameters
- stop searching after a certain length is found (parameter sweetLen)
## Release v0.7
1. Optimize code
2. Do statistical analysis to get linear presets.
3. Test sync.Pool compatability for xz and lzma Writer and Reader
3. Fuzz optimized code.
## Release v0.8
1. Support parallel go routines for writing and reading xz files.
2. Support a ReaderAt interface for xz files with small block sizes.
3. Improve compatibility between gxz and xz
4. Provide manual page for gxz
## Release v0.9
1. Improve documentation
2. Fuzz again
## Release v1.0
1. Full functioning gxz
2. Add godoc URL to README.md (godoc.org)
3. Resolve all issues.
4. Define release candidates.
5. Public announcement.
## Package lzma
### Release v0.6
- Rewrite Encoder into a simple greedy one-op-at-a-time encoder
including
+ simple scan at the dictionary head for the same byte
+ use the killer byte (requiring matches to get longer, the first
test should be the byte that would make the match longer)
## Optimizations
- There may be a lot of false sharing in lzma.State; check whether this
can be improved by reorganizing the internal structure of it.
- Check whether batching encoding and decoding improves speed.
### DAG optimizations
- Use full buffer to create minimal bit-length above range encoder.
- Might be too slow (see v0.4)
### Different match finders
- hashes with 2, 3 characters additional to 4 characters
- binary trees with 2-7 characters (uint64 as key, use uint32 as
pointers into a an array)
- rb-trees with 2-7 characters (uint64 as key, use uint32 as pointers
into an array with bit-steeling for the colors)
## Release Procedure
- execute goch -l for all packages; probably with lower param like 0.5.
- check orthography with gospell
- Write release notes in doc/relnotes.
- Update README.md
- xb copyright . in xz directory to ensure all new files have Copyright
header
- VERSION=<version> go generate github.com/ulikunitz/xz/... to update
version files
- Execute test for Linux/amd64, Linux/x86 and Windows/amd64.
- Update TODO.md - write short log entry
- git checkout master && git merge dev
- git tag -a <version>
- git push
## Log
### 2018-10-28
Release v0.5.5 fixes issues #19 observing ErrLimit outputs.
### 2017-06-05
Release v0.5.4 fixes issues #15 of another problem with the padding size
check for the xz block header. I removed the check completely.
### 2017-02-15
Release v0.5.3 fixes issue #12 regarding the decompression of an empty
XZ stream. Many thanks to Tomasz Kłak, who reported the issue.
### 2016-12-02
Release v0.5.2 became necessary to allow the decoding of xz files with
4-byte padding in the block header. Many thanks to Greg, who reported
the issue.
### 2016-07-23
Release v0.5.1 became necessary to fix problems with 32-bit platforms.
Many thanks to Bruno Brigas, who reported the issue.
### 2016-07-04
Release v0.5 provides improvements to the compressor and provides support for
the decompression of xz files with multiple xz streams.
### 2016-01-31
Another compression rate increase by checking the byte at length of the
best match first, before checking the whole prefix. This makes the
compressor even faster. We have now a large time budget to beat the
compression ratio of the xz tool. For enwik8 we have now over 40 seconds
to reduce the compressed file size for another 7 MiB.
### 2016-01-30
I simplified the encoder. Speed and compression rate increased
dramatically. A high compression rate affects also the decompression
speed. The approach with the buffer and optimizing for operation
compression rate has not been successful. Going for the maximum length
appears to be the best approach.
### 2016-01-28
The release v0.4 is ready. It provides a working xz implementation,
which is rather slow, but works and is interoperable with the xz tool.
It is an important milestone.
### 2016-01-10
I have the first working implementation of an xz reader and writer. I'm
happy about reaching this milestone.
### 2015-12-02
I'm now ready to implement xz because, I have a working LZMA2
implementation. I decided today that v0.4 will use the slow encoder
using the operations buffer to be able to go back, if I intend to do so.
### 2015-10-21
I have restarted the work on the library. While trying to implement
LZMA2, I discovered that I need to resimplify the encoder and decoder
functions. The option approach is too complicated. Using a limited byte
writer and not caring for written bytes at all and not to try to handle
uncompressed data simplifies the LZMA encoder and decoder much.
Processing uncompressed data and handling limits is a feature of the
LZMA2 format not of LZMA.
I learned an interesting method from the LZO format. If the last copy is
too far away they are moving the head one 2 bytes and not 1 byte to
reduce processing times.
### 2015-08-26
I have now reimplemented the lzma package. The code is reasonably fast,
but can still be optimized. The next step is to implement LZMA2 and then
xz.
### 2015-07-05
Created release v0.3. The version is the foundation for a full xz
implementation that is the target of v0.4.
### 2015-06-11
The gflag package has been developed because I couldn't use flag and
pflag for a fully compatible support of gzip's and lzma's options. It
seems to work now quite nicely.
### 2015-06-05
The overflow issue was interesting to research, however Henry S. Warren
Jr. Hacker's Delight book was very helpful as usual and had the issue
explained perfectly. Fefe's information on his website was based on the
C FAQ and quite bad, because it didn't address the issue of -MININT ==
MININT.
### 2015-06-04
It has been a productive day. I improved the interface of lzma.Reader
and lzma.Writer and fixed the error handling.
### 2015-06-01
By computing the bit length of the LZMA operations I was able to
improve the greedy algorithm implementation. By using an 8 MByte buffer
the compression rate was not as good as for xz but already better then
gzip default.
Compression is currently slow, but this is something we will be able to
improve over time.
### 2015-05-26
Checked the license of ogier/pflag. The binary lzmago binary should
include the license terms for the pflag library.
I added the endorsement clause as used by Google for the Go sources the
LICENSE file.
### 2015-05-22
The package lzb contains now the basic implementation for creating or
reading LZMA byte streams. It allows the support for the implementation
of the DAG-shortest-path algorithm for the compression function.
### 2015-04-23
Completed yesterday the lzbase classes. I'm a little bit concerned that
using the components may require too much code, but on the other hand
there is a lot of flexibility.
### 2015-04-22
Implemented Reader and Writer during the Bayern game against Porto. The
second half gave me enough time.
### 2015-04-21
While showering today morning I discovered that the design for OpEncoder
and OpDecoder doesn't work, because encoding/decoding might depend on
the current status of the dictionary. This is not exactly the right way
to start the day.
Therefore we need to keep the Reader and Writer design. This time around
we simplify it by ignoring size limits. These can be added by wrappers
around the Reader and Writer interfaces. The Parameters type isn't
needed anymore.
However I will implement a ReaderState and WriterState type to use
static typing to ensure the right State object is combined with the
right lzbase.Reader and lzbase.Writer.
As a start I have implemented ReaderState and WriterState to ensure
that the state for reading is only used by readers and WriterState only
used by Writers.
### 2015-04-20
Today I implemented the OpDecoder and tested OpEncoder and OpDecoder.
### 2015-04-08
Came up with a new simplified design for lzbase. I implemented already
the type State that replaces OpCodec.
### 2015-04-06
The new lzma package is now fully usable and lzmago is using it now. The
old lzma package has been completely removed.
### 2015-04-05
Implemented lzma.Reader and tested it.
### 2015-04-04
Implemented baseReader by adapting code form lzma.Reader.
### 2015-04-03
The opCodec has been copied yesterday to lzma2. opCodec has a high
number of dependencies on other files in lzma2. Therefore I had to copy
almost all files from lzma.
### 2015-03-31
Removed only a TODO item.
However in Francesco Campoy's presentation "Go for Javaneros
(Javaïstes?)" is the the idea that using an embedded field E, all the
methods of E will be defined on T. If E is an interface T satisfies E.
https://talks.golang.org/2014/go4java.slide#51
I have never used this, but it seems to be a cool idea.
### 2015-03-30
Finished the type writerDict and wrote a simple test.
### 2015-03-25
I started to implement the writerDict.
### 2015-03-24
After thinking long about the LZMA2 code and several false starts, I
have now a plan to create a self-sufficient lzma2 package that supports
the classic LZMA format as well as LZMA2. The core idea is to support a
baseReader and baseWriter type that support the basic LZMA stream
without any headers. Both types must support the reuse of dictionaries
and the opCodec.
### 2015-01-10
1. Implemented simple lzmago tool
2. Tested tool against large 4.4G file
- compression worked correctly; tested decompression with lzma
- decompression hits a full buffer condition
3. Fixed a bug in the compressor and wrote a test for it
4. Executed full cycle for 4.4 GB file; performance can be improved ;-)
### 2015-01-11
- Release v0.2 because of the working LZMA encoder and decoder

View File

@ -1,74 +0,0 @@
// Copyright 2014-2017 Ulrich Kunitz. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
package xz
import (
"errors"
"io"
)
// putUint32LE puts the little-endian representation of x into the first
// four bytes of p.
func putUint32LE(p []byte, x uint32) {
p[0] = byte(x)
p[1] = byte(x >> 8)
p[2] = byte(x >> 16)
p[3] = byte(x >> 24)
}
// putUint64LE puts the little-endian representation of x into the first
// eight bytes of p.
func putUint64LE(p []byte, x uint64) {
p[0] = byte(x)
p[1] = byte(x >> 8)
p[2] = byte(x >> 16)
p[3] = byte(x >> 24)
p[4] = byte(x >> 32)
p[5] = byte(x >> 40)
p[6] = byte(x >> 48)
p[7] = byte(x >> 56)
}
// uint32LE converts a little endian representation to an uint32 value.
func uint32LE(p []byte) uint32 {
return uint32(p[0]) | uint32(p[1])<<8 | uint32(p[2])<<16 |
uint32(p[3])<<24
}
// putUvarint puts a uvarint representation of x into the byte slice.
func putUvarint(p []byte, x uint64) int {
i := 0
for x >= 0x80 {
p[i] = byte(x) | 0x80
x >>= 7
i++
}
p[i] = byte(x)
return i + 1
}
// errOverflow indicates an overflow of the 64-bit unsigned integer.
var errOverflowU64 = errors.New("xz: uvarint overflows 64-bit unsigned integer")
// readUvarint reads a uvarint from the given byte reader.
func readUvarint(r io.ByteReader) (x uint64, n int, err error) {
var s uint
i := 0
for {
b, err := r.ReadByte()
if err != nil {
return x, i, err
}
i++
if b < 0x80 {
if i > 10 || i == 10 && b > 1 {
return x, i, errOverflowU64
}
return x | uint64(b)<<s, i, nil
}
x |= uint64(b&0x7f) << s
s += 7
}
}

View File

@ -1,54 +0,0 @@
// Copyright 2014-2017 Ulrich Kunitz. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
package xz
import (
"hash"
"hash/crc32"
"hash/crc64"
)
// crc32Hash implements the hash.Hash32 interface with Sum returning the
// crc32 value in little-endian encoding.
type crc32Hash struct {
hash.Hash32
}
// Sum returns the crc32 value as little endian.
func (h crc32Hash) Sum(b []byte) []byte {
p := make([]byte, 4)
putUint32LE(p, h.Hash32.Sum32())
b = append(b, p...)
return b
}
// newCRC32 returns a CRC-32 hash that returns the 64-bit value in
// little-endian encoding using the IEEE polynomial.
func newCRC32() hash.Hash {
return crc32Hash{Hash32: crc32.NewIEEE()}
}
// crc64Hash implements the Hash64 interface with Sum returning the
// CRC-64 value in little-endian encoding.
type crc64Hash struct {
hash.Hash64
}
// Sum returns the CRC-64 value in little-endian encoding.
func (h crc64Hash) Sum(b []byte) []byte {
p := make([]byte, 8)
putUint64LE(p, h.Hash64.Sum64())
b = append(b, p...)
return b
}
// crc64Table is used to create a CRC-64 hash.
var crc64Table = crc64.MakeTable(crc64.ECMA)
// newCRC64 returns a CRC-64 hash that returns the 64-bit value in
// little-endian encoding using the ECMA polynomial.
func newCRC64() hash.Hash {
return crc64Hash{Hash64: crc64.New(crc64Table)}
}

View File

@ -1,40 +0,0 @@
// Copyright 2014-2017 Ulrich Kunitz. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
// +build ignore
package main
import (
"bytes"
"io"
"log"
"os"
"github.com/ulikunitz/xz"
)
func main() {
const text = "The quick brown fox jumps over the lazy dog.\n"
var buf bytes.Buffer
// compress text
w, err := xz.NewWriter(&buf)
if err != nil {
log.Fatalf("xz.NewWriter error %s", err)
}
if _, err := io.WriteString(w, text); err != nil {
log.Fatalf("WriteString error %s", err)
}
if err := w.Close(); err != nil {
log.Fatalf("w.Close error %s", err)
}
// decompress buffer and write output to stdout
r, err := xz.NewReader(&buf)
if err != nil {
log.Fatalf("NewReader error %s", err)
}
if _, err = io.Copy(os.Stdout, r); err != nil {
log.Fatalf("io.Copy error %s", err)
}
}

View File

@ -1,728 +0,0 @@
// Copyright 2014-2017 Ulrich Kunitz. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
package xz
import (
"bytes"
"crypto/sha256"
"errors"
"fmt"
"hash"
"hash/crc32"
"io"
"github.com/ulikunitz/xz/lzma"
)
// allZeros checks whether a given byte slice has only zeros.
func allZeros(p []byte) bool {
for _, c := range p {
if c != 0 {
return false
}
}
return true
}
// padLen returns the length of the padding required for the given
// argument.
func padLen(n int64) int {
k := int(n % 4)
if k > 0 {
k = 4 - k
}
return k
}
/*** Header ***/
// headerMagic stores the magic bytes for the header
var headerMagic = []byte{0xfd, '7', 'z', 'X', 'Z', 0x00}
// HeaderLen provides the length of the xz file header.
const HeaderLen = 12
// Constants for the checksum methods supported by xz.
const (
CRC32 byte = 0x1
CRC64 = 0x4
SHA256 = 0xa
)
// errInvalidFlags indicates that flags are invalid.
var errInvalidFlags = errors.New("xz: invalid flags")
// verifyFlags returns the error errInvalidFlags if the value is
// invalid.
func verifyFlags(flags byte) error {
switch flags {
case CRC32, CRC64, SHA256:
return nil
default:
return errInvalidFlags
}
}
// flagstrings maps flag values to strings.
var flagstrings = map[byte]string{
CRC32: "CRC-32",
CRC64: "CRC-64",
SHA256: "SHA-256",
}
// flagString returns the string representation for the given flags.
func flagString(flags byte) string {
s, ok := flagstrings[flags]
if !ok {
return "invalid"
}
return s
}
// newHashFunc returns a function that creates hash instances for the
// hash method encoded in flags.
func newHashFunc(flags byte) (newHash func() hash.Hash, err error) {
switch flags {
case CRC32:
newHash = newCRC32
case CRC64:
newHash = newCRC64
case SHA256:
newHash = sha256.New
default:
err = errInvalidFlags
}
return
}
// header provides the actual content of the xz file header: the flags.
type header struct {
flags byte
}
// Errors returned by readHeader.
var errHeaderMagic = errors.New("xz: invalid header magic bytes")
// ValidHeader checks whether data is a correct xz file header. The
// length of data must be HeaderLen.
func ValidHeader(data []byte) bool {
var h header
err := h.UnmarshalBinary(data)
return err == nil
}
// String returns a string representation of the flags.
func (h header) String() string {
return flagString(h.flags)
}
// UnmarshalBinary reads header from the provided data slice.
func (h *header) UnmarshalBinary(data []byte) error {
// header length
if len(data) != HeaderLen {
return errors.New("xz: wrong file header length")
}
// magic header
if !bytes.Equal(headerMagic, data[:6]) {
return errHeaderMagic
}
// checksum
crc := crc32.NewIEEE()
crc.Write(data[6:8])
if uint32LE(data[8:]) != crc.Sum32() {
return errors.New("xz: invalid checksum for file header")
}
// stream flags
if data[6] != 0 {
return errInvalidFlags
}
flags := data[7]
if err := verifyFlags(flags); err != nil {
return err
}
h.flags = flags
return nil
}
// MarshalBinary generates the xz file header.
func (h *header) MarshalBinary() (data []byte, err error) {
if err = verifyFlags(h.flags); err != nil {
return nil, err
}
data = make([]byte, 12)
copy(data, headerMagic)
data[7] = h.flags
crc := crc32.NewIEEE()
crc.Write(data[6:8])
putUint32LE(data[8:], crc.Sum32())
return data, nil
}
/*** Footer ***/
// footerLen defines the length of the footer.
const footerLen = 12
// footerMagic contains the footer magic bytes.
var footerMagic = []byte{'Y', 'Z'}
// footer represents the content of the xz file footer.
type footer struct {
indexSize int64
flags byte
}
// String prints a string representation of the footer structure.
func (f footer) String() string {
return fmt.Sprintf("%s index size %d", flagString(f.flags), f.indexSize)
}
// Minimum and maximum for the size of the index (backward size).
const (
minIndexSize = 4
maxIndexSize = (1 << 32) * 4
)
// MarshalBinary converts footer values into an xz file footer. Note
// that the footer value is checked for correctness.
func (f *footer) MarshalBinary() (data []byte, err error) {
if err = verifyFlags(f.flags); err != nil {
return nil, err
}
if !(minIndexSize <= f.indexSize && f.indexSize <= maxIndexSize) {
return nil, errors.New("xz: index size out of range")
}
if f.indexSize%4 != 0 {
return nil, errors.New(
"xz: index size not aligned to four bytes")
}
data = make([]byte, footerLen)
// backward size (index size)
s := (f.indexSize / 4) - 1
putUint32LE(data[4:], uint32(s))
// flags
data[9] = f.flags
// footer magic
copy(data[10:], footerMagic)
// CRC-32
crc := crc32.NewIEEE()
crc.Write(data[4:10])
putUint32LE(data, crc.Sum32())
return data, nil
}
// UnmarshalBinary sets the footer value by unmarshalling an xz file
// footer.
func (f *footer) UnmarshalBinary(data []byte) error {
if len(data) != footerLen {
return errors.New("xz: wrong footer length")
}
// magic bytes
if !bytes.Equal(data[10:], footerMagic) {
return errors.New("xz: footer magic invalid")
}
// CRC-32
crc := crc32.NewIEEE()
crc.Write(data[4:10])
if uint32LE(data) != crc.Sum32() {
return errors.New("xz: footer checksum error")
}
var g footer
// backward size (index size)
g.indexSize = (int64(uint32LE(data[4:])) + 1) * 4
// flags
if data[8] != 0 {
return errInvalidFlags
}
g.flags = data[9]
if err := verifyFlags(g.flags); err != nil {
return err
}
*f = g
return nil
}
/*** Block Header ***/
// blockHeader represents the content of an xz block header.
type blockHeader struct {
compressedSize int64
uncompressedSize int64
filters []filter
}
// String converts the block header into a string.
func (h blockHeader) String() string {
var buf bytes.Buffer
first := true
if h.compressedSize >= 0 {
fmt.Fprintf(&buf, "compressed size %d", h.compressedSize)
first = false
}
if h.uncompressedSize >= 0 {
if !first {
buf.WriteString(" ")
}
fmt.Fprintf(&buf, "uncompressed size %d", h.uncompressedSize)
first = false
}
for _, f := range h.filters {
if !first {
buf.WriteString(" ")
}
fmt.Fprintf(&buf, "filter %s", f)
first = false
}
return buf.String()
}
// Masks for the block flags.
const (
filterCountMask = 0x03
compressedSizePresent = 0x40
uncompressedSizePresent = 0x80
reservedBlockFlags = 0x3C
)
// errIndexIndicator signals that an index indicator (0x00) has been found
// instead of an expected block header indicator.
var errIndexIndicator = errors.New("xz: found index indicator")
// readBlockHeader reads the block header.
func readBlockHeader(r io.Reader) (h *blockHeader, n int, err error) {
var buf bytes.Buffer
buf.Grow(20)
// block header size
z, err := io.CopyN(&buf, r, 1)
n = int(z)
if err != nil {
return nil, n, err
}
s := buf.Bytes()[0]
if s == 0 {
return nil, n, errIndexIndicator
}
// read complete header
headerLen := (int(s) + 1) * 4
buf.Grow(headerLen - 1)
z, err = io.CopyN(&buf, r, int64(headerLen-1))
n += int(z)
if err != nil {
return nil, n, err
}
// unmarshal block header
h = new(blockHeader)
if err = h.UnmarshalBinary(buf.Bytes()); err != nil {
return nil, n, err
}
return h, n, nil
}
// readSizeInBlockHeader reads the uncompressed or compressed size
// fields in the block header. The present value informs the function
// whether the respective field is actually present in the header.
func readSizeInBlockHeader(r io.ByteReader, present bool) (n int64, err error) {
if !present {
return -1, nil
}
x, _, err := readUvarint(r)
if err != nil {
return 0, err
}
if x >= 1<<63 {
return 0, errors.New("xz: size overflow in block header")
}
return int64(x), nil
}
// UnmarshalBinary unmarshals the block header.
func (h *blockHeader) UnmarshalBinary(data []byte) error {
// Check header length
s := data[0]
if data[0] == 0 {
return errIndexIndicator
}
headerLen := (int(s) + 1) * 4
if len(data) != headerLen {
return fmt.Errorf("xz: data length %d; want %d", len(data),
headerLen)
}
n := headerLen - 4
// Check CRC-32
crc := crc32.NewIEEE()
crc.Write(data[:n])
if crc.Sum32() != uint32LE(data[n:]) {
return errors.New("xz: checksum error for block header")
}
// Block header flags
flags := data[1]
if flags&reservedBlockFlags != 0 {
return errors.New("xz: reserved block header flags set")
}
r := bytes.NewReader(data[2:n])
// Compressed size
var err error
h.compressedSize, err = readSizeInBlockHeader(
r, flags&compressedSizePresent != 0)
if err != nil {
return err
}
// Uncompressed size
h.uncompressedSize, err = readSizeInBlockHeader(
r, flags&uncompressedSizePresent != 0)
if err != nil {
return err
}
h.filters, err = readFilters(r, int(flags&filterCountMask)+1)
if err != nil {
return err
}
// Check padding
// Since headerLen is a multiple of 4 we don't need to check
// alignment.
k := r.Len()
// The standard spec says that the padding should have not more
// than 3 bytes. However we found paddings of 4 or 5 in the
// wild. See https://github.com/ulikunitz/xz/pull/11 and
// https://github.com/ulikunitz/xz/issues/15
//
// The only reasonable approach seems to be to ignore the
// padding size. We still check that all padding bytes are zero.
if !allZeros(data[n-k : n]) {
return errPadding
}
return nil
}
// MarshalBinary marshals the binary header.
func (h *blockHeader) MarshalBinary() (data []byte, err error) {
if !(minFilters <= len(h.filters) && len(h.filters) <= maxFilters) {
return nil, errors.New("xz: filter count wrong")
}
for i, f := range h.filters {
if i < len(h.filters)-1 {
if f.id() == lzmaFilterID {
return nil, errors.New(
"xz: LZMA2 filter is not the last")
}
} else {
// last filter
if f.id() != lzmaFilterID {
return nil, errors.New("xz: " +
"last filter must be the LZMA2 filter")
}
}
}
var buf bytes.Buffer
// header size must set at the end
buf.WriteByte(0)
// flags
flags := byte(len(h.filters) - 1)
if h.compressedSize >= 0 {
flags |= compressedSizePresent
}
if h.uncompressedSize >= 0 {
flags |= uncompressedSizePresent
}
buf.WriteByte(flags)
p := make([]byte, 10)
if h.compressedSize >= 0 {
k := putUvarint(p, uint64(h.compressedSize))
buf.Write(p[:k])
}
if h.uncompressedSize >= 0 {
k := putUvarint(p, uint64(h.uncompressedSize))
buf.Write(p[:k])
}
for _, f := range h.filters {
fp, err := f.MarshalBinary()
if err != nil {
return nil, err
}
buf.Write(fp)
}
// padding
for i := padLen(int64(buf.Len())); i > 0; i-- {
buf.WriteByte(0)
}
// crc place holder
buf.Write(p[:4])
data = buf.Bytes()
if len(data)%4 != 0 {
panic("data length not aligned")
}
s := len(data)/4 - 1
if !(1 < s && s <= 255) {
panic("wrong block header size")
}
data[0] = byte(s)
crc := crc32.NewIEEE()
crc.Write(data[:len(data)-4])
putUint32LE(data[len(data)-4:], crc.Sum32())
return data, nil
}
// Constants used for marshalling and unmarshalling filters in the xz
// block header.
const (
minFilters = 1
maxFilters = 4
minReservedID = 1 << 62
)
// filter represents a filter in the block header.
type filter interface {
id() uint64
UnmarshalBinary(data []byte) error
MarshalBinary() (data []byte, err error)
reader(r io.Reader, c *ReaderConfig) (fr io.Reader, err error)
writeCloser(w io.WriteCloser, c *WriterConfig) (fw io.WriteCloser, err error)
// filter must be last filter
last() bool
}
// readFilter reads a block filter from the block header. At this point
// in time only the LZMA2 filter is supported.
func readFilter(r io.Reader) (f filter, err error) {
br := lzma.ByteReader(r)
// index
id, _, err := readUvarint(br)
if err != nil {
return nil, err
}
var data []byte
switch id {
case lzmaFilterID:
data = make([]byte, lzmaFilterLen)
data[0] = lzmaFilterID
if _, err = io.ReadFull(r, data[1:]); err != nil {
return nil, err
}
f = new(lzmaFilter)
default:
if id >= minReservedID {
return nil, errors.New(
"xz: reserved filter id in block stream header")
}
return nil, errors.New("xz: invalid filter id")
}
if err = f.UnmarshalBinary(data); err != nil {
return nil, err
}
return f, err
}
// readFilters reads count filters. At this point in time only the count
// 1 is supported.
func readFilters(r io.Reader, count int) (filters []filter, err error) {
if count != 1 {
return nil, errors.New("xz: unsupported filter count")
}
f, err := readFilter(r)
if err != nil {
return nil, err
}
return []filter{f}, err
}
// writeFilters writes the filters.
func writeFilters(w io.Writer, filters []filter) (n int, err error) {
for _, f := range filters {
p, err := f.MarshalBinary()
if err != nil {
return n, err
}
k, err := w.Write(p)
n += k
if err != nil {
return n, err
}
}
return n, nil
}
/*** Index ***/
// record describes a block in the xz file index.
type record struct {
unpaddedSize int64
uncompressedSize int64
}
// readRecord reads an index record.
func readRecord(r io.ByteReader) (rec record, n int, err error) {
u, k, err := readUvarint(r)
n += k
if err != nil {
return rec, n, err
}
rec.unpaddedSize = int64(u)
if rec.unpaddedSize < 0 {
return rec, n, errors.New("xz: unpadded size negative")
}
u, k, err = readUvarint(r)
n += k
if err != nil {
return rec, n, err
}
rec.uncompressedSize = int64(u)
if rec.uncompressedSize < 0 {
return rec, n, errors.New("xz: uncompressed size negative")
}
return rec, n, nil
}
// MarshalBinary converts an index record in its binary encoding.
func (rec *record) MarshalBinary() (data []byte, err error) {
// maximum length of a uvarint is 10
p := make([]byte, 20)
n := putUvarint(p, uint64(rec.unpaddedSize))
n += putUvarint(p[n:], uint64(rec.uncompressedSize))
return p[:n], nil
}
// writeIndex writes the index, a sequence of records.
func writeIndex(w io.Writer, index []record) (n int64, err error) {
crc := crc32.NewIEEE()
mw := io.MultiWriter(w, crc)
// index indicator
k, err := mw.Write([]byte{0})
n += int64(k)
if err != nil {
return n, err
}
// number of records
p := make([]byte, 10)
k = putUvarint(p, uint64(len(index)))
k, err = mw.Write(p[:k])
n += int64(k)
if err != nil {
return n, err
}
// list of records
for _, rec := range index {
p, err := rec.MarshalBinary()
if err != nil {
return n, err
}
k, err = mw.Write(p)
n += int64(k)
if err != nil {
return n, err
}
}
// index padding
k, err = mw.Write(make([]byte, padLen(int64(n))))
n += int64(k)
if err != nil {
return n, err
}
// crc32 checksum
putUint32LE(p, crc.Sum32())
k, err = w.Write(p[:4])
n += int64(k)
return n, err
}
// readIndexBody reads the index from the reader. It assumes that the
// index indicator has already been read.
func readIndexBody(r io.Reader) (records []record, n int64, err error) {
crc := crc32.NewIEEE()
// index indicator
crc.Write([]byte{0})
br := lzma.ByteReader(io.TeeReader(r, crc))
// number of records
u, k, err := readUvarint(br)
n += int64(k)
if err != nil {
return nil, n, err
}
recLen := int(u)
if recLen < 0 || uint64(recLen) != u {
return nil, n, errors.New("xz: record number overflow")
}
// list of records
records = make([]record, recLen)
for i := range records {
records[i], k, err = readRecord(br)
n += int64(k)
if err != nil {
return nil, n, err
}
}
p := make([]byte, padLen(int64(n+1)), 4)
k, err = io.ReadFull(br.(io.Reader), p)
n += int64(k)
if err != nil {
return nil, n, err
}
if !allZeros(p) {
return nil, n, errors.New("xz: non-zero byte in index padding")
}
// crc32
s := crc.Sum32()
p = p[:4]
k, err = io.ReadFull(br.(io.Reader), p)
n += int64(k)
if err != nil {
return records, n, err
}
if uint32LE(p) != s {
return nil, n, errors.New("xz: wrong checksum for index")
}
return records, n, nil
}

BIN
vendor/github.com/ulikunitz/xz/fox.xz generated vendored

Binary file not shown.

View File

@ -1,181 +0,0 @@
// Copyright 2014-2017 Ulrich Kunitz. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
package hash
// CyclicPoly provides a cyclic polynomial rolling hash.
type CyclicPoly struct {
h uint64
p []uint64
i int
}
// ror rotates the unsigned 64-bit integer to right. The argument s must be
// less than 64.
func ror(x uint64, s uint) uint64 {
return (x >> s) | (x << (64 - s))
}
// NewCyclicPoly creates a new instance of the CyclicPoly structure. The
// argument n gives the number of bytes for which a hash will be executed.
// This number must be positive; the method panics if this isn't the case.
func NewCyclicPoly(n int) *CyclicPoly {
if n < 1 {
panic("argument n must be positive")
}
return &CyclicPoly{p: make([]uint64, 0, n)}
}
// Len returns the length of the byte sequence for which a hash is generated.
func (r *CyclicPoly) Len() int {
return cap(r.p)
}
// RollByte hashes the next byte and returns a hash value. The complete becomes
// available after at least Len() bytes have been hashed.
func (r *CyclicPoly) RollByte(x byte) uint64 {
y := hash[x]
if len(r.p) < cap(r.p) {
r.h = ror(r.h, 1) ^ y
r.p = append(r.p, y)
} else {
r.h ^= ror(r.p[r.i], uint(cap(r.p)-1))
r.h = ror(r.h, 1) ^ y
r.p[r.i] = y
r.i = (r.i + 1) % cap(r.p)
}
return r.h
}
// Stores the hash for the individual bytes.
var hash = [256]uint64{
0x2e4fc3f904065142, 0xc790984cfbc99527,
0x879f95eb8c62f187, 0x3b61be86b5021ef2,
0x65a896a04196f0a5, 0xc5b307b80470b59e,
0xd3bff376a70df14b, 0xc332f04f0b3f1701,
0x753b5f0e9abf3e0d, 0xb41538fdfe66ef53,
0x1906a10c2c1c0208, 0xfb0c712a03421c0d,
0x38be311a65c9552b, 0xfee7ee4ca6445c7e,
0x71aadeded184f21e, 0xd73426fccda23b2d,
0x29773fb5fb9600b5, 0xce410261cd32981a,
0xfe2848b3c62dbc2d, 0x459eaaff6e43e11c,
0xc13e35fc9c73a887, 0xf30ed5c201e76dbc,
0xa5f10b3910482cea, 0x2945d59be02dfaad,
0x06ee334ff70571b5, 0xbabf9d8070f44380,
0xee3e2e9912ffd27c, 0x2a7118d1ea6b8ea7,
0x26183cb9f7b1664c, 0xea71dac7da068f21,
0xea92eca5bd1d0bb7, 0x415595862defcd75,
0x248a386023c60648, 0x9cf021ab284b3c8a,
0xfc9372df02870f6c, 0x2b92d693eeb3b3fc,
0x73e799d139dc6975, 0x7b15ae312486363c,
0xb70e5454a2239c80, 0x208e3fb31d3b2263,
0x01f563cabb930f44, 0x2ac4533d2a3240d8,
0x84231ed1064f6f7c, 0xa9f020977c2a6d19,
0x213c227271c20122, 0x09fe8a9a0a03d07a,
0x4236dc75bcaf910c, 0x460a8b2bead8f17e,
0xd9b27be1aa07055f, 0xd202d5dc4b11c33e,
0x70adb010543bea12, 0xcdae938f7ea6f579,
0x3f3d870208672f4d, 0x8e6ccbce9d349536,
0xe4c0871a389095ae, 0xf5f2a49152bca080,
0x9a43f9b97269934e, 0xc17b3753cb6f475c,
0xd56d941e8e206bd4, 0xac0a4f3e525eda00,
0xa06d5a011912a550, 0x5537ed19537ad1df,
0xa32fe713d611449d, 0x2a1d05b47c3b579f,
0x991d02dbd30a2a52, 0x39e91e7e28f93eb0,
0x40d06adb3e92c9ac, 0x9b9d3afde1c77c97,
0x9a3f3f41c02c616f, 0x22ecd4ba00f60c44,
0x0b63d5d801708420, 0x8f227ca8f37ffaec,
0x0256278670887c24, 0x107e14877dbf540b,
0x32c19f2786ac1c05, 0x1df5b12bb4bc9c61,
0xc0cac129d0d4c4e2, 0x9fdb52ee9800b001,
0x31f601d5d31c48c4, 0x72ff3c0928bcaec7,
0xd99264421147eb03, 0x535a2d6d38aefcfe,
0x6ba8b4454a916237, 0xfa39366eaae4719c,
0x10f00fd7bbb24b6f, 0x5bd23185c76c84d4,
0xb22c3d7e1b00d33f, 0x3efc20aa6bc830a8,
0xd61c2503fe639144, 0x30ce625441eb92d3,
0xe5d34cf359e93100, 0xa8e5aa13f2b9f7a5,
0x5c2b8d851ca254a6, 0x68fb6c5e8b0d5fdf,
0xc7ea4872c96b83ae, 0x6dd5d376f4392382,
0x1be88681aaa9792f, 0xfef465ee1b6c10d9,
0x1f98b65ed43fcb2e, 0x4d1ca11eb6e9a9c9,
0x7808e902b3857d0b, 0x171c9c4ea4607972,
0x58d66274850146df, 0x42b311c10d3981d1,
0x647fa8c621c41a4c, 0xf472771c66ddfedc,
0x338d27e3f847b46b, 0x6402ce3da97545ce,
0x5162db616fc38638, 0x9c83be97bc22a50e,
0x2d3d7478a78d5e72, 0xe621a9b938fd5397,
0x9454614eb0f81c45, 0x395fb6e742ed39b6,
0x77dd9179d06037bf, 0xc478d0fee4d2656d,
0x35d9d6cb772007af, 0x83a56e92c883f0f6,
0x27937453250c00a1, 0x27bd6ebc3a46a97d,
0x9f543bf784342d51, 0xd158f38c48b0ed52,
0x8dd8537c045f66b4, 0x846a57230226f6d5,
0x6b13939e0c4e7cdf, 0xfca25425d8176758,
0x92e5fc6cd52788e6, 0x9992e13d7a739170,
0x518246f7a199e8ea, 0xf104c2a71b9979c7,
0x86b3ffaabea4768f, 0x6388061cf3e351ad,
0x09d9b5295de5bbb5, 0x38bf1638c2599e92,
0x1d759846499e148d, 0x4c0ff015e5f96ef4,
0xa41a94cfa270f565, 0x42d76f9cb2326c0b,
0x0cf385dd3c9c23ba, 0x0508a6c7508d6e7a,
0x337523aabbe6cf8d, 0x646bb14001d42b12,
0xc178729d138adc74, 0xf900ef4491f24086,
0xee1a90d334bb5ac4, 0x9755c92247301a50,
0xb999bf7c4ff1b610, 0x6aeeb2f3b21e8fc9,
0x0fa8084cf91ac6ff, 0x10d226cf136e6189,
0xd302057a07d4fb21, 0x5f03800e20a0fcc3,
0x80118d4ae46bd210, 0x58ab61a522843733,
0x51edd575c5432a4b, 0x94ee6ff67f9197f7,
0x765669e0e5e8157b, 0xa5347830737132f0,
0x3ba485a69f01510c, 0x0b247d7b957a01c3,
0x1b3d63449fd807dc, 0x0fdc4721c30ad743,
0x8b535ed3829b2b14, 0xee41d0cad65d232c,
0xe6a99ed97a6a982f, 0x65ac6194c202003d,
0x692accf3a70573eb, 0xcc3c02c3e200d5af,
0x0d419e8b325914a3, 0x320f160f42c25e40,
0x00710d647a51fe7a, 0x3c947692330aed60,
0x9288aa280d355a7a, 0xa1806a9b791d1696,
0x5d60e38496763da1, 0x6c69e22e613fd0f4,
0x977fc2a5aadffb17, 0xfb7bd063fc5a94ba,
0x460c17992cbaece1, 0xf7822c5444d3297f,
0x344a9790c69b74aa, 0xb80a42e6cae09dce,
0x1b1361eaf2b1e757, 0xd84c1e758e236f01,
0x88e0b7be347627cc, 0x45246009b7a99490,
0x8011c6dd3fe50472, 0xc341d682bffb99d7,
0x2511be93808e2d15, 0xd5bc13d7fd739840,
0x2a3cd030679ae1ec, 0x8ad9898a4b9ee157,
0x3245fef0a8eaf521, 0x3d6d8dbbb427d2b0,
0x1ed146d8968b3981, 0x0c6a28bf7d45f3fc,
0x4a1fd3dbcee3c561, 0x4210ff6a476bf67e,
0xa559cce0d9199aac, 0xde39d47ef3723380,
0xe5b69d848ce42e35, 0xefa24296f8e79f52,
0x70190b59db9a5afc, 0x26f166cdb211e7bf,
0x4deaf2df3c6b8ef5, 0xf171dbdd670f1017,
0xb9059b05e9420d90, 0x2f0da855c9388754,
0x611d5e9ab77949cc, 0x2912038ac01163f4,
0x0231df50402b2fba, 0x45660fc4f3245f58,
0xb91cc97c7c8dac50, 0xb72d2aafe4953427,
0xfa6463f87e813d6b, 0x4515f7ee95d5c6a2,
0x1310e1c1a48d21c3, 0xad48a7810cdd8544,
0x4d5bdfefd5c9e631, 0xa43ed43f1fdcb7de,
0xe70cfc8fe1ee9626, 0xef4711b0d8dda442,
0xb80dd9bd4dab6c93, 0xa23be08d31ba4d93,
0x9b37db9d0335a39c, 0x494b6f870f5cfebc,
0x6d1b3c1149dda943, 0x372c943a518c1093,
0xad27af45e77c09c4, 0x3b6f92b646044604,
0xac2917909f5fcf4f, 0x2069a60e977e5557,
0x353a469e71014de5, 0x24be356281f55c15,
0x2b6d710ba8e9adea, 0x404ad1751c749c29,
0xed7311bf23d7f185, 0xba4f6976b4acc43e,
0x32d7198d2bc39000, 0xee667019014d6e01,
0x494ef3e128d14c83, 0x1f95a152baecd6be,
0x201648dff1f483a5, 0x68c28550c8384af6,
0x5fc834a6824a7f48, 0x7cd06cb7365eaf28,
0xd82bbd95e9b30909, 0x234f0d1694c53f6d,
0xd2fb7f4a96d83f4a, 0xff0d5da83acac05e,
0xf8f6b97f5585080a, 0x74236084be57b95b,
0xa25e40c03bbc36ad, 0x6b6e5c14ce88465b,
0x4378ffe93e1528c5, 0x94ca92a17118e2d2,
}

View File

@ -1,14 +0,0 @@
// Copyright 2014-2017 Ulrich Kunitz. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
/*
Package hash provides rolling hashes.
Rolling hashes have to be used for maintaining the positions of n-byte
sequences in the dictionary buffer.
The package provides currently the Rabin-Karp rolling hash and a Cyclic
Polynomial hash. Both support the Hashes method to be used with an interface.
*/
package hash

View File

@ -1,66 +0,0 @@
// Copyright 2014-2017 Ulrich Kunitz. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
package hash
// A is the default constant for Robin-Karp rolling hash. This is a random
// prime.
const A = 0x97b548add41d5da1
// RabinKarp supports the computation of a rolling hash.
type RabinKarp struct {
A uint64
// a^n
aOldest uint64
h uint64
p []byte
i int
}
// NewRabinKarp creates a new RabinKarp value. The argument n defines the
// length of the byte sequence to be hashed. The default constant will will be
// used.
func NewRabinKarp(n int) *RabinKarp {
return NewRabinKarpConst(n, A)
}
// NewRabinKarpConst creates a new RabinKarp value. The argument n defines the
// length of the byte sequence to be hashed. The argument a provides the
// constant used to compute the hash.
func NewRabinKarpConst(n int, a uint64) *RabinKarp {
if n <= 0 {
panic("number of bytes n must be positive")
}
aOldest := uint64(1)
// There are faster methods. For the small n required by the LZMA
// compressor O(n) is sufficient.
for i := 0; i < n; i++ {
aOldest *= a
}
return &RabinKarp{
A: a, aOldest: aOldest,
p: make([]byte, 0, n),
}
}
// Len returns the length of the byte sequence.
func (r *RabinKarp) Len() int {
return cap(r.p)
}
// RollByte computes the hash after x has been added.
func (r *RabinKarp) RollByte(x byte) uint64 {
if len(r.p) < cap(r.p) {
r.h += uint64(x)
r.h *= r.A
r.p = append(r.p, x)
} else {
r.h -= uint64(r.p[r.i]) * r.aOldest
r.h += uint64(x)
r.h *= r.A
r.p[r.i] = x
r.i = (r.i + 1) % cap(r.p)
}
return r.h
}

View File

@ -1,29 +0,0 @@
// Copyright 2014-2017 Ulrich Kunitz. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
package hash
// Roller provides an interface for rolling hashes. The hash value will become
// valid after hash has been called Len times.
type Roller interface {
Len() int
RollByte(x byte) uint64
}
// Hashes computes all hash values for the array p. Note that the state of the
// roller is changed.
func Hashes(r Roller, p []byte) []uint64 {
n := r.Len()
if len(p) < n {
return nil
}
h := make([]uint64, len(p)-n+1)
for i := 0; i < n-1; i++ {
r.RollByte(p[i])
}
for i := range h {
h[i] = r.RollByte(p[i+n-1])
}
return h
}

View File

@ -1,457 +0,0 @@
// Copyright 2014-2017 Ulrich Kunitz. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
// Package xlog provides a simple logging package that allows to disable
// certain message categories. It defines a type, Logger, with multiple
// methods for formatting output. The package has also a predefined
// 'standard' Logger accessible through helper function Print[f|ln],
// Fatal[f|ln], Panic[f|ln], Warn[f|ln], Print[f|ln] and Debug[f|ln]
// that are easier to use then creating a Logger manually. That logger
// writes to standard error and prints the date and time of each logged
// message, which can be configured using the function SetFlags.
//
// The Fatal functions call os.Exit(1) after the message is output
// unless not suppressed by the flags. The Panic functions call panic
// after the writing the log message unless suppressed.
package xlog
import (
"fmt"
"io"
"os"
"runtime"
"sync"
"time"
)
// The flags define what information is prefixed to each log entry
// generated by the Logger. The Lno* versions allow the suppression of
// specific output. The bits are or'ed together to control what will be
// printed. There is no control over the order of the items printed and
// the format. The full format is:
//
// 2009-01-23 01:23:23.123123 /a/b/c/d.go:23: message
//
const (
Ldate = 1 << iota // the date: 2009-01-23
Ltime // the time: 01:23:23
Lmicroseconds // microsecond resolution: 01:23:23.123123
Llongfile // full file name and line number: /a/b/c/d.go:23
Lshortfile // final file name element and line number: d.go:23
Lnopanic // suppresses output from Panic[f|ln] but not the panic call
Lnofatal // suppresses output from Fatal[f|ln] but not the exit
Lnowarn // suppresses output from Warn[f|ln]
Lnoprint // suppresses output from Print[f|ln]
Lnodebug // suppresses output from Debug[f|ln]
// initial values for the standard logger
Lstdflags = Ldate | Ltime | Lnodebug
)
// A Logger represents an active logging object that generates lines of
// output to an io.Writer. Each logging operation if not suppressed
// makes a single call to the Writer's Write method. A Logger can be
// used simultaneously from multiple goroutines; it guarantees to
// serialize access to the Writer.
type Logger struct {
mu sync.Mutex // ensures atomic writes; and protects the following
// fields
prefix string // prefix to write at beginning of each line
flag int // properties
out io.Writer // destination for output
buf []byte // for accumulating text to write
}
// New creates a new Logger. The out argument sets the destination to
// which the log output will be written. The prefix appears at the
// beginning of each log line. The flag argument defines the logging
// properties.
func New(out io.Writer, prefix string, flag int) *Logger {
return &Logger{out: out, prefix: prefix, flag: flag}
}
// std is the standard logger used by the package scope functions.
var std = New(os.Stderr, "", Lstdflags)
// itoa converts the integer to ASCII. A negative widths will avoid
// zero-padding. The function supports only non-negative integers.
func itoa(buf *[]byte, i int, wid int) {
var u = uint(i)
if u == 0 && wid <= 1 {
*buf = append(*buf, '0')
return
}
var b [32]byte
bp := len(b)
for ; u > 0 || wid > 0; u /= 10 {
bp--
wid--
b[bp] = byte(u%10) + '0'
}
*buf = append(*buf, b[bp:]...)
}
// formatHeader puts the header into the buf field of the buffer.
func (l *Logger) formatHeader(t time.Time, file string, line int) {
l.buf = append(l.buf, l.prefix...)
if l.flag&(Ldate|Ltime|Lmicroseconds) != 0 {
if l.flag&Ldate != 0 {
year, month, day := t.Date()
itoa(&l.buf, year, 4)
l.buf = append(l.buf, '-')
itoa(&l.buf, int(month), 2)
l.buf = append(l.buf, '-')
itoa(&l.buf, day, 2)
l.buf = append(l.buf, ' ')
}
if l.flag&(Ltime|Lmicroseconds) != 0 {
hour, min, sec := t.Clock()
itoa(&l.buf, hour, 2)
l.buf = append(l.buf, ':')
itoa(&l.buf, min, 2)
l.buf = append(l.buf, ':')
itoa(&l.buf, sec, 2)
if l.flag&Lmicroseconds != 0 {
l.buf = append(l.buf, '.')
itoa(&l.buf, t.Nanosecond()/1e3, 6)
}
l.buf = append(l.buf, ' ')
}
}
if l.flag&(Lshortfile|Llongfile) != 0 {
if l.flag&Lshortfile != 0 {
short := file
for i := len(file) - 1; i > 0; i-- {
if file[i] == '/' {
short = file[i+1:]
break
}
}
file = short
}
l.buf = append(l.buf, file...)
l.buf = append(l.buf, ':')
itoa(&l.buf, line, -1)
l.buf = append(l.buf, ": "...)
}
}
func (l *Logger) output(calldepth int, now time.Time, s string) error {
var file string
var line int
if l.flag&(Lshortfile|Llongfile) != 0 {
l.mu.Unlock()
var ok bool
_, file, line, ok = runtime.Caller(calldepth)
if !ok {
file = "???"
line = 0
}
l.mu.Lock()
}
l.buf = l.buf[:0]
l.formatHeader(now, file, line)
l.buf = append(l.buf, s...)
if len(s) == 0 || s[len(s)-1] != '\n' {
l.buf = append(l.buf, '\n')
}
_, err := l.out.Write(l.buf)
return err
}
// Output writes the string s with the header controlled by the flags to
// the l.out writer. A newline will be appended if s doesn't end in a
// newline. Calldepth is used to recover the PC, although all current
// calls of Output use the call depth 2. Access to the function is serialized.
func (l *Logger) Output(calldepth, noflag int, v ...interface{}) error {
now := time.Now()
l.mu.Lock()
defer l.mu.Unlock()
if l.flag&noflag != 0 {
return nil
}
s := fmt.Sprint(v...)
return l.output(calldepth+1, now, s)
}
// Outputf works like output but formats the output like Printf.
func (l *Logger) Outputf(calldepth int, noflag int, format string, v ...interface{}) error {
now := time.Now()
l.mu.Lock()
defer l.mu.Unlock()
if l.flag&noflag != 0 {
return nil
}
s := fmt.Sprintf(format, v...)
return l.output(calldepth+1, now, s)
}
// Outputln works like output but formats the output like Println.
func (l *Logger) Outputln(calldepth int, noflag int, v ...interface{}) error {
now := time.Now()
l.mu.Lock()
defer l.mu.Unlock()
if l.flag&noflag != 0 {
return nil
}
s := fmt.Sprintln(v...)
return l.output(calldepth+1, now, s)
}
// Panic prints the message like Print and calls panic. The printing
// might be suppressed by the flag Lnopanic.
func (l *Logger) Panic(v ...interface{}) {
l.Output(2, Lnopanic, v...)
s := fmt.Sprint(v...)
panic(s)
}
// Panic prints the message like Print and calls panic. The printing
// might be suppressed by the flag Lnopanic.
func Panic(v ...interface{}) {
std.Output(2, Lnopanic, v...)
s := fmt.Sprint(v...)
panic(s)
}
// Panicf prints the message like Printf and calls panic. The printing
// might be suppressed by the flag Lnopanic.
func (l *Logger) Panicf(format string, v ...interface{}) {
l.Outputf(2, Lnopanic, format, v...)
s := fmt.Sprintf(format, v...)
panic(s)
}
// Panicf prints the message like Printf and calls panic. The printing
// might be suppressed by the flag Lnopanic.
func Panicf(format string, v ...interface{}) {
std.Outputf(2, Lnopanic, format, v...)
s := fmt.Sprintf(format, v...)
panic(s)
}
// Panicln prints the message like Println and calls panic. The printing
// might be suppressed by the flag Lnopanic.
func (l *Logger) Panicln(v ...interface{}) {
l.Outputln(2, Lnopanic, v...)
s := fmt.Sprintln(v...)
panic(s)
}
// Panicln prints the message like Println and calls panic. The printing
// might be suppressed by the flag Lnopanic.
func Panicln(v ...interface{}) {
std.Outputln(2, Lnopanic, v...)
s := fmt.Sprintln(v...)
panic(s)
}
// Fatal prints the message like Print and calls os.Exit(1). The
// printing might be suppressed by the flag Lnofatal.
func (l *Logger) Fatal(v ...interface{}) {
l.Output(2, Lnofatal, v...)
os.Exit(1)
}
// Fatal prints the message like Print and calls os.Exit(1). The
// printing might be suppressed by the flag Lnofatal.
func Fatal(v ...interface{}) {
std.Output(2, Lnofatal, v...)
os.Exit(1)
}
// Fatalf prints the message like Printf and calls os.Exit(1). The
// printing might be suppressed by the flag Lnofatal.
func (l *Logger) Fatalf(format string, v ...interface{}) {
l.Outputf(2, Lnofatal, format, v...)
os.Exit(1)
}
// Fatalf prints the message like Printf and calls os.Exit(1). The
// printing might be suppressed by the flag Lnofatal.
func Fatalf(format string, v ...interface{}) {
std.Outputf(2, Lnofatal, format, v...)
os.Exit(1)
}
// Fatalln prints the message like Println and calls os.Exit(1). The
// printing might be suppressed by the flag Lnofatal.
func (l *Logger) Fatalln(format string, v ...interface{}) {
l.Outputln(2, Lnofatal, v...)
os.Exit(1)
}
// Fatalln prints the message like Println and calls os.Exit(1). The
// printing might be suppressed by the flag Lnofatal.
func Fatalln(format string, v ...interface{}) {
std.Outputln(2, Lnofatal, v...)
os.Exit(1)
}
// Warn prints the message like Print. The printing might be suppressed
// by the flag Lnowarn.
func (l *Logger) Warn(v ...interface{}) {
l.Output(2, Lnowarn, v...)
}
// Warn prints the message like Print. The printing might be suppressed
// by the flag Lnowarn.
func Warn(v ...interface{}) {
std.Output(2, Lnowarn, v...)
}
// Warnf prints the message like Printf. The printing might be suppressed
// by the flag Lnowarn.
func (l *Logger) Warnf(format string, v ...interface{}) {
l.Outputf(2, Lnowarn, format, v...)
}
// Warnf prints the message like Printf. The printing might be suppressed
// by the flag Lnowarn.
func Warnf(format string, v ...interface{}) {
std.Outputf(2, Lnowarn, format, v...)
}
// Warnln prints the message like Println. The printing might be suppressed
// by the flag Lnowarn.
func (l *Logger) Warnln(v ...interface{}) {
l.Outputln(2, Lnowarn, v...)
}
// Warnln prints the message like Println. The printing might be suppressed
// by the flag Lnowarn.
func Warnln(v ...interface{}) {
std.Outputln(2, Lnowarn, v...)
}
// Print prints the message like fmt.Print. The printing might be suppressed
// by the flag Lnoprint.
func (l *Logger) Print(v ...interface{}) {
l.Output(2, Lnoprint, v...)
}
// Print prints the message like fmt.Print. The printing might be suppressed
// by the flag Lnoprint.
func Print(v ...interface{}) {
std.Output(2, Lnoprint, v...)
}
// Printf prints the message like fmt.Printf. The printing might be suppressed
// by the flag Lnoprint.
func (l *Logger) Printf(format string, v ...interface{}) {
l.Outputf(2, Lnoprint, format, v...)
}
// Printf prints the message like fmt.Printf. The printing might be suppressed
// by the flag Lnoprint.
func Printf(format string, v ...interface{}) {
std.Outputf(2, Lnoprint, format, v...)
}
// Println prints the message like fmt.Println. The printing might be
// suppressed by the flag Lnoprint.
func (l *Logger) Println(v ...interface{}) {
l.Outputln(2, Lnoprint, v...)
}
// Println prints the message like fmt.Println. The printing might be
// suppressed by the flag Lnoprint.
func Println(v ...interface{}) {
std.Outputln(2, Lnoprint, v...)
}
// Debug prints the message like Print. The printing might be suppressed
// by the flag Lnodebug.
func (l *Logger) Debug(v ...interface{}) {
l.Output(2, Lnodebug, v...)
}
// Debug prints the message like Print. The printing might be suppressed
// by the flag Lnodebug.
func Debug(v ...interface{}) {
std.Output(2, Lnodebug, v...)
}
// Debugf prints the message like Printf. The printing might be suppressed
// by the flag Lnodebug.
func (l *Logger) Debugf(format string, v ...interface{}) {
l.Outputf(2, Lnodebug, format, v...)
}
// Debugf prints the message like Printf. The printing might be suppressed
// by the flag Lnodebug.
func Debugf(format string, v ...interface{}) {
std.Outputf(2, Lnodebug, format, v...)
}
// Debugln prints the message like Println. The printing might be suppressed
// by the flag Lnodebug.
func (l *Logger) Debugln(v ...interface{}) {
l.Outputln(2, Lnodebug, v...)
}
// Debugln prints the message like Println. The printing might be suppressed
// by the flag Lnodebug.
func Debugln(v ...interface{}) {
std.Outputln(2, Lnodebug, v...)
}
// Flags returns the current flags used by the logger.
func (l *Logger) Flags() int {
l.mu.Lock()
defer l.mu.Unlock()
return l.flag
}
// Flags returns the current flags used by the standard logger.
func Flags() int {
return std.Flags()
}
// SetFlags sets the flags of the logger.
func (l *Logger) SetFlags(flag int) {
l.mu.Lock()
defer l.mu.Unlock()
l.flag = flag
}
// SetFlags sets the flags for the standard logger.
func SetFlags(flag int) {
std.SetFlags(flag)
}
// Prefix returns the prefix used by the logger.
func (l *Logger) Prefix() string {
l.mu.Lock()
defer l.mu.Unlock()
return l.prefix
}
// Prefix returns the prefix used by the standard logger of the package.
func Prefix() string {
return std.Prefix()
}
// SetPrefix sets the prefix for the logger.
func (l *Logger) SetPrefix(prefix string) {
l.mu.Lock()
defer l.mu.Unlock()
l.prefix = prefix
}
// SetPrefix sets the prefix of the standard logger of the package.
func SetPrefix(prefix string) {
std.SetPrefix(prefix)
}
// SetOutput sets the output of the logger.
func (l *Logger) SetOutput(w io.Writer) {
l.mu.Lock()
defer l.mu.Unlock()
l.out = w
}
// SetOutput sets the output for the standard logger of the package.
func SetOutput(w io.Writer) {
std.SetOutput(w)
}

View File

@ -1,523 +0,0 @@
// Copyright 2014-2017 Ulrich Kunitz. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
package lzma
import (
"bufio"
"errors"
"fmt"
"io"
"unicode"
)
// node represents a node in the binary tree.
type node struct {
// x is the search value
x uint32
// p parent node
p uint32
// l left child
l uint32
// r right child
r uint32
}
// wordLen is the number of bytes represented by the v field of a node.
const wordLen = 4
// binTree supports the identification of the next operation based on a
// binary tree.
//
// Nodes will be identified by their index into the ring buffer.
type binTree struct {
dict *encoderDict
// ring buffer of nodes
node []node
// absolute offset of the entry for the next node. Position 4
// byte larger.
hoff int64
// front position in the node ring buffer
front uint32
// index of the root node
root uint32
// current x value
x uint32
// preallocated array
data []byte
}
// null represents the nonexistent index. We can't use zero because it
// would always exist or we would need to decrease the index for each
// reference.
const null uint32 = 1<<32 - 1
// newBinTree initializes the binTree structure. The capacity defines
// the size of the buffer and defines the maximum distance for which
// matches will be found.
func newBinTree(capacity int) (t *binTree, err error) {
if capacity < 1 {
return nil, errors.New(
"newBinTree: capacity must be larger than zero")
}
if int64(capacity) >= int64(null) {
return nil, errors.New(
"newBinTree: capacity must less 2^{32}-1")
}
t = &binTree{
node: make([]node, capacity),
hoff: -int64(wordLen),
root: null,
data: make([]byte, maxMatchLen),
}
return t, nil
}
func (t *binTree) SetDict(d *encoderDict) { t.dict = d }
// WriteByte writes a single byte into the binary tree.
func (t *binTree) WriteByte(c byte) error {
t.x = (t.x << 8) | uint32(c)
t.hoff++
if t.hoff < 0 {
return nil
}
v := t.front
if int64(v) < t.hoff {
// We are overwriting old nodes stored in the tree.
t.remove(v)
}
t.node[v].x = t.x
t.add(v)
t.front++
if int64(t.front) >= int64(len(t.node)) {
t.front = 0
}
return nil
}
// Writes writes a sequence of bytes into the binTree structure.
func (t *binTree) Write(p []byte) (n int, err error) {
for _, c := range p {
t.WriteByte(c)
}
return len(p), nil
}
// add puts the node v into the tree. The node must not be part of the
// tree before.
func (t *binTree) add(v uint32) {
vn := &t.node[v]
// Set left and right to null indices.
vn.l, vn.r = null, null
// If the binary tree is empty make v the root.
if t.root == null {
t.root = v
vn.p = null
return
}
x := vn.x
p := t.root
// Search for the right leave link and add the new node.
for {
pn := &t.node[p]
if x <= pn.x {
if pn.l == null {
pn.l = v
vn.p = p
return
}
p = pn.l
} else {
if pn.r == null {
pn.r = v
vn.p = p
return
}
p = pn.r
}
}
}
// parent returns the parent node index of v and the pointer to v value
// in the parent.
func (t *binTree) parent(v uint32) (p uint32, ptr *uint32) {
if t.root == v {
return null, &t.root
}
p = t.node[v].p
if t.node[p].l == v {
ptr = &t.node[p].l
} else {
ptr = &t.node[p].r
}
return
}
// Remove node v.
func (t *binTree) remove(v uint32) {
vn := &t.node[v]
p, ptr := t.parent(v)
l, r := vn.l, vn.r
if l == null {
// Move the right child up.
*ptr = r
if r != null {
t.node[r].p = p
}
return
}
if r == null {
// Move the left child up.
*ptr = l
t.node[l].p = p
return
}
// Search the in-order predecessor u.
un := &t.node[l]
ur := un.r
if ur == null {
// In order predecessor is l. Move it up.
un.r = r
t.node[r].p = l
un.p = p
*ptr = l
return
}
var u uint32
for {
// Look for the max value in the tree where l is root.
u = ur
ur = t.node[u].r
if ur == null {
break
}
}
// replace u with ul
un = &t.node[u]
ul := un.l
up := un.p
t.node[up].r = ul
if ul != null {
t.node[ul].p = up
}
// replace v by u
un.l, un.r = l, r
t.node[l].p = u
t.node[r].p = u
*ptr = u
un.p = p
}
// search looks for the node that have the value x or for the nodes that
// brace it. The node highest in the tree with the value x will be
// returned. All other nodes with the same value live in left subtree of
// the returned node.
func (t *binTree) search(v uint32, x uint32) (a, b uint32) {
a, b = null, null
if v == null {
return
}
for {
vn := &t.node[v]
if x <= vn.x {
if x == vn.x {
return v, v
}
b = v
if vn.l == null {
return
}
v = vn.l
} else {
a = v
if vn.r == null {
return
}
v = vn.r
}
}
}
// max returns the node with maximum value in the subtree with v as
// root.
func (t *binTree) max(v uint32) uint32 {
if v == null {
return null
}
for {
r := t.node[v].r
if r == null {
return v
}
v = r
}
}
// min returns the node with the minimum value in the subtree with v as
// root.
func (t *binTree) min(v uint32) uint32 {
if v == null {
return null
}
for {
l := t.node[v].l
if l == null {
return v
}
v = l
}
}
// pred returns the in-order predecessor of node v.
func (t *binTree) pred(v uint32) uint32 {
if v == null {
return null
}
u := t.max(t.node[v].l)
if u != null {
return u
}
for {
p := t.node[v].p
if p == null {
return null
}
if t.node[p].r == v {
return p
}
v = p
}
}
// succ returns the in-order successor of node v.
func (t *binTree) succ(v uint32) uint32 {
if v == null {
return null
}
u := t.min(t.node[v].r)
if u != null {
return u
}
for {
p := t.node[v].p
if p == null {
return null
}
if t.node[p].l == v {
return p
}
v = p
}
}
// xval converts the first four bytes of a into an 32-bit unsigned
// integer in big-endian order.
func xval(a []byte) uint32 {
var x uint32
switch len(a) {
default:
x |= uint32(a[3])
fallthrough
case 3:
x |= uint32(a[2]) << 8
fallthrough
case 2:
x |= uint32(a[1]) << 16
fallthrough
case 1:
x |= uint32(a[0]) << 24
case 0:
}
return x
}
// dumpX converts value x into a four-letter string.
func dumpX(x uint32) string {
a := make([]byte, 4)
for i := 0; i < 4; i++ {
c := byte(x >> uint((3-i)*8))
if unicode.IsGraphic(rune(c)) {
a[i] = c
} else {
a[i] = '.'
}
}
return string(a)
}
// dumpNode writes a representation of the node v into the io.Writer.
func (t *binTree) dumpNode(w io.Writer, v uint32, indent int) {
if v == null {
return
}
vn := &t.node[v]
t.dumpNode(w, vn.r, indent+2)
for i := 0; i < indent; i++ {
fmt.Fprint(w, " ")
}
if vn.p == null {
fmt.Fprintf(w, "node %d %q parent null\n", v, dumpX(vn.x))
} else {
fmt.Fprintf(w, "node %d %q parent %d\n", v, dumpX(vn.x), vn.p)
}
t.dumpNode(w, vn.l, indent+2)
}
// dump prints a representation of the binary tree into the writer.
func (t *binTree) dump(w io.Writer) error {
bw := bufio.NewWriter(w)
t.dumpNode(bw, t.root, 0)
return bw.Flush()
}
func (t *binTree) distance(v uint32) int {
dist := int(t.front) - int(v)
if dist <= 0 {
dist += len(t.node)
}
return dist
}
type matchParams struct {
rep [4]uint32
// length when match will be accepted
nAccept int
// nodes to check
check int
// finish if length get shorter
stopShorter bool
}
func (t *binTree) match(m match, distIter func() (int, bool), p matchParams,
) (r match, checked int, accepted bool) {
buf := &t.dict.buf
for {
if checked >= p.check {
return m, checked, true
}
dist, ok := distIter()
if !ok {
return m, checked, false
}
checked++
if m.n > 0 {
i := buf.rear - dist + m.n - 1
if i < 0 {
i += len(buf.data)
} else if i >= len(buf.data) {
i -= len(buf.data)
}
if buf.data[i] != t.data[m.n-1] {
if p.stopShorter {
return m, checked, false
}
continue
}
}
n := buf.matchLen(dist, t.data)
switch n {
case 0:
if p.stopShorter {
return m, checked, false
}
continue
case 1:
if uint32(dist-minDistance) != p.rep[0] {
continue
}
}
if n < m.n || (n == m.n && int64(dist) >= m.distance) {
continue
}
m = match{int64(dist), n}
if n >= p.nAccept {
return m, checked, true
}
}
}
func (t *binTree) NextOp(rep [4]uint32) operation {
// retrieve maxMatchLen data
n, _ := t.dict.buf.Peek(t.data[:maxMatchLen])
if n == 0 {
panic("no data in buffer")
}
t.data = t.data[:n]
var (
m match
x, u, v uint32
iterPred, iterSucc func() (int, bool)
)
p := matchParams{
rep: rep,
nAccept: maxMatchLen,
check: 32,
}
i := 4
iterSmall := func() (dist int, ok bool) {
i--
if i <= 0 {
return 0, false
}
return i, true
}
m, checked, accepted := t.match(m, iterSmall, p)
if accepted {
goto end
}
p.check -= checked
x = xval(t.data)
u, v = t.search(t.root, x)
if u == v && len(t.data) == 4 {
iter := func() (dist int, ok bool) {
if u == null {
return 0, false
}
dist = t.distance(u)
u, v = t.search(t.node[u].l, x)
if u != v {
u = null
}
return dist, true
}
m, _, _ = t.match(m, iter, p)
goto end
}
p.stopShorter = true
iterSucc = func() (dist int, ok bool) {
if v == null {
return 0, false
}
dist = t.distance(v)
v = t.succ(v)
return dist, true
}
m, checked, accepted = t.match(m, iterSucc, p)
if accepted {
goto end
}
p.check -= checked
iterPred = func() (dist int, ok bool) {
if u == null {
return 0, false
}
dist = t.distance(u)
u = t.pred(u)
return dist, true
}
m, _, _ = t.match(m, iterPred, p)
end:
if m.n == 0 {
return lit{t.data[0]}
}
return m
}

View File

@ -1,45 +0,0 @@
// Copyright 2014-2017 Ulrich Kunitz. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
package lzma
/* Naming conventions follows the CodeReviewComments in the Go Wiki. */
// ntz32Const is used by the functions NTZ and NLZ.
const ntz32Const = 0x04d7651f
// ntz32Table is a helper table for de Bruijn algorithm by Danny Dubé.
// See Henry S. Warren, Jr. "Hacker's Delight" section 5-1 figure 5-26.
var ntz32Table = [32]int8{
0, 1, 2, 24, 3, 19, 6, 25,
22, 4, 20, 10, 16, 7, 12, 26,
31, 23, 18, 5, 21, 9, 15, 11,
30, 17, 8, 14, 29, 13, 28, 27,
}
// ntz32 computes the number of trailing zeros for an unsigned 32-bit integer.
func ntz32(x uint32) int {
if x == 0 {
return 32
}
x = (x & -x) * ntz32Const
return int(ntz32Table[x>>27])
}
// nlz32 computes the number of leading zeros for an unsigned 32-bit integer.
func nlz32(x uint32) int {
// Smear left most bit to the right
x |= x >> 1
x |= x >> 2
x |= x >> 4
x |= x >> 8
x |= x >> 16
// Use ntz mechanism to calculate nlz.
x++
if x == 0 {
return 0
}
x *= ntz32Const
return 32 - int(ntz32Table[x>>27])
}

View File

@ -1,39 +0,0 @@
// Copyright 2014-2017 Ulrich Kunitz. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
package lzma
import (
"errors"
"io"
)
// breader provides the ReadByte function for a Reader. It doesn't read
// more data from the reader than absolutely necessary.
type breader struct {
io.Reader
// helper slice to save allocations
p []byte
}
// ByteReader converts an io.Reader into an io.ByteReader.
func ByteReader(r io.Reader) io.ByteReader {
br, ok := r.(io.ByteReader)
if !ok {
return &breader{r, make([]byte, 1)}
}
return br
}
// ReadByte read byte function.
func (r *breader) ReadByte() (c byte, err error) {
n, err := r.Reader.Read(r.p)
if n < 1 {
if err == nil {
err = errors.New("breader.ReadByte: no data")
}
return 0, err
}
return r.p[0], nil
}

View File

@ -1,171 +0,0 @@
// Copyright 2014-2017 Ulrich Kunitz. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
package lzma
import (
"errors"
)
// buffer provides a circular buffer of bytes. If the front index equals
// the rear index the buffer is empty. As a consequence front cannot be
// equal rear for a full buffer. So a full buffer has a length that is
// one byte less the the length of the data slice.
type buffer struct {
data []byte
front int
rear int
}
// newBuffer creates a buffer with the given size.
func newBuffer(size int) *buffer {
return &buffer{data: make([]byte, size+1)}
}
// Cap returns the capacity of the buffer.
func (b *buffer) Cap() int {
return len(b.data) - 1
}
// Resets the buffer. The front and rear index are set to zero.
func (b *buffer) Reset() {
b.front = 0
b.rear = 0
}
// Buffered returns the number of bytes buffered.
func (b *buffer) Buffered() int {
delta := b.front - b.rear
if delta < 0 {
delta += len(b.data)
}
return delta
}
// Available returns the number of bytes available for writing.
func (b *buffer) Available() int {
delta := b.rear - 1 - b.front
if delta < 0 {
delta += len(b.data)
}
return delta
}
// addIndex adds a non-negative integer to the index i and returns the
// resulting index. The function takes care of wrapping the index as
// well as potential overflow situations.
func (b *buffer) addIndex(i int, n int) int {
// subtraction of len(b.data) prevents overflow
i += n - len(b.data)
if i < 0 {
i += len(b.data)
}
return i
}
// Read reads bytes from the buffer into p and returns the number of
// bytes read. The function never returns an error but might return less
// data than requested.
func (b *buffer) Read(p []byte) (n int, err error) {
n, err = b.Peek(p)
b.rear = b.addIndex(b.rear, n)
return n, err
}
// Peek reads bytes from the buffer into p without changing the buffer.
// Peek will never return an error but might return less data than
// requested.
func (b *buffer) Peek(p []byte) (n int, err error) {
m := b.Buffered()
n = len(p)
if m < n {
n = m
p = p[:n]
}
k := copy(p, b.data[b.rear:])
if k < n {
copy(p[k:], b.data)
}
return n, nil
}
// Discard skips the n next bytes to read from the buffer, returning the
// bytes discarded.
//
// If Discards skips fewer than n bytes, it returns an error.
func (b *buffer) Discard(n int) (discarded int, err error) {
if n < 0 {
return 0, errors.New("buffer.Discard: negative argument")
}
m := b.Buffered()
if m < n {
n = m
err = errors.New(
"buffer.Discard: discarded less bytes then requested")
}
b.rear = b.addIndex(b.rear, n)
return n, err
}
// ErrNoSpace indicates that there is insufficient space for the Write
// operation.
var ErrNoSpace = errors.New("insufficient space")
// Write puts data into the buffer. If less bytes are written than
// requested ErrNoSpace is returned.
func (b *buffer) Write(p []byte) (n int, err error) {
m := b.Available()
n = len(p)
if m < n {
n = m
p = p[:m]
err = ErrNoSpace
}
k := copy(b.data[b.front:], p)
if k < n {
copy(b.data, p[k:])
}
b.front = b.addIndex(b.front, n)
return n, err
}
// WriteByte writes a single byte into the buffer. The error ErrNoSpace
// is returned if no single byte is available in the buffer for writing.
func (b *buffer) WriteByte(c byte) error {
if b.Available() < 1 {
return ErrNoSpace
}
b.data[b.front] = c
b.front = b.addIndex(b.front, 1)
return nil
}
// prefixLen returns the length of the common prefix of a and b.
func prefixLen(a, b []byte) int {
if len(a) > len(b) {
a, b = b, a
}
for i, c := range a {
if b[i] != c {
return i
}
}
return len(a)
}
// matchLen returns the length of the common prefix for the given
// distance from the rear and the byte slice p.
func (b *buffer) matchLen(distance int, p []byte) int {
var n int
i := b.rear - distance
if i < 0 {
if n = prefixLen(p, b.data[len(b.data)+i:]); n < -i {
return n
}
p = p[n:]
i = 0
}
n += prefixLen(p, b.data[i:])
return n
}

View File

@ -1,37 +0,0 @@
// Copyright 2014-2017 Ulrich Kunitz. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
package lzma
import (
"errors"
"io"
)
// ErrLimit indicates that the limit of the LimitedByteWriter has been
// reached.
var ErrLimit = errors.New("limit reached")
// LimitedByteWriter provides a byte writer that can be written until a
// limit is reached. The field N provides the number of remaining
// bytes.
type LimitedByteWriter struct {
BW io.ByteWriter
N int64
}
// WriteByte writes a single byte to the limited byte writer. It returns
// ErrLimit if the limit has been reached. If the byte is successfully
// written the field N of the LimitedByteWriter will be decremented by
// one.
func (l *LimitedByteWriter) WriteByte(c byte) error {
if l.N <= 0 {
return ErrLimit
}
if err := l.BW.WriteByte(c); err != nil {
return err
}
l.N--
return nil
}

View File

@ -1,277 +0,0 @@
// Copyright 2014-2017 Ulrich Kunitz. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
package lzma
import (
"errors"
"fmt"
"io"
)
// decoder decodes a raw LZMA stream without any header.
type decoder struct {
// dictionary; the rear pointer of the buffer will be used for
// reading the data.
Dict *decoderDict
// decoder state
State *state
// range decoder
rd *rangeDecoder
// start stores the head value of the dictionary for the LZMA
// stream
start int64
// size of uncompressed data
size int64
// end-of-stream encountered
eos bool
// EOS marker found
eosMarker bool
}
// newDecoder creates a new decoder instance. The parameter size provides
// the expected byte size of the decompressed data. If the size is
// unknown use a negative value. In that case the decoder will look for
// a terminating end-of-stream marker.
func newDecoder(br io.ByteReader, state *state, dict *decoderDict, size int64) (d *decoder, err error) {
rd, err := newRangeDecoder(br)
if err != nil {
return nil, err
}
d = &decoder{
State: state,
Dict: dict,
rd: rd,
size: size,
start: dict.pos(),
}
return d, nil
}
// Reopen restarts the decoder with a new byte reader and a new size. Reopen
// resets the Decompressed counter to zero.
func (d *decoder) Reopen(br io.ByteReader, size int64) error {
var err error
if d.rd, err = newRangeDecoder(br); err != nil {
return err
}
d.start = d.Dict.pos()
d.size = size
d.eos = false
return nil
}
// decodeLiteral decodes a single literal from the LZMA stream.
func (d *decoder) decodeLiteral() (op operation, err error) {
litState := d.State.litState(d.Dict.byteAt(1), d.Dict.head)
match := d.Dict.byteAt(int(d.State.rep[0]) + 1)
s, err := d.State.litCodec.Decode(d.rd, d.State.state, match, litState)
if err != nil {
return nil, err
}
return lit{s}, nil
}
// errEOS indicates that an EOS marker has been found.
var errEOS = errors.New("EOS marker found")
// readOp decodes the next operation from the compressed stream. It
// returns the operation. If an explicit end of stream marker is
// identified the eos error is returned.
func (d *decoder) readOp() (op operation, err error) {
// Value of the end of stream (EOS) marker
const eosDist = 1<<32 - 1
state, state2, posState := d.State.states(d.Dict.head)
b, err := d.State.isMatch[state2].Decode(d.rd)
if err != nil {
return nil, err
}
if b == 0 {
// literal
op, err := d.decodeLiteral()
if err != nil {
return nil, err
}
d.State.updateStateLiteral()
return op, nil
}
b, err = d.State.isRep[state].Decode(d.rd)
if err != nil {
return nil, err
}
if b == 0 {
// simple match
d.State.rep[3], d.State.rep[2], d.State.rep[1] =
d.State.rep[2], d.State.rep[1], d.State.rep[0]
d.State.updateStateMatch()
// The length decoder returns the length offset.
n, err := d.State.lenCodec.Decode(d.rd, posState)
if err != nil {
return nil, err
}
// The dist decoder returns the distance offset. The actual
// distance is 1 higher.
d.State.rep[0], err = d.State.distCodec.Decode(d.rd, n)
if err != nil {
return nil, err
}
if d.State.rep[0] == eosDist {
d.eosMarker = true
return nil, errEOS
}
op = match{n: int(n) + minMatchLen,
distance: int64(d.State.rep[0]) + minDistance}
return op, nil
}
b, err = d.State.isRepG0[state].Decode(d.rd)
if err != nil {
return nil, err
}
dist := d.State.rep[0]
if b == 0 {
// rep match 0
b, err = d.State.isRepG0Long[state2].Decode(d.rd)
if err != nil {
return nil, err
}
if b == 0 {
d.State.updateStateShortRep()
op = match{n: 1, distance: int64(dist) + minDistance}
return op, nil
}
} else {
b, err = d.State.isRepG1[state].Decode(d.rd)
if err != nil {
return nil, err
}
if b == 0 {
dist = d.State.rep[1]
} else {
b, err = d.State.isRepG2[state].Decode(d.rd)
if err != nil {
return nil, err
}
if b == 0 {
dist = d.State.rep[2]
} else {
dist = d.State.rep[3]
d.State.rep[3] = d.State.rep[2]
}
d.State.rep[2] = d.State.rep[1]
}
d.State.rep[1] = d.State.rep[0]
d.State.rep[0] = dist
}
n, err := d.State.repLenCodec.Decode(d.rd, posState)
if err != nil {
return nil, err
}
d.State.updateStateRep()
op = match{n: int(n) + minMatchLen, distance: int64(dist) + minDistance}
return op, nil
}
// apply takes the operation and transforms the decoder dictionary accordingly.
func (d *decoder) apply(op operation) error {
var err error
switch x := op.(type) {
case match:
err = d.Dict.writeMatch(x.distance, x.n)
case lit:
err = d.Dict.WriteByte(x.b)
default:
panic("op is neither a match nor a literal")
}
return err
}
// decompress fills the dictionary unless no space for new data is
// available. If the end of the LZMA stream has been reached io.EOF will
// be returned.
func (d *decoder) decompress() error {
if d.eos {
return io.EOF
}
for d.Dict.Available() >= maxMatchLen {
op, err := d.readOp()
switch err {
case nil:
break
case errEOS:
d.eos = true
if !d.rd.possiblyAtEnd() {
return errDataAfterEOS
}
if d.size >= 0 && d.size != d.Decompressed() {
return errSize
}
return io.EOF
case io.EOF:
d.eos = true
return io.ErrUnexpectedEOF
default:
return err
}
if err = d.apply(op); err != nil {
return err
}
if d.size >= 0 && d.Decompressed() >= d.size {
d.eos = true
if d.Decompressed() > d.size {
return errSize
}
if !d.rd.possiblyAtEnd() {
switch _, err = d.readOp(); err {
case nil:
return errSize
case io.EOF:
return io.ErrUnexpectedEOF
case errEOS:
break
default:
return err
}
}
return io.EOF
}
}
return nil
}
// Errors that may be returned while decoding data.
var (
errDataAfterEOS = errors.New("lzma: data after end of stream marker")
errSize = errors.New("lzma: wrong uncompressed data size")
)
// Read reads data from the buffer. If no more data is available io.EOF is
// returned.
func (d *decoder) Read(p []byte) (n int, err error) {
var k int
for {
// Read of decoder dict never returns an error.
k, err = d.Dict.Read(p[n:])
if err != nil {
panic(fmt.Errorf("dictionary read error %s", err))
}
if k == 0 && d.eos {
return n, io.EOF
}
n += k
if n >= len(p) {
return n, nil
}
if err = d.decompress(); err != nil && err != io.EOF {
return n, err
}
}
}
// Decompressed returns the number of bytes decompressed by the decoder.
func (d *decoder) Decompressed() int64 {
return d.Dict.pos() - d.start
}

View File

@ -1,135 +0,0 @@
// Copyright 2014-2017 Ulrich Kunitz. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
package lzma
import (
"errors"
"fmt"
)
// decoderDict provides the dictionary for the decoder. The whole
// dictionary is used as reader buffer.
type decoderDict struct {
buf buffer
head int64
}
// newDecoderDict creates a new decoder dictionary. The whole dictionary
// will be used as reader buffer.
func newDecoderDict(dictCap int) (d *decoderDict, err error) {
// lower limit supports easy test cases
if !(1 <= dictCap && int64(dictCap) <= MaxDictCap) {
return nil, errors.New("lzma: dictCap out of range")
}
d = &decoderDict{buf: *newBuffer(dictCap)}
return d, nil
}
// Reset clears the dictionary. The read buffer is not changed, so the
// buffered data can still be read.
func (d *decoderDict) Reset() {
d.head = 0
}
// WriteByte writes a single byte into the dictionary. It is used to
// write literals into the dictionary.
func (d *decoderDict) WriteByte(c byte) error {
if err := d.buf.WriteByte(c); err != nil {
return err
}
d.head++
return nil
}
// pos returns the position of the dictionary head.
func (d *decoderDict) pos() int64 { return d.head }
// dictLen returns the actual length of the dictionary.
func (d *decoderDict) dictLen() int {
capacity := d.buf.Cap()
if d.head >= int64(capacity) {
return capacity
}
return int(d.head)
}
// byteAt returns a byte stored in the dictionary. If the distance is
// non-positive or exceeds the current length of the dictionary the zero
// byte is returned.
func (d *decoderDict) byteAt(dist int) byte {
if !(0 < dist && dist <= d.dictLen()) {
return 0
}
i := d.buf.front - dist
if i < 0 {
i += len(d.buf.data)
}
return d.buf.data[i]
}
// writeMatch writes the match at the top of the dictionary. The given
// distance must point in the current dictionary and the length must not
// exceed the maximum length 273 supported in LZMA.
//
// The error value ErrNoSpace indicates that no space is available in
// the dictionary for writing. You need to read from the dictionary
// first.
func (d *decoderDict) writeMatch(dist int64, length int) error {
if !(0 < dist && dist <= int64(d.dictLen())) {
return errors.New("writeMatch: distance out of range")
}
if !(0 < length && length <= maxMatchLen) {
return errors.New("writeMatch: length out of range")
}
if length > d.buf.Available() {
return ErrNoSpace
}
d.head += int64(length)
i := d.buf.front - int(dist)
if i < 0 {
i += len(d.buf.data)
}
for length > 0 {
var p []byte
if i >= d.buf.front {
p = d.buf.data[i:]
i = 0
} else {
p = d.buf.data[i:d.buf.front]
i = d.buf.front
}
if len(p) > length {
p = p[:length]
}
if _, err := d.buf.Write(p); err != nil {
panic(fmt.Errorf("d.buf.Write returned error %s", err))
}
length -= len(p)
}
return nil
}
// Write writes the given bytes into the dictionary and advances the
// head.
func (d *decoderDict) Write(p []byte) (n int, err error) {
n, err = d.buf.Write(p)
d.head += int64(n)
return n, err
}
// Available returns the number of available bytes for writing into the
// decoder dictionary.
func (d *decoderDict) Available() int { return d.buf.Available() }
// Read reads data from the buffer contained in the decoder dictionary.
func (d *decoderDict) Read(p []byte) (n int, err error) { return d.buf.Read(p) }
// Buffered returns the number of bytes currently buffered in the
// decoder dictionary.
func (d *decoderDict) buffered() int { return d.buf.Buffered() }
// Peek gets data from the buffer without advancing the rear index.
func (d *decoderDict) peek(p []byte) (n int, err error) { return d.buf.Peek(p) }

View File

@ -1,49 +0,0 @@
// Copyright 2014-2017 Ulrich Kunitz. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
package lzma
import "fmt"
// directCodec allows the encoding and decoding of values with a fixed number
// of bits. The number of bits must be in the range [1,32].
type directCodec byte
// makeDirectCodec creates a directCodec. The function panics if the number of
// bits is not in the range [1,32].
func makeDirectCodec(bits int) directCodec {
if !(1 <= bits && bits <= 32) {
panic(fmt.Errorf("bits=%d out of range", bits))
}
return directCodec(bits)
}
// Bits returns the number of bits supported by this codec.
func (dc directCodec) Bits() int {
return int(dc)
}
// Encode uses the range encoder to encode a value with the fixed number of
// bits. The most-significant bit is encoded first.
func (dc directCodec) Encode(e *rangeEncoder, v uint32) error {
for i := int(dc) - 1; i >= 0; i-- {
if err := e.DirectEncodeBit(v >> uint(i)); err != nil {
return err
}
}
return nil
}
// Decode uses the range decoder to decode a value with the given number of
// given bits. The most-significant bit is decoded first.
func (dc directCodec) Decode(d *rangeDecoder) (v uint32, err error) {
for i := int(dc) - 1; i >= 0; i-- {
x, err := d.DirectDecodeBit()
if err != nil {
return 0, err
}
v = (v << 1) | x
}
return v, nil
}

View File

@ -1,156 +0,0 @@
// Copyright 2014-2017 Ulrich Kunitz. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
package lzma
// Constants used by the distance codec.
const (
// minimum supported distance
minDistance = 1
// maximum supported distance, value is used for the eos marker.
maxDistance = 1 << 32
// number of the supported len states
lenStates = 4
// start for the position models
startPosModel = 4
// first index with align bits support
endPosModel = 14
// bits for the position slots
posSlotBits = 6
// number of align bits
alignBits = 4
// maximum position slot
maxPosSlot = 63
)
// distCodec provides encoding and decoding of distance values.
type distCodec struct {
posSlotCodecs [lenStates]treeCodec
posModel [endPosModel - startPosModel]treeReverseCodec
alignCodec treeReverseCodec
}
// deepcopy initializes dc as deep copy of the source.
func (dc *distCodec) deepcopy(src *distCodec) {
if dc == src {
return
}
for i := range dc.posSlotCodecs {
dc.posSlotCodecs[i].deepcopy(&src.posSlotCodecs[i])
}
for i := range dc.posModel {
dc.posModel[i].deepcopy(&src.posModel[i])
}
dc.alignCodec.deepcopy(&src.alignCodec)
}
// distBits returns the number of bits required to encode dist.
func distBits(dist uint32) int {
if dist < startPosModel {
return 6
}
// slot s > 3, dist d
// s = 2(bits(d)-1) + bit(d, bits(d)-2)
// s>>1 = bits(d)-1
// bits(d) = 32-nlz32(d)
// s>>1=31-nlz32(d)
// n = 5 + (s>>1) = 36 - nlz32(d)
return 36 - nlz32(dist)
}
// newDistCodec creates a new distance codec.
func (dc *distCodec) init() {
for i := range dc.posSlotCodecs {
dc.posSlotCodecs[i] = makeTreeCodec(posSlotBits)
}
for i := range dc.posModel {
posSlot := startPosModel + i
bits := (posSlot >> 1) - 1
dc.posModel[i] = makeTreeReverseCodec(bits)
}
dc.alignCodec = makeTreeReverseCodec(alignBits)
}
// lenState converts the value l to a supported lenState value.
func lenState(l uint32) uint32 {
if l >= lenStates {
l = lenStates - 1
}
return l
}
// Encode encodes the distance using the parameter l. Dist can have values from
// the full range of uint32 values. To get the distance offset the actual match
// distance has to be decreased by 1. A distance offset of 0xffffffff (eos)
// indicates the end of the stream.
func (dc *distCodec) Encode(e *rangeEncoder, dist uint32, l uint32) (err error) {
// Compute the posSlot using nlz32
var posSlot uint32
var bits uint32
if dist < startPosModel {
posSlot = dist
} else {
bits = uint32(30 - nlz32(dist))
posSlot = startPosModel - 2 + (bits << 1)
posSlot += (dist >> uint(bits)) & 1
}
if err = dc.posSlotCodecs[lenState(l)].Encode(e, posSlot); err != nil {
return
}
switch {
case posSlot < startPosModel:
return nil
case posSlot < endPosModel:
tc := &dc.posModel[posSlot-startPosModel]
return tc.Encode(dist, e)
}
dic := directCodec(bits - alignBits)
if err = dic.Encode(e, dist>>alignBits); err != nil {
return
}
return dc.alignCodec.Encode(dist, e)
}
// Decode decodes the distance offset using the parameter l. The dist value
// 0xffffffff (eos) indicates the end of the stream. Add one to the distance
// offset to get the actual match distance.
func (dc *distCodec) Decode(d *rangeDecoder, l uint32) (dist uint32, err error) {
posSlot, err := dc.posSlotCodecs[lenState(l)].Decode(d)
if err != nil {
return
}
// posSlot equals distance
if posSlot < startPosModel {
return posSlot, nil
}
// posSlot uses the individual models
bits := (posSlot >> 1) - 1
dist = (2 | (posSlot & 1)) << bits
var u uint32
if posSlot < endPosModel {
tc := &dc.posModel[posSlot-startPosModel]
if u, err = tc.Decode(d); err != nil {
return 0, err
}
dist += u
return dist, nil
}
// posSlots use direct encoding and a single model for the four align
// bits.
dic := directCodec(bits - alignBits)
if u, err = dic.Decode(d); err != nil {
return 0, err
}
dist += u << alignBits
if u, err = dc.alignCodec.Decode(d); err != nil {
return 0, err
}
dist += u
return dist, nil
}

View File

@ -1,268 +0,0 @@
// Copyright 2014-2017 Ulrich Kunitz. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
package lzma
import (
"fmt"
"io"
)
// opLenMargin provides the upper limit of the number of bytes required
// to encode a single operation.
const opLenMargin = 16
// compressFlags control the compression process.
type compressFlags uint32
// Values for compressFlags.
const (
// all data should be compressed, even if compression is not
// optimal.
all compressFlags = 1 << iota
)
// encoderFlags provide the flags for an encoder.
type encoderFlags uint32
// Flags for the encoder.
const (
// eosMarker requests an EOS marker to be written.
eosMarker encoderFlags = 1 << iota
)
// Encoder compresses data buffered in the encoder dictionary and writes
// it into a byte writer.
type encoder struct {
dict *encoderDict
state *state
re *rangeEncoder
start int64
// generate eos marker
marker bool
limit bool
margin int
}
// newEncoder creates a new encoder. If the byte writer must be
// limited use LimitedByteWriter provided by this package. The flags
// argument supports the eosMarker flag, controlling whether a
// terminating end-of-stream marker must be written.
func newEncoder(bw io.ByteWriter, state *state, dict *encoderDict,
flags encoderFlags) (e *encoder, err error) {
re, err := newRangeEncoder(bw)
if err != nil {
return nil, err
}
e = &encoder{
dict: dict,
state: state,
re: re,
marker: flags&eosMarker != 0,
start: dict.Pos(),
margin: opLenMargin,
}
if e.marker {
e.margin += 5
}
return e, nil
}
// Write writes the bytes from p into the dictionary. If not enough
// space is available the data in the dictionary buffer will be
// compressed to make additional space available. If the limit of the
// underlying writer has been reached ErrLimit will be returned.
func (e *encoder) Write(p []byte) (n int, err error) {
for {
k, err := e.dict.Write(p[n:])
n += k
if err == ErrNoSpace {
if err = e.compress(0); err != nil {
return n, err
}
continue
}
return n, err
}
}
// Reopen reopens the encoder with a new byte writer.
func (e *encoder) Reopen(bw io.ByteWriter) error {
var err error
if e.re, err = newRangeEncoder(bw); err != nil {
return err
}
e.start = e.dict.Pos()
e.limit = false
return nil
}
// writeLiteral writes a literal into the LZMA stream
func (e *encoder) writeLiteral(l lit) error {
var err error
state, state2, _ := e.state.states(e.dict.Pos())
if err = e.state.isMatch[state2].Encode(e.re, 0); err != nil {
return err
}
litState := e.state.litState(e.dict.ByteAt(1), e.dict.Pos())
match := e.dict.ByteAt(int(e.state.rep[0]) + 1)
err = e.state.litCodec.Encode(e.re, l.b, state, match, litState)
if err != nil {
return err
}
e.state.updateStateLiteral()
return nil
}
// iverson implements the Iverson operator as proposed by Donald Knuth in his
// book Concrete Mathematics.
func iverson(ok bool) uint32 {
if ok {
return 1
}
return 0
}
// writeMatch writes a repetition operation into the operation stream
func (e *encoder) writeMatch(m match) error {
var err error
if !(minDistance <= m.distance && m.distance <= maxDistance) {
panic(fmt.Errorf("match distance %d out of range", m.distance))
}
dist := uint32(m.distance - minDistance)
if !(minMatchLen <= m.n && m.n <= maxMatchLen) &&
!(dist == e.state.rep[0] && m.n == 1) {
panic(fmt.Errorf(
"match length %d out of range; dist %d rep[0] %d",
m.n, dist, e.state.rep[0]))
}
state, state2, posState := e.state.states(e.dict.Pos())
if err = e.state.isMatch[state2].Encode(e.re, 1); err != nil {
return err
}
g := 0
for ; g < 4; g++ {
if e.state.rep[g] == dist {
break
}
}
b := iverson(g < 4)
if err = e.state.isRep[state].Encode(e.re, b); err != nil {
return err
}
n := uint32(m.n - minMatchLen)
if b == 0 {
// simple match
e.state.rep[3], e.state.rep[2], e.state.rep[1], e.state.rep[0] =
e.state.rep[2], e.state.rep[1], e.state.rep[0], dist
e.state.updateStateMatch()
if err = e.state.lenCodec.Encode(e.re, n, posState); err != nil {
return err
}
return e.state.distCodec.Encode(e.re, dist, n)
}
b = iverson(g != 0)
if err = e.state.isRepG0[state].Encode(e.re, b); err != nil {
return err
}
if b == 0 {
// g == 0
b = iverson(m.n != 1)
if err = e.state.isRepG0Long[state2].Encode(e.re, b); err != nil {
return err
}
if b == 0 {
e.state.updateStateShortRep()
return nil
}
} else {
// g in {1,2,3}
b = iverson(g != 1)
if err = e.state.isRepG1[state].Encode(e.re, b); err != nil {
return err
}
if b == 1 {
// g in {2,3}
b = iverson(g != 2)
err = e.state.isRepG2[state].Encode(e.re, b)
if err != nil {
return err
}
if b == 1 {
e.state.rep[3] = e.state.rep[2]
}
e.state.rep[2] = e.state.rep[1]
}
e.state.rep[1] = e.state.rep[0]
e.state.rep[0] = dist
}
e.state.updateStateRep()
return e.state.repLenCodec.Encode(e.re, n, posState)
}
// writeOp writes a single operation to the range encoder. The function
// checks whether there is enough space available to close the LZMA
// stream.
func (e *encoder) writeOp(op operation) error {
if e.re.Available() < int64(e.margin) {
return ErrLimit
}
switch x := op.(type) {
case lit:
return e.writeLiteral(x)
case match:
return e.writeMatch(x)
default:
panic("unexpected operation")
}
}
// compress compressed data from the dictionary buffer. If the flag all
// is set, all data in the dictionary buffer will be compressed. The
// function returns ErrLimit if the underlying writer has reached its
// limit.
func (e *encoder) compress(flags compressFlags) error {
n := 0
if flags&all == 0 {
n = maxMatchLen - 1
}
d := e.dict
m := d.m
for d.Buffered() > n {
op := m.NextOp(e.state.rep)
if err := e.writeOp(op); err != nil {
return err
}
d.Discard(op.Len())
}
return nil
}
// eosMatch is a pseudo operation that indicates the end of the stream.
var eosMatch = match{distance: maxDistance, n: minMatchLen}
// Close terminates the LZMA stream. If requested the end-of-stream
// marker will be written. If the byte writer limit has been or will be
// reached during compression of the remaining data in the buffer the
// LZMA stream will be closed and data will remain in the buffer.
func (e *encoder) Close() error {
err := e.compress(all)
if err != nil && err != ErrLimit {
return err
}
if e.marker {
if err := e.writeMatch(eosMatch); err != nil {
return err
}
}
err = e.re.Close()
return err
}
// Compressed returns the number bytes of the input data that been
// compressed.
func (e *encoder) Compressed() int64 {
return e.dict.Pos() - e.start
}

View File

@ -1,149 +0,0 @@
// Copyright 2014-2017 Ulrich Kunitz. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
package lzma
import (
"errors"
"fmt"
"io"
)
// matcher is an interface that supports the identification of the next
// operation.
type matcher interface {
io.Writer
SetDict(d *encoderDict)
NextOp(rep [4]uint32) operation
}
// encoderDict provides the dictionary of the encoder. It includes an
// addtional buffer atop of the actual dictionary.
type encoderDict struct {
buf buffer
m matcher
head int64
capacity int
// preallocated array
data [maxMatchLen]byte
}
// newEncoderDict creates the encoder dictionary. The argument bufSize
// defines the size of the additional buffer.
func newEncoderDict(dictCap, bufSize int, m matcher) (d *encoderDict, err error) {
if !(1 <= dictCap && int64(dictCap) <= MaxDictCap) {
return nil, errors.New(
"lzma: dictionary capacity out of range")
}
if bufSize < 1 {
return nil, errors.New(
"lzma: buffer size must be larger than zero")
}
d = &encoderDict{
buf: *newBuffer(dictCap + bufSize),
capacity: dictCap,
m: m,
}
m.SetDict(d)
return d, nil
}
// Discard discards n bytes. Note that n must not be larger than
// MaxMatchLen.
func (d *encoderDict) Discard(n int) {
p := d.data[:n]
k, _ := d.buf.Read(p)
if k < n {
panic(fmt.Errorf("lzma: can't discard %d bytes", n))
}
d.head += int64(n)
d.m.Write(p)
}
// Len returns the data available in the encoder dictionary.
func (d *encoderDict) Len() int {
n := d.buf.Available()
if int64(n) > d.head {
return int(d.head)
}
return n
}
// DictLen returns the actual length of data in the dictionary.
func (d *encoderDict) DictLen() int {
if d.head < int64(d.capacity) {
return int(d.head)
}
return d.capacity
}
// Available returns the number of bytes that can be written by a
// following Write call.
func (d *encoderDict) Available() int {
return d.buf.Available() - d.DictLen()
}
// Write writes data into the dictionary buffer. Note that the position
// of the dictionary head will not be moved. If there is not enough
// space in the buffer ErrNoSpace will be returned.
func (d *encoderDict) Write(p []byte) (n int, err error) {
m := d.Available()
if len(p) > m {
p = p[:m]
err = ErrNoSpace
}
var e error
if n, e = d.buf.Write(p); e != nil {
err = e
}
return n, err
}
// Pos returns the position of the head.
func (d *encoderDict) Pos() int64 { return d.head }
// ByteAt returns the byte at the given distance.
func (d *encoderDict) ByteAt(distance int) byte {
if !(0 < distance && distance <= d.Len()) {
return 0
}
i := d.buf.rear - distance
if i < 0 {
i += len(d.buf.data)
}
return d.buf.data[i]
}
// CopyN copies the last n bytes from the dictionary into the provided
// writer. This is used for copying uncompressed data into an
// uncompressed segment.
func (d *encoderDict) CopyN(w io.Writer, n int) (written int, err error) {
if n <= 0 {
return 0, nil
}
m := d.Len()
if n > m {
n = m
err = ErrNoSpace
}
i := d.buf.rear - n
var e error
if i < 0 {
i += len(d.buf.data)
if written, e = w.Write(d.buf.data[i:]); e != nil {
return written, e
}
i = 0
}
var k int
k, e = w.Write(d.buf.data[i:d.buf.rear])
written += k
if e != nil {
err = e
}
return written, err
}
// Buffered returns the number of bytes in the buffer.
func (d *encoderDict) Buffered() int { return d.buf.Buffered() }

Binary file not shown.

View File

@ -1,309 +0,0 @@
// Copyright 2014-2017 Ulrich Kunitz. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
package lzma
import (
"errors"
"fmt"
"github.com/ulikunitz/xz/internal/hash"
)
/* For compression we need to find byte sequences that match the byte
* sequence at the dictionary head. A hash table is a simple method to
* provide this capability.
*/
// maxMatches limits the number of matches requested from the Matches
// function. This controls the speed of the overall encoding.
const maxMatches = 16
// shortDists defines the number of short distances supported by the
// implementation.
const shortDists = 8
// The minimum is somehow arbitrary but the maximum is limited by the
// memory requirements of the hash table.
const (
minTableExponent = 9
maxTableExponent = 20
)
// newRoller contains the function used to create an instance of the
// hash.Roller.
var newRoller = func(n int) hash.Roller { return hash.NewCyclicPoly(n) }
// hashTable stores the hash table including the rolling hash method.
//
// We implement chained hashing into a circular buffer. Each entry in
// the circular buffer stores the delta distance to the next position with a
// word that has the same hash value.
type hashTable struct {
dict *encoderDict
// actual hash table
t []int64
// circular list data with the offset to the next word
data []uint32
front int
// mask for computing the index for the hash table
mask uint64
// hash offset; initial value is -int64(wordLen)
hoff int64
// length of the hashed word
wordLen int
// hash roller for computing the hash values for the Write
// method
wr hash.Roller
// hash roller for computing arbitrary hashes
hr hash.Roller
// preallocated slices
p [maxMatches]int64
distances [maxMatches + shortDists]int
}
// hashTableExponent derives the hash table exponent from the dictionary
// capacity.
func hashTableExponent(n uint32) int {
e := 30 - nlz32(n)
switch {
case e < minTableExponent:
e = minTableExponent
case e > maxTableExponent:
e = maxTableExponent
}
return e
}
// newHashTable creates a new hash table for words of length wordLen
func newHashTable(capacity int, wordLen int) (t *hashTable, err error) {
if !(0 < capacity) {
return nil, errors.New(
"newHashTable: capacity must not be negative")
}
exp := hashTableExponent(uint32(capacity))
if !(1 <= wordLen && wordLen <= 4) {
return nil, errors.New("newHashTable: " +
"argument wordLen out of range")
}
n := 1 << uint(exp)
if n <= 0 {
panic("newHashTable: exponent is too large")
}
t = &hashTable{
t: make([]int64, n),
data: make([]uint32, capacity),
mask: (uint64(1) << uint(exp)) - 1,
hoff: -int64(wordLen),
wordLen: wordLen,
wr: newRoller(wordLen),
hr: newRoller(wordLen),
}
return t, nil
}
func (t *hashTable) SetDict(d *encoderDict) { t.dict = d }
// buffered returns the number of bytes that are currently hashed.
func (t *hashTable) buffered() int {
n := t.hoff + 1
switch {
case n <= 0:
return 0
case n >= int64(len(t.data)):
return len(t.data)
}
return int(n)
}
// addIndex adds n to an index ensuring that is stays inside the
// circular buffer for the hash chain.
func (t *hashTable) addIndex(i, n int) int {
i += n - len(t.data)
if i < 0 {
i += len(t.data)
}
return i
}
// putDelta puts the delta instance at the current front of the circular
// chain buffer.
func (t *hashTable) putDelta(delta uint32) {
t.data[t.front] = delta
t.front = t.addIndex(t.front, 1)
}
// putEntry puts a new entry into the hash table. If there is already a
// value stored it is moved into the circular chain buffer.
func (t *hashTable) putEntry(h uint64, pos int64) {
if pos < 0 {
return
}
i := h & t.mask
old := t.t[i] - 1
t.t[i] = pos + 1
var delta int64
if old >= 0 {
delta = pos - old
if delta > 1<<32-1 || delta > int64(t.buffered()) {
delta = 0
}
}
t.putDelta(uint32(delta))
}
// WriteByte converts a single byte into a hash and puts them into the hash
// table.
func (t *hashTable) WriteByte(b byte) error {
h := t.wr.RollByte(b)
t.hoff++
t.putEntry(h, t.hoff)
return nil
}
// Write converts the bytes provided into hash tables and stores the
// abbreviated offsets into the hash table. The method will never return an
// error.
func (t *hashTable) Write(p []byte) (n int, err error) {
for _, b := range p {
// WriteByte doesn't generate an error.
t.WriteByte(b)
}
return len(p), nil
}
// getMatches the matches for a specific hash. The functions returns the
// number of positions found.
//
// TODO: Make a getDistances because that we are actually interested in.
func (t *hashTable) getMatches(h uint64, positions []int64) (n int) {
if t.hoff < 0 || len(positions) == 0 {
return 0
}
buffered := t.buffered()
tailPos := t.hoff + 1 - int64(buffered)
rear := t.front - buffered
if rear >= 0 {
rear -= len(t.data)
}
// get the slot for the hash
pos := t.t[h&t.mask] - 1
delta := pos - tailPos
for {
if delta < 0 {
return n
}
positions[n] = tailPos + delta
n++
if n >= len(positions) {
return n
}
i := rear + int(delta)
if i < 0 {
i += len(t.data)
}
u := t.data[i]
if u == 0 {
return n
}
delta -= int64(u)
}
}
// hash computes the rolling hash for the word stored in p. For correct
// results its length must be equal to t.wordLen.
func (t *hashTable) hash(p []byte) uint64 {
var h uint64
for _, b := range p {
h = t.hr.RollByte(b)
}
return h
}
// Matches fills the positions slice with potential matches. The
// functions returns the number of positions filled into positions. The
// byte slice p must have word length of the hash table.
func (t *hashTable) Matches(p []byte, positions []int64) int {
if len(p) != t.wordLen {
panic(fmt.Errorf(
"byte slice must have length %d", t.wordLen))
}
h := t.hash(p)
return t.getMatches(h, positions)
}
// NextOp identifies the next operation using the hash table.
//
// TODO: Use all repetitions to find matches.
func (t *hashTable) NextOp(rep [4]uint32) operation {
// get positions
data := t.dict.data[:maxMatchLen]
n, _ := t.dict.buf.Peek(data)
data = data[:n]
var p []int64
if n < t.wordLen {
p = t.p[:0]
} else {
p = t.p[:maxMatches]
n = t.Matches(data[:t.wordLen], p)
p = p[:n]
}
// convert positions in potential distances
head := t.dict.head
dists := append(t.distances[:0], 1, 2, 3, 4, 5, 6, 7, 8)
for _, pos := range p {
dis := int(head - pos)
if dis > shortDists {
dists = append(dists, dis)
}
}
// check distances
var m match
dictLen := t.dict.DictLen()
for _, dist := range dists {
if dist > dictLen {
continue
}
// Here comes a trick. We are only interested in matches
// that are longer than the matches we have been found
// before. So before we test the whole byte sequence at
// the given distance, we test the first byte that would
// make the match longer. If it doesn't match the byte
// to match, we don't to care any longer.
i := t.dict.buf.rear - dist + m.n
if i < 0 {
i += len(t.dict.buf.data)
}
if t.dict.buf.data[i] != data[m.n] {
// We can't get a longer match. Jump to the next
// distance.
continue
}
n := t.dict.buf.matchLen(dist, data)
switch n {
case 0:
continue
case 1:
if uint32(dist-minDistance) != rep[0] {
continue
}
}
if n > m.n {
m = match{int64(dist), n}
if n == len(data) {
// No better match will be found.
break
}
}
}
if m.n == 0 {
return lit{data[0]}
}
return m
}

View File

@ -1,167 +0,0 @@
// Copyright 2014-2017 Ulrich Kunitz. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
package lzma
import (
"errors"
"fmt"
)
// uint32LE reads an uint32 integer from a byte slice
func uint32LE(b []byte) uint32 {
x := uint32(b[3]) << 24
x |= uint32(b[2]) << 16
x |= uint32(b[1]) << 8
x |= uint32(b[0])
return x
}
// uint64LE converts the uint64 value stored as little endian to an uint64
// value.
func uint64LE(b []byte) uint64 {
x := uint64(b[7]) << 56
x |= uint64(b[6]) << 48
x |= uint64(b[5]) << 40
x |= uint64(b[4]) << 32
x |= uint64(b[3]) << 24
x |= uint64(b[2]) << 16
x |= uint64(b[1]) << 8
x |= uint64(b[0])
return x
}
// putUint32LE puts an uint32 integer into a byte slice that must have at least
// a length of 4 bytes.
func putUint32LE(b []byte, x uint32) {
b[0] = byte(x)
b[1] = byte(x >> 8)
b[2] = byte(x >> 16)
b[3] = byte(x >> 24)
}
// putUint64LE puts the uint64 value into the byte slice as little endian
// value. The byte slice b must have at least place for 8 bytes.
func putUint64LE(b []byte, x uint64) {
b[0] = byte(x)
b[1] = byte(x >> 8)
b[2] = byte(x >> 16)
b[3] = byte(x >> 24)
b[4] = byte(x >> 32)
b[5] = byte(x >> 40)
b[6] = byte(x >> 48)
b[7] = byte(x >> 56)
}
// noHeaderSize defines the value of the length field in the LZMA header.
const noHeaderSize uint64 = 1<<64 - 1
// HeaderLen provides the length of the LZMA file header.
const HeaderLen = 13
// header represents the header of an LZMA file.
type header struct {
properties Properties
dictCap int
// uncompressed size; negative value if no size is given
size int64
}
// marshalBinary marshals the header.
func (h *header) marshalBinary() (data []byte, err error) {
if err = h.properties.verify(); err != nil {
return nil, err
}
if !(0 <= h.dictCap && int64(h.dictCap) <= MaxDictCap) {
return nil, fmt.Errorf("lzma: DictCap %d out of range",
h.dictCap)
}
data = make([]byte, 13)
// property byte
data[0] = h.properties.Code()
// dictionary capacity
putUint32LE(data[1:5], uint32(h.dictCap))
// uncompressed size
var s uint64
if h.size > 0 {
s = uint64(h.size)
} else {
s = noHeaderSize
}
putUint64LE(data[5:], s)
return data, nil
}
// unmarshalBinary unmarshals the header.
func (h *header) unmarshalBinary(data []byte) error {
if len(data) != HeaderLen {
return errors.New("lzma.unmarshalBinary: data has wrong length")
}
// properties
var err error
if h.properties, err = PropertiesForCode(data[0]); err != nil {
return err
}
// dictionary capacity
h.dictCap = int(uint32LE(data[1:]))
if h.dictCap < 0 {
return errors.New(
"LZMA header: dictionary capacity exceeds maximum " +
"integer")
}
// uncompressed size
s := uint64LE(data[5:])
if s == noHeaderSize {
h.size = -1
} else {
h.size = int64(s)
if h.size < 0 {
return errors.New(
"LZMA header: uncompressed size " +
"out of int64 range")
}
}
return nil
}
// validDictCap checks whether the dictionary capacity is correct. This
// is used to weed out wrong file headers.
func validDictCap(dictcap int) bool {
if int64(dictcap) == MaxDictCap {
return true
}
for n := uint(10); n < 32; n++ {
if dictcap == 1<<n {
return true
}
if dictcap == 1<<n+1<<(n-1) {
return true
}
}
return false
}
// ValidHeader checks for a valid LZMA file header. It allows only
// dictionary sizes of 2^n or 2^n+2^(n-1) with n >= 10 or 2^32-1. If
// there is an explicit size it must not exceed 256 GiB. The length of
// the data argument must be HeaderLen.
func ValidHeader(data []byte) bool {
var h header
if err := h.unmarshalBinary(data); err != nil {
return false
}
if !validDictCap(h.dictCap) {
return false
}
return h.size < 0 || h.size <= 1<<38
}

View File

@ -1,398 +0,0 @@
// Copyright 2014-2017 Ulrich Kunitz. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
package lzma
import (
"errors"
"fmt"
"io"
)
const (
// maximum size of compressed data in a chunk
maxCompressed = 1 << 16
// maximum size of uncompressed data in a chunk
maxUncompressed = 1 << 21
)
// chunkType represents the type of an LZMA2 chunk. Note that this
// value is an internal representation and no actual encoding of a LZMA2
// chunk header.
type chunkType byte
// Possible values for the chunk type.
const (
// end of stream
cEOS chunkType = iota
// uncompressed; reset dictionary
cUD
// uncompressed; no reset of dictionary
cU
// LZMA compressed; no reset
cL
// LZMA compressed; reset state
cLR
// LZMA compressed; reset state; new property value
cLRN
// LZMA compressed; reset state; new property value; reset dictionary
cLRND
)
// chunkTypeStrings provide a string representation for the chunk types.
var chunkTypeStrings = [...]string{
cEOS: "EOS",
cU: "U",
cUD: "UD",
cL: "L",
cLR: "LR",
cLRN: "LRN",
cLRND: "LRND",
}
// String returns a string representation of the chunk type.
func (c chunkType) String() string {
if !(cEOS <= c && c <= cLRND) {
return "unknown"
}
return chunkTypeStrings[c]
}
// Actual encodings for the chunk types in the value. Note that the high
// uncompressed size bits are stored in the header byte additionally.
const (
hEOS = 0
hUD = 1
hU = 2
hL = 1 << 7
hLR = 1<<7 | 1<<5
hLRN = 1<<7 | 1<<6
hLRND = 1<<7 | 1<<6 | 1<<5
)
// errHeaderByte indicates an unsupported value for the chunk header
// byte. These bytes starts the variable-length chunk header.
var errHeaderByte = errors.New("lzma: unsupported chunk header byte")
// headerChunkType converts the header byte into a chunk type. It
// ignores the uncompressed size bits in the chunk header byte.
func headerChunkType(h byte) (c chunkType, err error) {
if h&hL == 0 {
// no compression
switch h {
case hEOS:
c = cEOS
case hUD:
c = cUD
case hU:
c = cU
default:
return 0, errHeaderByte
}
return
}
switch h & hLRND {
case hL:
c = cL
case hLR:
c = cLR
case hLRN:
c = cLRN
case hLRND:
c = cLRND
default:
return 0, errHeaderByte
}
return
}
// uncompressedHeaderLen provides the length of an uncompressed header
const uncompressedHeaderLen = 3
// headerLen returns the length of the LZMA2 header for a given chunk
// type.
func headerLen(c chunkType) int {
switch c {
case cEOS:
return 1
case cU, cUD:
return uncompressedHeaderLen
case cL, cLR:
return 5
case cLRN, cLRND:
return 6
}
panic(fmt.Errorf("unsupported chunk type %d", c))
}
// chunkHeader represents the contents of a chunk header.
type chunkHeader struct {
ctype chunkType
uncompressed uint32
compressed uint16
props Properties
}
// String returns a string representation of the chunk header.
func (h *chunkHeader) String() string {
return fmt.Sprintf("%s %d %d %s", h.ctype, h.uncompressed,
h.compressed, &h.props)
}
// UnmarshalBinary reads the content of the chunk header from the data
// slice. The slice must have the correct length.
func (h *chunkHeader) UnmarshalBinary(data []byte) error {
if len(data) == 0 {
return errors.New("no data")
}
c, err := headerChunkType(data[0])
if err != nil {
return err
}
n := headerLen(c)
if len(data) < n {
return errors.New("incomplete data")
}
if len(data) > n {
return errors.New("invalid data length")
}
*h = chunkHeader{ctype: c}
if c == cEOS {
return nil
}
h.uncompressed = uint32(uint16BE(data[1:3]))
if c <= cU {
return nil
}
h.uncompressed |= uint32(data[0]&^hLRND) << 16
h.compressed = uint16BE(data[3:5])
if c <= cLR {
return nil
}
h.props, err = PropertiesForCode(data[5])
return err
}
// MarshalBinary encodes the chunk header value. The function checks
// whether the content of the chunk header is correct.
func (h *chunkHeader) MarshalBinary() (data []byte, err error) {
if h.ctype > cLRND {
return nil, errors.New("invalid chunk type")
}
if err = h.props.verify(); err != nil {
return nil, err
}
data = make([]byte, headerLen(h.ctype))
switch h.ctype {
case cEOS:
return data, nil
case cUD:
data[0] = hUD
case cU:
data[0] = hU
case cL:
data[0] = hL
case cLR:
data[0] = hLR
case cLRN:
data[0] = hLRN
case cLRND:
data[0] = hLRND
}
putUint16BE(data[1:3], uint16(h.uncompressed))
if h.ctype <= cU {
return data, nil
}
data[0] |= byte(h.uncompressed>>16) &^ hLRND
putUint16BE(data[3:5], h.compressed)
if h.ctype <= cLR {
return data, nil
}
data[5] = h.props.Code()
return data, nil
}
// readChunkHeader reads the chunk header from the IO reader.
func readChunkHeader(r io.Reader) (h *chunkHeader, err error) {
p := make([]byte, 1, 6)
if _, err = io.ReadFull(r, p); err != nil {
return
}
c, err := headerChunkType(p[0])
if err != nil {
return
}
p = p[:headerLen(c)]
if _, err = io.ReadFull(r, p[1:]); err != nil {
return
}
h = new(chunkHeader)
if err = h.UnmarshalBinary(p); err != nil {
return nil, err
}
return h, nil
}
// uint16BE converts a big-endian uint16 representation to an uint16
// value.
func uint16BE(p []byte) uint16 {
return uint16(p[0])<<8 | uint16(p[1])
}
// putUint16BE puts the big-endian uint16 presentation into the given
// slice.
func putUint16BE(p []byte, x uint16) {
p[0] = byte(x >> 8)
p[1] = byte(x)
}
// chunkState is used to manage the state of the chunks
type chunkState byte
// start and stop define the initial and terminating state of the chunk
// state
const (
start chunkState = 'S'
stop = 'T'
)
// errors for the chunk state handling
var (
errChunkType = errors.New("lzma: unexpected chunk type")
errState = errors.New("lzma: wrong chunk state")
)
// next transitions state based on chunk type input
func (c *chunkState) next(ctype chunkType) error {
switch *c {
// start state
case 'S':
switch ctype {
case cEOS:
*c = 'T'
case cUD:
*c = 'R'
case cLRND:
*c = 'L'
default:
return errChunkType
}
// normal LZMA mode
case 'L':
switch ctype {
case cEOS:
*c = 'T'
case cUD:
*c = 'R'
case cU:
*c = 'U'
case cL, cLR, cLRN, cLRND:
break
default:
return errChunkType
}
// reset required
case 'R':
switch ctype {
case cEOS:
*c = 'T'
case cUD, cU:
break
case cLRN, cLRND:
*c = 'L'
default:
return errChunkType
}
// uncompressed
case 'U':
switch ctype {
case cEOS:
*c = 'T'
case cUD:
*c = 'R'
case cU:
break
case cL, cLR, cLRN, cLRND:
*c = 'L'
default:
return errChunkType
}
// terminal state
case 'T':
return errChunkType
default:
return errState
}
return nil
}
// defaultChunkType returns the default chunk type for each chunk state.
func (c chunkState) defaultChunkType() chunkType {
switch c {
case 'S':
return cLRND
case 'L', 'U':
return cL
case 'R':
return cLRN
default:
// no error
return cEOS
}
}
// maxDictCap defines the maximum dictionary capacity supported by the
// LZMA2 dictionary capacity encoding.
const maxDictCap = 1<<32 - 1
// maxDictCapCode defines the maximum dictionary capacity code.
const maxDictCapCode = 40
// The function decodes the dictionary capacity byte, but doesn't change
// for the correct range of the given byte.
func decodeDictCap(c byte) int64 {
return (2 | int64(c)&1) << (11 + (c>>1)&0x1f)
}
// DecodeDictCap decodes the encoded dictionary capacity. The function
// returns an error if the code is out of range.
func DecodeDictCap(c byte) (n int64, err error) {
if c >= maxDictCapCode {
if c == maxDictCapCode {
return maxDictCap, nil
}
return 0, errors.New("lzma: invalid dictionary size code")
}
return decodeDictCap(c), nil
}
// EncodeDictCap encodes a dictionary capacity. The function returns the
// code for the capacity that is greater or equal n. If n exceeds the
// maximum support dictionary capacity, the maximum value is returned.
func EncodeDictCap(n int64) byte {
a, b := byte(0), byte(40)
for a < b {
c := a + (b-a)>>1
m := decodeDictCap(c)
if n <= m {
if n == m {
return c
}
b = c
} else {
a = c + 1
}
}
return a
}

View File

@ -1,129 +0,0 @@
// Copyright 2014-2017 Ulrich Kunitz. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
package lzma
import "errors"
// maxPosBits defines the number of bits of the position value that are used to
// to compute the posState value. The value is used to select the tree codec
// for length encoding and decoding.
const maxPosBits = 4
// minMatchLen and maxMatchLen give the minimum and maximum values for
// encoding and decoding length values. minMatchLen is also used as base
// for the encoded length values.
const (
minMatchLen = 2
maxMatchLen = minMatchLen + 16 + 256 - 1
)
// lengthCodec support the encoding of the length value.
type lengthCodec struct {
choice [2]prob
low [1 << maxPosBits]treeCodec
mid [1 << maxPosBits]treeCodec
high treeCodec
}
// deepcopy initializes the lc value as deep copy of the source value.
func (lc *lengthCodec) deepcopy(src *lengthCodec) {
if lc == src {
return
}
lc.choice = src.choice
for i := range lc.low {
lc.low[i].deepcopy(&src.low[i])
}
for i := range lc.mid {
lc.mid[i].deepcopy(&src.mid[i])
}
lc.high.deepcopy(&src.high)
}
// init initializes a new length codec.
func (lc *lengthCodec) init() {
for i := range lc.choice {
lc.choice[i] = probInit
}
for i := range lc.low {
lc.low[i] = makeTreeCodec(3)
}
for i := range lc.mid {
lc.mid[i] = makeTreeCodec(3)
}
lc.high = makeTreeCodec(8)
}
// lBits gives the number of bits used for the encoding of the l value
// provided to the range encoder.
func lBits(l uint32) int {
switch {
case l < 8:
return 4
case l < 16:
return 5
default:
return 10
}
}
// Encode encodes the length offset. The length offset l can be compute by
// subtracting minMatchLen (2) from the actual length.
//
// l = length - minMatchLen
//
func (lc *lengthCodec) Encode(e *rangeEncoder, l uint32, posState uint32,
) (err error) {
if l > maxMatchLen-minMatchLen {
return errors.New("lengthCodec.Encode: l out of range")
}
if l < 8 {
if err = lc.choice[0].Encode(e, 0); err != nil {
return
}
return lc.low[posState].Encode(e, l)
}
if err = lc.choice[0].Encode(e, 1); err != nil {
return
}
if l < 16 {
if err = lc.choice[1].Encode(e, 0); err != nil {
return
}
return lc.mid[posState].Encode(e, l-8)
}
if err = lc.choice[1].Encode(e, 1); err != nil {
return
}
if err = lc.high.Encode(e, l-16); err != nil {
return
}
return nil
}
// Decode reads the length offset. Add minMatchLen to compute the actual length
// to the length offset l.
func (lc *lengthCodec) Decode(d *rangeDecoder, posState uint32,
) (l uint32, err error) {
var b uint32
if b, err = lc.choice[0].Decode(d); err != nil {
return
}
if b == 0 {
l, err = lc.low[posState].Decode(d)
return
}
if b, err = lc.choice[1].Decode(d); err != nil {
return
}
if b == 0 {
l, err = lc.mid[posState].Decode(d)
l += 8
return
}
l, err = lc.high.Decode(d)
l += 16
return
}

View File

@ -1,132 +0,0 @@
// Copyright 2014-2017 Ulrich Kunitz. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
package lzma
// literalCodec supports the encoding of literal. It provides 768 probability
// values per literal state. The upper 512 probabilities are used with the
// context of a match bit.
type literalCodec struct {
probs []prob
}
// deepcopy initializes literal codec c as a deep copy of the source.
func (c *literalCodec) deepcopy(src *literalCodec) {
if c == src {
return
}
c.probs = make([]prob, len(src.probs))
copy(c.probs, src.probs)
}
// init initializes the literal codec.
func (c *literalCodec) init(lc, lp int) {
switch {
case !(minLC <= lc && lc <= maxLC):
panic("lc out of range")
case !(minLP <= lp && lp <= maxLP):
panic("lp out of range")
}
c.probs = make([]prob, 0x300<<uint(lc+lp))
for i := range c.probs {
c.probs[i] = probInit
}
}
// Encode encodes the byte s using a range encoder as well as the current LZMA
// encoder state, a match byte and the literal state.
func (c *literalCodec) Encode(e *rangeEncoder, s byte,
state uint32, match byte, litState uint32,
) (err error) {
k := litState * 0x300
probs := c.probs[k : k+0x300]
symbol := uint32(1)
r := uint32(s)
if state >= 7 {
m := uint32(match)
for {
matchBit := (m >> 7) & 1
m <<= 1
bit := (r >> 7) & 1
r <<= 1
i := ((1 + matchBit) << 8) | symbol
if err = probs[i].Encode(e, bit); err != nil {
return
}
symbol = (symbol << 1) | bit
if matchBit != bit {
break
}
if symbol >= 0x100 {
break
}
}
}
for symbol < 0x100 {
bit := (r >> 7) & 1
r <<= 1
if err = probs[symbol].Encode(e, bit); err != nil {
return
}
symbol = (symbol << 1) | bit
}
return nil
}
// Decode decodes a literal byte using the range decoder as well as the LZMA
// state, a match byte, and the literal state.
func (c *literalCodec) Decode(d *rangeDecoder,
state uint32, match byte, litState uint32,
) (s byte, err error) {
k := litState * 0x300
probs := c.probs[k : k+0x300]
symbol := uint32(1)
if state >= 7 {
m := uint32(match)
for {
matchBit := (m >> 7) & 1
m <<= 1
i := ((1 + matchBit) << 8) | symbol
bit, err := d.DecodeBit(&probs[i])
if err != nil {
return 0, err
}
symbol = (symbol << 1) | bit
if matchBit != bit {
break
}
if symbol >= 0x100 {
break
}
}
}
for symbol < 0x100 {
bit, err := d.DecodeBit(&probs[symbol])
if err != nil {
return 0, err
}
symbol = (symbol << 1) | bit
}
s = byte(symbol - 0x100)
return s, nil
}
// minLC and maxLC define the range for LC values.
const (
minLC = 0
maxLC = 8
)
// minLC and maxLC define the range for LP values.
const (
minLP = 0
maxLP = 4
)
// minState and maxState define a range for the state values stored in
// the State values.
const (
minState = 0
maxState = 11
)

View File

@ -1,52 +0,0 @@
// Copyright 2014-2017 Ulrich Kunitz. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
package lzma
import "errors"
// MatchAlgorithm identifies an algorithm to find matches in the
// dictionary.
type MatchAlgorithm byte
// Supported matcher algorithms.
const (
HashTable4 MatchAlgorithm = iota
BinaryTree
)
// maStrings are used by the String method.
var maStrings = map[MatchAlgorithm]string{
HashTable4: "HashTable4",
BinaryTree: "BinaryTree",
}
// String returns a string representation of the Matcher.
func (a MatchAlgorithm) String() string {
if s, ok := maStrings[a]; ok {
return s
}
return "unknown"
}
var errUnsupportedMatchAlgorithm = errors.New(
"lzma: unsupported match algorithm value")
// verify checks whether the matcher value is supported.
func (a MatchAlgorithm) verify() error {
if _, ok := maStrings[a]; !ok {
return errUnsupportedMatchAlgorithm
}
return nil
}
func (a MatchAlgorithm) new(dictCap int) (m matcher, err error) {
switch a {
case HashTable4:
return newHashTable(dictCap, 4)
case BinaryTree:
return newBinTree(dictCap)
}
return nil, errUnsupportedMatchAlgorithm
}

View File

@ -1,80 +0,0 @@
// Copyright 2014-2017 Ulrich Kunitz. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
package lzma
import (
"errors"
"fmt"
"unicode"
)
// operation represents an operation on the dictionary during encoding or
// decoding.
type operation interface {
Len() int
}
// rep represents a repetition at the given distance and the given length
type match struct {
// supports all possible distance values, including the eos marker
distance int64
// length
n int
}
// verify checks whether the match is valid. If that is not the case an
// error is returned.
func (m match) verify() error {
if !(minDistance <= m.distance && m.distance <= maxDistance) {
return errors.New("distance out of range")
}
if !(1 <= m.n && m.n <= maxMatchLen) {
return errors.New("length out of range")
}
return nil
}
// l return the l-value for the match, which is the difference of length
// n and 2.
func (m match) l() uint32 {
return uint32(m.n - minMatchLen)
}
// dist returns the dist value for the match, which is one less of the
// distance stored in the match.
func (m match) dist() uint32 {
return uint32(m.distance - minDistance)
}
// Len returns the number of bytes matched.
func (m match) Len() int {
return m.n
}
// String returns a string representation for the repetition.
func (m match) String() string {
return fmt.Sprintf("M{%d,%d}", m.distance, m.n)
}
// lit represents a single byte literal.
type lit struct {
b byte
}
// Len returns 1 for the single byte literal.
func (l lit) Len() int {
return 1
}
// String returns a string representation for the literal.
func (l lit) String() string {
var c byte
if unicode.IsPrint(rune(l.b)) {
c = l.b
} else {
c = '.'
}
return fmt.Sprintf("L{%c/%02x}", c, l.b)
}

View File

@ -1,53 +0,0 @@
// Copyright 2014-2017 Ulrich Kunitz. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
package lzma
// movebits defines the number of bits used for the updates of probability
// values.
const movebits = 5
// probbits defines the number of bits of a probability value.
const probbits = 11
// probInit defines 0.5 as initial value for prob values.
const probInit prob = 1 << (probbits - 1)
// Type prob represents probabilities. The type can also be used to encode and
// decode single bits.
type prob uint16
// Dec decreases the probability. The decrease is proportional to the
// probability value.
func (p *prob) dec() {
*p -= *p >> movebits
}
// Inc increases the probability. The Increase is proportional to the
// difference of 1 and the probability value.
func (p *prob) inc() {
*p += ((1 << probbits) - *p) >> movebits
}
// Computes the new bound for a given range using the probability value.
func (p prob) bound(r uint32) uint32 {
return (r >> probbits) * uint32(p)
}
// Bits returns 1. One is the number of bits that can be encoded or decoded
// with a single prob value.
func (p prob) Bits() int {
return 1
}
// Encode encodes the least-significant bit of v. Note that the p value will be
// changed.
func (p *prob) Encode(e *rangeEncoder, v uint32) error {
return e.EncodeBit(v, p)
}
// Decode decodes a single bit. Note that the p value will change.
func (p *prob) Decode(d *rangeDecoder) (v uint32, err error) {
return d.DecodeBit(p)
}

View File

@ -1,69 +0,0 @@
// Copyright 2014-2017 Ulrich Kunitz. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
package lzma
import (
"errors"
"fmt"
)
// maximum and minimum values for the LZMA properties.
const (
minPB = 0
maxPB = 4
)
// maxPropertyCode is the possible maximum of a properties code byte.
const maxPropertyCode = (maxPB+1)*(maxLP+1)*(maxLC+1) - 1
// Properties contains the parameters LC, LP and PB. The parameter LC
// defines the number of literal context bits; parameter LP the number
// of literal position bits and PB the number of position bits.
type Properties struct {
LC int
LP int
PB int
}
// String returns the properties in a string representation.
func (p *Properties) String() string {
return fmt.Sprintf("LC %d LP %d PB %d", p.LC, p.LP, p.PB)
}
// PropertiesForCode converts a properties code byte into a Properties value.
func PropertiesForCode(code byte) (p Properties, err error) {
if code > maxPropertyCode {
return p, errors.New("lzma: invalid properties code")
}
p.LC = int(code % 9)
code /= 9
p.LP = int(code % 5)
code /= 5
p.PB = int(code % 5)
return p, err
}
// verify checks the properties for correctness.
func (p *Properties) verify() error {
if p == nil {
return errors.New("lzma: properties are nil")
}
if !(minLC <= p.LC && p.LC <= maxLC) {
return errors.New("lzma: lc out of range")
}
if !(minLP <= p.LP && p.LP <= maxLP) {
return errors.New("lzma: lp out of range")
}
if !(minPB <= p.PB && p.PB <= maxPB) {
return errors.New("lzma: pb out of range")
}
return nil
}
// Code converts the properties to a byte. The function assumes that
// the properties components are all in range.
func (p Properties) Code() byte {
return byte((p.PB*5+p.LP)*9 + p.LC)
}

View File

@ -1,248 +0,0 @@
// Copyright 2014-2017 Ulrich Kunitz. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
package lzma
import (
"errors"
"io"
)
// rangeEncoder implements range encoding of single bits. The low value can
// overflow therefore we need uint64. The cache value is used to handle
// overflows.
type rangeEncoder struct {
lbw *LimitedByteWriter
nrange uint32
low uint64
cacheLen int64
cache byte
}
// maxInt64 provides the maximal value of the int64 type
const maxInt64 = 1<<63 - 1
// newRangeEncoder creates a new range encoder.
func newRangeEncoder(bw io.ByteWriter) (re *rangeEncoder, err error) {
lbw, ok := bw.(*LimitedByteWriter)
if !ok {
lbw = &LimitedByteWriter{BW: bw, N: maxInt64}
}
return &rangeEncoder{
lbw: lbw,
nrange: 0xffffffff,
cacheLen: 1}, nil
}
// Available returns the number of bytes that still can be written. The
// method takes the bytes that will be currently written by Close into
// account.
func (e *rangeEncoder) Available() int64 {
return e.lbw.N - (e.cacheLen + 4)
}
// writeByte writes a single byte to the underlying writer. An error is
// returned if the limit is reached. The written byte will be counted if
// the underlying writer doesn't return an error.
func (e *rangeEncoder) writeByte(c byte) error {
if e.Available() < 1 {
return ErrLimit
}
return e.lbw.WriteByte(c)
}
// DirectEncodeBit encodes the least-significant bit of b with probability 1/2.
func (e *rangeEncoder) DirectEncodeBit(b uint32) error {
e.nrange >>= 1
e.low += uint64(e.nrange) & (0 - (uint64(b) & 1))
// normalize
const top = 1 << 24
if e.nrange >= top {
return nil
}
e.nrange <<= 8
return e.shiftLow()
}
// EncodeBit encodes the least significant bit of b. The p value will be
// updated by the function depending on the bit encoded.
func (e *rangeEncoder) EncodeBit(b uint32, p *prob) error {
bound := p.bound(e.nrange)
if b&1 == 0 {
e.nrange = bound
p.inc()
} else {
e.low += uint64(bound)
e.nrange -= bound
p.dec()
}
// normalize
const top = 1 << 24
if e.nrange >= top {
return nil
}
e.nrange <<= 8
return e.shiftLow()
}
// Close writes a complete copy of the low value.
func (e *rangeEncoder) Close() error {
for i := 0; i < 5; i++ {
if err := e.shiftLow(); err != nil {
return err
}
}
return nil
}
// shiftLow shifts the low value for 8 bit. The shifted byte is written into
// the byte writer. The cache value is used to handle overflows.
func (e *rangeEncoder) shiftLow() error {
if uint32(e.low) < 0xff000000 || (e.low>>32) != 0 {
tmp := e.cache
for {
err := e.writeByte(tmp + byte(e.low>>32))
if err != nil {
return err
}
tmp = 0xff
e.cacheLen--
if e.cacheLen <= 0 {
if e.cacheLen < 0 {
panic("negative cacheLen")
}
break
}
}
e.cache = byte(uint32(e.low) >> 24)
}
e.cacheLen++
e.low = uint64(uint32(e.low) << 8)
return nil
}
// rangeDecoder decodes single bits of the range encoding stream.
type rangeDecoder struct {
br io.ByteReader
nrange uint32
code uint32
}
// init initializes the range decoder, by reading from the byte reader.
func (d *rangeDecoder) init() error {
d.nrange = 0xffffffff
d.code = 0
b, err := d.br.ReadByte()
if err != nil {
return err
}
if b != 0 {
return errors.New("newRangeDecoder: first byte not zero")
}
for i := 0; i < 4; i++ {
if err = d.updateCode(); err != nil {
return err
}
}
if d.code >= d.nrange {
return errors.New("newRangeDecoder: d.code >= d.nrange")
}
return nil
}
// newRangeDecoder initializes a range decoder. It reads five bytes from the
// reader and therefore may return an error.
func newRangeDecoder(br io.ByteReader) (d *rangeDecoder, err error) {
d = &rangeDecoder{br: br, nrange: 0xffffffff}
b, err := d.br.ReadByte()
if err != nil {
return nil, err
}
if b != 0 {
return nil, errors.New("newRangeDecoder: first byte not zero")
}
for i := 0; i < 4; i++ {
if err = d.updateCode(); err != nil {
return nil, err
}
}
if d.code >= d.nrange {
return nil, errors.New("newRangeDecoder: d.code >= d.nrange")
}
return d, nil
}
// possiblyAtEnd checks whether the decoder may be at the end of the stream.
func (d *rangeDecoder) possiblyAtEnd() bool {
return d.code == 0
}
// DirectDecodeBit decodes a bit with probability 1/2. The return value b will
// contain the bit at the least-significant position. All other bits will be
// zero.
func (d *rangeDecoder) DirectDecodeBit() (b uint32, err error) {
d.nrange >>= 1
d.code -= d.nrange
t := 0 - (d.code >> 31)
d.code += d.nrange & t
b = (t + 1) & 1
// d.code will stay less then d.nrange
// normalize
// assume d.code < d.nrange
const top = 1 << 24
if d.nrange >= top {
return b, nil
}
d.nrange <<= 8
// d.code < d.nrange will be maintained
return b, d.updateCode()
}
// decodeBit decodes a single bit. The bit will be returned at the
// least-significant position. All other bits will be zero. The probability
// value will be updated.
func (d *rangeDecoder) DecodeBit(p *prob) (b uint32, err error) {
bound := p.bound(d.nrange)
if d.code < bound {
d.nrange = bound
p.inc()
b = 0
} else {
d.code -= bound
d.nrange -= bound
p.dec()
b = 1
}
// normalize
// assume d.code < d.nrange
const top = 1 << 24
if d.nrange >= top {
return b, nil
}
d.nrange <<= 8
// d.code < d.nrange will be maintained
return b, d.updateCode()
}
// updateCode reads a new byte into the code.
func (d *rangeDecoder) updateCode() error {
b, err := d.br.ReadByte()
if err != nil {
return err
}
d.code = (d.code << 8) | uint32(b)
return nil
}

View File

@ -1,100 +0,0 @@
// Copyright 2014-2017 Ulrich Kunitz. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
// Package lzma supports the decoding and encoding of LZMA streams.
// Reader and Writer support the classic LZMA format. Reader2 and
// Writer2 support the decoding and encoding of LZMA2 streams.
//
// The package is written completely in Go and doesn't rely on any external
// library.
package lzma
import (
"errors"
"io"
)
// ReaderConfig stores the parameters for the reader of the classic LZMA
// format.
type ReaderConfig struct {
DictCap int
}
// fill converts the zero values of the configuration to the default values.
func (c *ReaderConfig) fill() {
if c.DictCap == 0 {
c.DictCap = 8 * 1024 * 1024
}
}
// Verify checks the reader configuration for errors. Zero values will
// be replaced by default values.
func (c *ReaderConfig) Verify() error {
c.fill()
if !(MinDictCap <= c.DictCap && int64(c.DictCap) <= MaxDictCap) {
return errors.New("lzma: dictionary capacity is out of range")
}
return nil
}
// Reader provides a reader for LZMA files or streams.
type Reader struct {
lzma io.Reader
h header
d *decoder
}
// NewReader creates a new reader for an LZMA stream using the classic
// format. NewReader reads and checks the header of the LZMA stream.
func NewReader(lzma io.Reader) (r *Reader, err error) {
return ReaderConfig{}.NewReader(lzma)
}
// NewReader creates a new reader for an LZMA stream in the classic
// format. The function reads and verifies the the header of the LZMA
// stream.
func (c ReaderConfig) NewReader(lzma io.Reader) (r *Reader, err error) {
if err = c.Verify(); err != nil {
return nil, err
}
data := make([]byte, HeaderLen)
if _, err := io.ReadFull(lzma, data); err != nil {
if err == io.EOF {
return nil, errors.New("lzma: unexpected EOF")
}
return nil, err
}
r = &Reader{lzma: lzma}
if err = r.h.unmarshalBinary(data); err != nil {
return nil, err
}
if r.h.dictCap < MinDictCap {
return nil, errors.New("lzma: dictionary capacity too small")
}
dictCap := r.h.dictCap
if c.DictCap > dictCap {
dictCap = c.DictCap
}
state := newState(r.h.properties)
dict, err := newDecoderDict(dictCap)
if err != nil {
return nil, err
}
r.d, err = newDecoder(ByteReader(lzma), state, dict, r.h.size)
if err != nil {
return nil, err
}
return r, nil
}
// EOSMarker indicates that an EOS marker has been encountered.
func (r *Reader) EOSMarker() bool {
return r.d.eosMarker
}
// Read returns uncompressed data.
func (r *Reader) Read(p []byte) (n int, err error) {
return r.d.Read(p)
}

View File

@ -1,232 +0,0 @@
// Copyright 2014-2017 Ulrich Kunitz. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
package lzma
import (
"errors"
"io"
"github.com/ulikunitz/xz/internal/xlog"
)
// Reader2Config stores the parameters for the LZMA2 reader.
// format.
type Reader2Config struct {
DictCap int
}
// fill converts the zero values of the configuration to the default values.
func (c *Reader2Config) fill() {
if c.DictCap == 0 {
c.DictCap = 8 * 1024 * 1024
}
}
// Verify checks the reader configuration for errors. Zero configuration values
// will be replaced by default values.
func (c *Reader2Config) Verify() error {
c.fill()
if !(MinDictCap <= c.DictCap && int64(c.DictCap) <= MaxDictCap) {
return errors.New("lzma: dictionary capacity is out of range")
}
return nil
}
// Reader2 supports the reading of LZMA2 chunk sequences. Note that the
// first chunk should have a dictionary reset and the first compressed
// chunk a properties reset. The chunk sequence may not be terminated by
// an end-of-stream chunk.
type Reader2 struct {
r io.Reader
err error
dict *decoderDict
ur *uncompressedReader
decoder *decoder
chunkReader io.Reader
cstate chunkState
ctype chunkType
}
// NewReader2 creates a reader for an LZMA2 chunk sequence.
func NewReader2(lzma2 io.Reader) (r *Reader2, err error) {
return Reader2Config{}.NewReader2(lzma2)
}
// NewReader2 creates an LZMA2 reader using the given configuration.
func (c Reader2Config) NewReader2(lzma2 io.Reader) (r *Reader2, err error) {
if err = c.Verify(); err != nil {
return nil, err
}
r = &Reader2{r: lzma2, cstate: start}
r.dict, err = newDecoderDict(c.DictCap)
if err != nil {
return nil, err
}
if err = r.startChunk(); err != nil {
r.err = err
}
return r, nil
}
// uncompressed tests whether the chunk type specifies an uncompressed
// chunk.
func uncompressed(ctype chunkType) bool {
return ctype == cU || ctype == cUD
}
// startChunk parses a new chunk.
func (r *Reader2) startChunk() error {
r.chunkReader = nil
header, err := readChunkHeader(r.r)
if err != nil {
if err == io.EOF {
err = io.ErrUnexpectedEOF
}
return err
}
xlog.Debugf("chunk header %v", header)
if err = r.cstate.next(header.ctype); err != nil {
return err
}
if r.cstate == stop {
return io.EOF
}
if header.ctype == cUD || header.ctype == cLRND {
r.dict.Reset()
}
size := int64(header.uncompressed) + 1
if uncompressed(header.ctype) {
if r.ur != nil {
r.ur.Reopen(r.r, size)
} else {
r.ur = newUncompressedReader(r.r, r.dict, size)
}
r.chunkReader = r.ur
return nil
}
br := ByteReader(io.LimitReader(r.r, int64(header.compressed)+1))
if r.decoder == nil {
state := newState(header.props)
r.decoder, err = newDecoder(br, state, r.dict, size)
if err != nil {
return err
}
r.chunkReader = r.decoder
return nil
}
switch header.ctype {
case cLR:
r.decoder.State.Reset()
case cLRN, cLRND:
r.decoder.State = newState(header.props)
}
err = r.decoder.Reopen(br, size)
if err != nil {
return err
}
r.chunkReader = r.decoder
return nil
}
// Read reads data from the LZMA2 chunk sequence.
func (r *Reader2) Read(p []byte) (n int, err error) {
if r.err != nil {
return 0, r.err
}
for n < len(p) {
var k int
k, err = r.chunkReader.Read(p[n:])
n += k
if err != nil {
if err == io.EOF {
err = r.startChunk()
if err == nil {
continue
}
}
r.err = err
return n, err
}
if k == 0 {
r.err = errors.New("lzma: Reader2 doesn't get data")
return n, r.err
}
}
return n, nil
}
// EOS returns whether the LZMA2 stream has been terminated by an
// end-of-stream chunk.
func (r *Reader2) EOS() bool {
return r.cstate == stop
}
// uncompressedReader is used to read uncompressed chunks.
type uncompressedReader struct {
lr io.LimitedReader
Dict *decoderDict
eof bool
err error
}
// newUncompressedReader initializes a new uncompressedReader.
func newUncompressedReader(r io.Reader, dict *decoderDict, size int64) *uncompressedReader {
ur := &uncompressedReader{
lr: io.LimitedReader{R: r, N: size},
Dict: dict,
}
return ur
}
// Reopen reinitializes an uncompressed reader.
func (ur *uncompressedReader) Reopen(r io.Reader, size int64) {
ur.err = nil
ur.eof = false
ur.lr = io.LimitedReader{R: r, N: size}
}
// fill reads uncompressed data into the dictionary.
func (ur *uncompressedReader) fill() error {
if !ur.eof {
n, err := io.CopyN(ur.Dict, &ur.lr, int64(ur.Dict.Available()))
if err != io.EOF {
return err
}
ur.eof = true
if n > 0 {
return nil
}
}
if ur.lr.N != 0 {
return io.ErrUnexpectedEOF
}
return io.EOF
}
// Read reads uncompressed data from the limited reader.
func (ur *uncompressedReader) Read(p []byte) (n int, err error) {
if ur.err != nil {
return 0, ur.err
}
for {
var k int
k, err = ur.Dict.Read(p[n:])
n += k
if n >= len(p) {
return n, nil
}
if err != nil {
break
}
err = ur.fill()
if err != nil {
break
}
}
ur.err = err
return n, err
}

View File

@ -1,151 +0,0 @@
// Copyright 2014-2017 Ulrich Kunitz. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
package lzma
// states defines the overall state count
const states = 12
// State maintains the full state of the operation encoding or decoding
// process.
type state struct {
rep [4]uint32
isMatch [states << maxPosBits]prob
isRepG0Long [states << maxPosBits]prob
isRep [states]prob
isRepG0 [states]prob
isRepG1 [states]prob
isRepG2 [states]prob
litCodec literalCodec
lenCodec lengthCodec
repLenCodec lengthCodec
distCodec distCodec
state uint32
posBitMask uint32
Properties Properties
}
// initProbSlice initializes a slice of probabilities.
func initProbSlice(p []prob) {
for i := range p {
p[i] = probInit
}
}
// Reset sets all state information to the original values.
func (s *state) Reset() {
p := s.Properties
*s = state{
Properties: p,
// dict: s.dict,
posBitMask: (uint32(1) << uint(p.PB)) - 1,
}
initProbSlice(s.isMatch[:])
initProbSlice(s.isRep[:])
initProbSlice(s.isRepG0[:])
initProbSlice(s.isRepG1[:])
initProbSlice(s.isRepG2[:])
initProbSlice(s.isRepG0Long[:])
s.litCodec.init(p.LC, p.LP)
s.lenCodec.init()
s.repLenCodec.init()
s.distCodec.init()
}
// initState initializes the state.
func initState(s *state, p Properties) {
*s = state{Properties: p}
s.Reset()
}
// newState creates a new state from the give Properties.
func newState(p Properties) *state {
s := &state{Properties: p}
s.Reset()
return s
}
// deepcopy initializes s as a deep copy of the source.
func (s *state) deepcopy(src *state) {
if s == src {
return
}
s.rep = src.rep
s.isMatch = src.isMatch
s.isRepG0Long = src.isRepG0Long
s.isRep = src.isRep
s.isRepG0 = src.isRepG0
s.isRepG1 = src.isRepG1
s.isRepG2 = src.isRepG2
s.litCodec.deepcopy(&src.litCodec)
s.lenCodec.deepcopy(&src.lenCodec)
s.repLenCodec.deepcopy(&src.repLenCodec)
s.distCodec.deepcopy(&src.distCodec)
s.state = src.state
s.posBitMask = src.posBitMask
s.Properties = src.Properties
}
// cloneState creates a new clone of the give state.
func cloneState(src *state) *state {
s := new(state)
s.deepcopy(src)
return s
}
// updateStateLiteral updates the state for a literal.
func (s *state) updateStateLiteral() {
switch {
case s.state < 4:
s.state = 0
return
case s.state < 10:
s.state -= 3
return
}
s.state -= 6
}
// updateStateMatch updates the state for a match.
func (s *state) updateStateMatch() {
if s.state < 7 {
s.state = 7
} else {
s.state = 10
}
}
// updateStateRep updates the state for a repetition.
func (s *state) updateStateRep() {
if s.state < 7 {
s.state = 8
} else {
s.state = 11
}
}
// updateStateShortRep updates the state for a short repetition.
func (s *state) updateStateShortRep() {
if s.state < 7 {
s.state = 9
} else {
s.state = 11
}
}
// states computes the states of the operation codec.
func (s *state) states(dictHead int64) (state1, state2, posState uint32) {
state1 = s.state
posState = uint32(dictHead) & s.posBitMask
state2 = (s.state << maxPosBits) | posState
return
}
// litState computes the literal state.
func (s *state) litState(prev byte, dictHead int64) uint32 {
lp, lc := uint(s.Properties.LP), uint(s.Properties.LC)
litState := ((uint32(dictHead) & ((1 << lp) - 1)) << lc) |
(uint32(prev) >> (8 - lc))
return litState
}

View File

@ -1,133 +0,0 @@
// Copyright 2014-2017 Ulrich Kunitz. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
package lzma
// treeCodec encodes or decodes values with a fixed bit size. It is using a
// tree of probability value. The root of the tree is the most-significant bit.
type treeCodec struct {
probTree
}
// makeTreeCodec makes a tree codec. The bits value must be inside the range
// [1,32].
func makeTreeCodec(bits int) treeCodec {
return treeCodec{makeProbTree(bits)}
}
// deepcopy initializes tc as a deep copy of the source.
func (tc *treeCodec) deepcopy(src *treeCodec) {
tc.probTree.deepcopy(&src.probTree)
}
// Encode uses the range encoder to encode a fixed-bit-size value.
func (tc *treeCodec) Encode(e *rangeEncoder, v uint32) (err error) {
m := uint32(1)
for i := int(tc.bits) - 1; i >= 0; i-- {
b := (v >> uint(i)) & 1
if err := e.EncodeBit(b, &tc.probs[m]); err != nil {
return err
}
m = (m << 1) | b
}
return nil
}
// Decodes uses the range decoder to decode a fixed-bit-size value. Errors may
// be caused by the range decoder.
func (tc *treeCodec) Decode(d *rangeDecoder) (v uint32, err error) {
m := uint32(1)
for j := 0; j < int(tc.bits); j++ {
b, err := d.DecodeBit(&tc.probs[m])
if err != nil {
return 0, err
}
m = (m << 1) | b
}
return m - (1 << uint(tc.bits)), nil
}
// treeReverseCodec is another tree codec, where the least-significant bit is
// the start of the probability tree.
type treeReverseCodec struct {
probTree
}
// deepcopy initializes the treeReverseCodec as a deep copy of the
// source.
func (tc *treeReverseCodec) deepcopy(src *treeReverseCodec) {
tc.probTree.deepcopy(&src.probTree)
}
// makeTreeReverseCodec creates treeReverseCodec value. The bits argument must
// be in the range [1,32].
func makeTreeReverseCodec(bits int) treeReverseCodec {
return treeReverseCodec{makeProbTree(bits)}
}
// Encode uses range encoder to encode a fixed-bit-size value. The range
// encoder may cause errors.
func (tc *treeReverseCodec) Encode(v uint32, e *rangeEncoder) (err error) {
m := uint32(1)
for i := uint(0); i < uint(tc.bits); i++ {
b := (v >> i) & 1
if err := e.EncodeBit(b, &tc.probs[m]); err != nil {
return err
}
m = (m << 1) | b
}
return nil
}
// Decodes uses the range decoder to decode a fixed-bit-size value. Errors
// returned by the range decoder will be returned.
func (tc *treeReverseCodec) Decode(d *rangeDecoder) (v uint32, err error) {
m := uint32(1)
for j := uint(0); j < uint(tc.bits); j++ {
b, err := d.DecodeBit(&tc.probs[m])
if err != nil {
return 0, err
}
m = (m << 1) | b
v |= b << j
}
return v, nil
}
// probTree stores enough probability values to be used by the treeEncode and
// treeDecode methods of the range coder types.
type probTree struct {
probs []prob
bits byte
}
// deepcopy initializes the probTree value as a deep copy of the source.
func (t *probTree) deepcopy(src *probTree) {
if t == src {
return
}
t.probs = make([]prob, len(src.probs))
copy(t.probs, src.probs)
t.bits = src.bits
}
// makeProbTree initializes a probTree structure.
func makeProbTree(bits int) probTree {
if !(1 <= bits && bits <= 32) {
panic("bits outside of range [1,32]")
}
t := probTree{
bits: byte(bits),
probs: make([]prob, 1<<uint(bits)),
}
for i := range t.probs {
t.probs[i] = probInit
}
return t
}
// Bits provides the number of bits for the values to de- or encode.
func (t *probTree) Bits() int {
return int(t.bits)
}

View File

@ -1,209 +0,0 @@
// Copyright 2014-2017 Ulrich Kunitz. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
package lzma
import (
"bufio"
"errors"
"io"
)
// MinDictCap and MaxDictCap provide the range of supported dictionary
// capacities.
const (
MinDictCap = 1 << 12
MaxDictCap = 1<<32 - 1
)
// WriterConfig defines the configuration parameter for a writer.
type WriterConfig struct {
// Properties for the encoding. If the it is nil the value
// {LC: 3, LP: 0, PB: 2} will be chosen.
Properties *Properties
// The capacity of the dictionary. If DictCap is zero, the value
// 8 MiB will be chosen.
DictCap int
// Size of the lookahead buffer; value 0 indicates default size
// 4096
BufSize int
// Match algorithm
Matcher MatchAlgorithm
// SizeInHeader indicates that the header will contain an
// explicit size.
SizeInHeader bool
// Size of the data to be encoded. A positive value will imply
// than an explicit size will be set in the header.
Size int64
// EOSMarker requests whether the EOSMarker needs to be written.
// If no explicit size is been given the EOSMarker will be
// set automatically.
EOSMarker bool
}
// fill converts zero-value fields to their explicit default values.
func (c *WriterConfig) fill() {
if c.Properties == nil {
c.Properties = &Properties{LC: 3, LP: 0, PB: 2}
}
if c.DictCap == 0 {
c.DictCap = 8 * 1024 * 1024
}
if c.BufSize == 0 {
c.BufSize = 4096
}
if c.Size > 0 {
c.SizeInHeader = true
}
if !c.SizeInHeader {
c.EOSMarker = true
}
}
// Verify checks WriterConfig for errors. Verify will replace zero
// values with default values.
func (c *WriterConfig) Verify() error {
c.fill()
var err error
if c == nil {
return errors.New("lzma: WriterConfig is nil")
}
if c.Properties == nil {
return errors.New("lzma: WriterConfig has no Properties set")
}
if err = c.Properties.verify(); err != nil {
return err
}
if !(MinDictCap <= c.DictCap && int64(c.DictCap) <= MaxDictCap) {
return errors.New("lzma: dictionary capacity is out of range")
}
if !(maxMatchLen <= c.BufSize) {
return errors.New("lzma: lookahead buffer size too small")
}
if c.SizeInHeader {
if c.Size < 0 {
return errors.New("lzma: negative size not supported")
}
} else if !c.EOSMarker {
return errors.New("lzma: EOS marker is required")
}
if err = c.Matcher.verify(); err != nil {
return err
}
return nil
}
// header returns the header structure for this configuration.
func (c *WriterConfig) header() header {
h := header{
properties: *c.Properties,
dictCap: c.DictCap,
size: -1,
}
if c.SizeInHeader {
h.size = c.Size
}
return h
}
// Writer writes an LZMA stream in the classic format.
type Writer struct {
h header
bw io.ByteWriter
buf *bufio.Writer
e *encoder
}
// NewWriter creates a new LZMA writer for the classic format. The
// method will write the header to the underlying stream.
func (c WriterConfig) NewWriter(lzma io.Writer) (w *Writer, err error) {
if err = c.Verify(); err != nil {
return nil, err
}
w = &Writer{h: c.header()}
var ok bool
w.bw, ok = lzma.(io.ByteWriter)
if !ok {
w.buf = bufio.NewWriter(lzma)
w.bw = w.buf
}
state := newState(w.h.properties)
m, err := c.Matcher.new(w.h.dictCap)
if err != nil {
return nil, err
}
dict, err := newEncoderDict(w.h.dictCap, c.BufSize, m)
if err != nil {
return nil, err
}
var flags encoderFlags
if c.EOSMarker {
flags = eosMarker
}
if w.e, err = newEncoder(w.bw, state, dict, flags); err != nil {
return nil, err
}
if err = w.writeHeader(); err != nil {
return nil, err
}
return w, nil
}
// NewWriter creates a new LZMA writer using the classic format. The
// function writes the header to the underlying stream.
func NewWriter(lzma io.Writer) (w *Writer, err error) {
return WriterConfig{}.NewWriter(lzma)
}
// writeHeader writes the LZMA header into the stream.
func (w *Writer) writeHeader() error {
data, err := w.h.marshalBinary()
if err != nil {
return err
}
_, err = w.bw.(io.Writer).Write(data)
return err
}
// Write puts data into the Writer.
func (w *Writer) Write(p []byte) (n int, err error) {
if w.h.size >= 0 {
m := w.h.size
m -= w.e.Compressed() + int64(w.e.dict.Buffered())
if m < 0 {
m = 0
}
if m < int64(len(p)) {
p = p[:m]
err = ErrNoSpace
}
}
var werr error
if n, werr = w.e.Write(p); werr != nil {
err = werr
}
return n, err
}
// Close closes the writer stream. It ensures that all data from the
// buffer will be compressed and the LZMA stream will be finished.
func (w *Writer) Close() error {
if w.h.size >= 0 {
n := w.e.Compressed() + int64(w.e.dict.Buffered())
if n != w.h.size {
return errSize
}
}
err := w.e.Close()
if w.buf != nil {
ferr := w.buf.Flush()
if err == nil {
err = ferr
}
}
return err
}

View File

@ -1,305 +0,0 @@
// Copyright 2014-2017 Ulrich Kunitz. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
package lzma
import (
"bytes"
"errors"
"io"
)
// Writer2Config is used to create a Writer2 using parameters.
type Writer2Config struct {
// The properties for the encoding. If the it is nil the value
// {LC: 3, LP: 0, PB: 2} will be chosen.
Properties *Properties
// The capacity of the dictionary. If DictCap is zero, the value
// 8 MiB will be chosen.
DictCap int
// Size of the lookahead buffer; value 0 indicates default size
// 4096
BufSize int
// Match algorithm
Matcher MatchAlgorithm
}
// fill replaces zero values with default values.
func (c *Writer2Config) fill() {
if c.Properties == nil {
c.Properties = &Properties{LC: 3, LP: 0, PB: 2}
}
if c.DictCap == 0 {
c.DictCap = 8 * 1024 * 1024
}
if c.BufSize == 0 {
c.BufSize = 4096
}
}
// Verify checks the Writer2Config for correctness. Zero values will be
// replaced by default values.
func (c *Writer2Config) Verify() error {
c.fill()
var err error
if c == nil {
return errors.New("lzma: WriterConfig is nil")
}
if c.Properties == nil {
return errors.New("lzma: WriterConfig has no Properties set")
}
if err = c.Properties.verify(); err != nil {
return err
}
if !(MinDictCap <= c.DictCap && int64(c.DictCap) <= MaxDictCap) {
return errors.New("lzma: dictionary capacity is out of range")
}
if !(maxMatchLen <= c.BufSize) {
return errors.New("lzma: lookahead buffer size too small")
}
if c.Properties.LC+c.Properties.LP > 4 {
return errors.New("lzma: sum of lc and lp exceeds 4")
}
if err = c.Matcher.verify(); err != nil {
return err
}
return nil
}
// Writer2 supports the creation of an LZMA2 stream. But note that
// written data is buffered, so call Flush or Close to write data to the
// underlying writer. The Close method writes the end-of-stream marker
// to the stream. So you may be able to concatenate the output of two
// writers as long the output of the first writer has only been flushed
// but not closed.
//
// Any change to the fields Properties, DictCap must be done before the
// first call to Write, Flush or Close.
type Writer2 struct {
w io.Writer
start *state
encoder *encoder
cstate chunkState
ctype chunkType
buf bytes.Buffer
lbw LimitedByteWriter
}
// NewWriter2 creates an LZMA2 chunk sequence writer with the default
// parameters and options.
func NewWriter2(lzma2 io.Writer) (w *Writer2, err error) {
return Writer2Config{}.NewWriter2(lzma2)
}
// NewWriter2 creates a new LZMA2 writer using the given configuration.
func (c Writer2Config) NewWriter2(lzma2 io.Writer) (w *Writer2, err error) {
if err = c.Verify(); err != nil {
return nil, err
}
w = &Writer2{
w: lzma2,
start: newState(*c.Properties),
cstate: start,
ctype: start.defaultChunkType(),
}
w.buf.Grow(maxCompressed)
w.lbw = LimitedByteWriter{BW: &w.buf, N: maxCompressed}
m, err := c.Matcher.new(c.DictCap)
if err != nil {
return nil, err
}
d, err := newEncoderDict(c.DictCap, c.BufSize, m)
if err != nil {
return nil, err
}
w.encoder, err = newEncoder(&w.lbw, cloneState(w.start), d, 0)
if err != nil {
return nil, err
}
return w, nil
}
// written returns the number of bytes written to the current chunk
func (w *Writer2) written() int {
if w.encoder == nil {
return 0
}
return int(w.encoder.Compressed()) + w.encoder.dict.Buffered()
}
// errClosed indicates that the writer is closed.
var errClosed = errors.New("lzma: writer closed")
// Writes data to LZMA2 stream. Note that written data will be buffered.
// Use Flush or Close to ensure that data is written to the underlying
// writer.
func (w *Writer2) Write(p []byte) (n int, err error) {
if w.cstate == stop {
return 0, errClosed
}
for n < len(p) {
m := maxUncompressed - w.written()
if m <= 0 {
panic("lzma: maxUncompressed reached")
}
var q []byte
if n+m < len(p) {
q = p[n : n+m]
} else {
q = p[n:]
}
k, err := w.encoder.Write(q)
n += k
if err != nil && err != ErrLimit {
return n, err
}
if err == ErrLimit || k == m {
if err = w.flushChunk(); err != nil {
return n, err
}
}
}
return n, nil
}
// writeUncompressedChunk writes an uncompressed chunk to the LZMA2
// stream.
func (w *Writer2) writeUncompressedChunk() error {
u := w.encoder.Compressed()
if u <= 0 {
return errors.New("lzma: can't write empty uncompressed chunk")
}
if u > maxUncompressed {
panic("overrun of uncompressed data limit")
}
switch w.ctype {
case cLRND:
w.ctype = cUD
default:
w.ctype = cU
}
w.encoder.state = w.start
header := chunkHeader{
ctype: w.ctype,
uncompressed: uint32(u - 1),
}
hdata, err := header.MarshalBinary()
if err != nil {
return err
}
if _, err = w.w.Write(hdata); err != nil {
return err
}
_, err = w.encoder.dict.CopyN(w.w, int(u))
return err
}
// writeCompressedChunk writes a compressed chunk to the underlying
// writer.
func (w *Writer2) writeCompressedChunk() error {
if w.ctype == cU || w.ctype == cUD {
panic("chunk type uncompressed")
}
u := w.encoder.Compressed()
if u <= 0 {
return errors.New("writeCompressedChunk: empty chunk")
}
if u > maxUncompressed {
panic("overrun of uncompressed data limit")
}
c := w.buf.Len()
if c <= 0 {
panic("no compressed data")
}
if c > maxCompressed {
panic("overrun of compressed data limit")
}
header := chunkHeader{
ctype: w.ctype,
uncompressed: uint32(u - 1),
compressed: uint16(c - 1),
props: w.encoder.state.Properties,
}
hdata, err := header.MarshalBinary()
if err != nil {
return err
}
if _, err = w.w.Write(hdata); err != nil {
return err
}
_, err = io.Copy(w.w, &w.buf)
return err
}
// writes a single chunk to the underlying writer.
func (w *Writer2) writeChunk() error {
u := int(uncompressedHeaderLen + w.encoder.Compressed())
c := headerLen(w.ctype) + w.buf.Len()
if u < c {
return w.writeUncompressedChunk()
}
return w.writeCompressedChunk()
}
// flushChunk terminates the current chunk. The encoder will be reset
// to support the next chunk.
func (w *Writer2) flushChunk() error {
if w.written() == 0 {
return nil
}
var err error
if err = w.encoder.Close(); err != nil {
return err
}
if err = w.writeChunk(); err != nil {
return err
}
w.buf.Reset()
w.lbw.N = maxCompressed
if err = w.encoder.Reopen(&w.lbw); err != nil {
return err
}
if err = w.cstate.next(w.ctype); err != nil {
return err
}
w.ctype = w.cstate.defaultChunkType()
w.start = cloneState(w.encoder.state)
return nil
}
// Flush writes all buffered data out to the underlying stream. This
// could result in multiple chunks to be created.
func (w *Writer2) Flush() error {
if w.cstate == stop {
return errClosed
}
for w.written() > 0 {
if err := w.flushChunk(); err != nil {
return err
}
}
return nil
}
// Close terminates the LZMA2 stream with an EOS chunk.
func (w *Writer2) Close() error {
if w.cstate == stop {
return errClosed
}
if err := w.Flush(); err != nil {
return nil
}
// write zero byte EOS chunk
_, err := w.w.Write([]byte{0})
if err != nil {
return err
}
w.cstate = stop
return nil
}

View File

@ -1,117 +0,0 @@
// Copyright 2014-2017 Ulrich Kunitz. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
package xz
import (
"errors"
"fmt"
"io"
"github.com/ulikunitz/xz/lzma"
)
// LZMA filter constants.
const (
lzmaFilterID = 0x21
lzmaFilterLen = 3
)
// lzmaFilter declares the LZMA2 filter information stored in an xz
// block header.
type lzmaFilter struct {
dictCap int64
}
// String returns a representation of the LZMA filter.
func (f lzmaFilter) String() string {
return fmt.Sprintf("LZMA dict cap %#x", f.dictCap)
}
// id returns the ID for the LZMA2 filter.
func (f lzmaFilter) id() uint64 { return lzmaFilterID }
// MarshalBinary converts the lzmaFilter in its encoded representation.
func (f lzmaFilter) MarshalBinary() (data []byte, err error) {
c := lzma.EncodeDictCap(f.dictCap)
return []byte{lzmaFilterID, 1, c}, nil
}
// UnmarshalBinary unmarshals the given data representation of the LZMA2
// filter.
func (f *lzmaFilter) UnmarshalBinary(data []byte) error {
if len(data) != lzmaFilterLen {
return errors.New("xz: data for LZMA2 filter has wrong length")
}
if data[0] != lzmaFilterID {
return errors.New("xz: wrong LZMA2 filter id")
}
if data[1] != 1 {
return errors.New("xz: wrong LZMA2 filter size")
}
dc, err := lzma.DecodeDictCap(data[2])
if err != nil {
return errors.New("xz: wrong LZMA2 dictionary size property")
}
f.dictCap = dc
return nil
}
// reader creates a new reader for the LZMA2 filter.
func (f lzmaFilter) reader(r io.Reader, c *ReaderConfig) (fr io.Reader,
err error) {
config := new(lzma.Reader2Config)
if c != nil {
config.DictCap = c.DictCap
}
dc := int(f.dictCap)
if dc < 1 {
return nil, errors.New("xz: LZMA2 filter parameter " +
"dictionary capacity overflow")
}
if dc > config.DictCap {
config.DictCap = dc
}
fr, err = config.NewReader2(r)
if err != nil {
return nil, err
}
return fr, nil
}
// writeCloser creates a io.WriteCloser for the LZMA2 filter.
func (f lzmaFilter) writeCloser(w io.WriteCloser, c *WriterConfig,
) (fw io.WriteCloser, err error) {
config := new(lzma.Writer2Config)
if c != nil {
*config = lzma.Writer2Config{
Properties: c.Properties,
DictCap: c.DictCap,
BufSize: c.BufSize,
Matcher: c.Matcher,
}
}
dc := int(f.dictCap)
if dc < 1 {
return nil, errors.New("xz: LZMA2 filter parameter " +
"dictionary capacity overflow")
}
if dc > config.DictCap {
config.DictCap = dc
}
fw, err = config.NewWriter2(w)
if err != nil {
return nil, err
}
return fw, nil
}
// last returns true, because an LZMA2 filter must be the last filter in
// the filter list.
func (f lzmaFilter) last() bool { return true }

View File

@ -1,5 +0,0 @@
#!/bin/sh
set -x
pandoc -t html5 -f markdown -s --css=doc/md.css -o README.html README.md
pandoc -t html5 -f markdown -s --css=doc/md.css -o TODO.html TODO.md

View File

@ -1,373 +0,0 @@
// Copyright 2014-2017 Ulrich Kunitz. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
// Package xz supports the compression and decompression of xz files. It
// supports version 1.0.4 of the specification without the non-LZMA2
// filters. See http://tukaani.org/xz/xz-file-format-1.0.4.txt
package xz
import (
"bytes"
"errors"
"fmt"
"hash"
"io"
"github.com/ulikunitz/xz/internal/xlog"
"github.com/ulikunitz/xz/lzma"
)
// ReaderConfig defines the parameters for the xz reader. The
// SingleStream parameter requests the reader to assume that the
// underlying stream contains only a single stream.
type ReaderConfig struct {
DictCap int
SingleStream bool
}
// fill replaces all zero values with their default values.
func (c *ReaderConfig) fill() {
if c.DictCap == 0 {
c.DictCap = 8 * 1024 * 1024
}
}
// Verify checks the reader parameters for Validity. Zero values will be
// replaced by default values.
func (c *ReaderConfig) Verify() error {
if c == nil {
return errors.New("xz: reader parameters are nil")
}
lc := lzma.Reader2Config{DictCap: c.DictCap}
if err := lc.Verify(); err != nil {
return err
}
return nil
}
// Reader supports the reading of one or multiple xz streams.
type Reader struct {
ReaderConfig
xz io.Reader
sr *streamReader
}
// streamReader decodes a single xz stream
type streamReader struct {
ReaderConfig
xz io.Reader
br *blockReader
newHash func() hash.Hash
h header
index []record
}
// NewReader creates a new xz reader using the default parameters.
// The function reads and checks the header of the first XZ stream. The
// reader will process multiple streams including padding.
func NewReader(xz io.Reader) (r *Reader, err error) {
return ReaderConfig{}.NewReader(xz)
}
// NewReader creates an xz stream reader. The created reader will be
// able to process multiple streams and padding unless a SingleStream
// has been set in the reader configuration c.
func (c ReaderConfig) NewReader(xz io.Reader) (r *Reader, err error) {
if err = c.Verify(); err != nil {
return nil, err
}
r = &Reader{
ReaderConfig: c,
xz: xz,
}
if r.sr, err = c.newStreamReader(xz); err != nil {
if err == io.EOF {
err = io.ErrUnexpectedEOF
}
return nil, err
}
return r, nil
}
var errUnexpectedData = errors.New("xz: unexpected data after stream")
// Read reads uncompressed data from the stream.
func (r *Reader) Read(p []byte) (n int, err error) {
for n < len(p) {
if r.sr == nil {
if r.SingleStream {
data := make([]byte, 1)
_, err = io.ReadFull(r.xz, data)
if err != io.EOF {
return n, errUnexpectedData
}
return n, io.EOF
}
for {
r.sr, err = r.ReaderConfig.newStreamReader(r.xz)
if err != errPadding {
break
}
}
if err != nil {
return n, err
}
}
k, err := r.sr.Read(p[n:])
n += k
if err != nil {
if err == io.EOF {
r.sr = nil
continue
}
return n, err
}
}
return n, nil
}
var errPadding = errors.New("xz: padding (4 zero bytes) encountered")
// newStreamReader creates a new xz stream reader using the given configuration
// parameters. NewReader reads and checks the header of the xz stream.
func (c ReaderConfig) newStreamReader(xz io.Reader) (r *streamReader, err error) {
if err = c.Verify(); err != nil {
return nil, err
}
data := make([]byte, HeaderLen)
if _, err := io.ReadFull(xz, data[:4]); err != nil {
return nil, err
}
if bytes.Equal(data[:4], []byte{0, 0, 0, 0}) {
return nil, errPadding
}
if _, err = io.ReadFull(xz, data[4:]); err != nil {
if err == io.EOF {
err = io.ErrUnexpectedEOF
}
return nil, err
}
r = &streamReader{
ReaderConfig: c,
xz: xz,
index: make([]record, 0, 4),
}
if err = r.h.UnmarshalBinary(data); err != nil {
return nil, err
}
xlog.Debugf("xz header %s", r.h)
if r.newHash, err = newHashFunc(r.h.flags); err != nil {
return nil, err
}
return r, nil
}
// errIndex indicates an error with the xz file index.
var errIndex = errors.New("xz: error in xz file index")
// readTail reads the index body and the xz footer.
func (r *streamReader) readTail() error {
index, n, err := readIndexBody(r.xz)
if err != nil {
if err == io.EOF {
err = io.ErrUnexpectedEOF
}
return err
}
if len(index) != len(r.index) {
return fmt.Errorf("xz: index length is %d; want %d",
len(index), len(r.index))
}
for i, rec := range r.index {
if rec != index[i] {
return fmt.Errorf("xz: record %d is %v; want %v",
i, rec, index[i])
}
}
p := make([]byte, footerLen)
if _, err = io.ReadFull(r.xz, p); err != nil {
if err == io.EOF {
err = io.ErrUnexpectedEOF
}
return err
}
var f footer
if err = f.UnmarshalBinary(p); err != nil {
return err
}
xlog.Debugf("xz footer %s", f)
if f.flags != r.h.flags {
return errors.New("xz: footer flags incorrect")
}
if f.indexSize != int64(n)+1 {
return errors.New("xz: index size in footer wrong")
}
return nil
}
// Read reads actual data from the xz stream.
func (r *streamReader) Read(p []byte) (n int, err error) {
for n < len(p) {
if r.br == nil {
bh, hlen, err := readBlockHeader(r.xz)
if err != nil {
if err == errIndexIndicator {
if err = r.readTail(); err != nil {
return n, err
}
return n, io.EOF
}
return n, err
}
xlog.Debugf("block %v", *bh)
r.br, err = r.ReaderConfig.newBlockReader(r.xz, bh,
hlen, r.newHash())
if err != nil {
return n, err
}
}
k, err := r.br.Read(p[n:])
n += k
if err != nil {
if err == io.EOF {
r.index = append(r.index, r.br.record())
r.br = nil
} else {
return n, err
}
}
}
return n, nil
}
// countingReader is a reader that counts the bytes read.
type countingReader struct {
r io.Reader
n int64
}
// Read reads data from the wrapped reader and adds it to the n field.
func (lr *countingReader) Read(p []byte) (n int, err error) {
n, err = lr.r.Read(p)
lr.n += int64(n)
return n, err
}
// blockReader supports the reading of a block.
type blockReader struct {
lxz countingReader
header *blockHeader
headerLen int
n int64
hash hash.Hash
r io.Reader
err error
}
// newBlockReader creates a new block reader.
func (c *ReaderConfig) newBlockReader(xz io.Reader, h *blockHeader,
hlen int, hash hash.Hash) (br *blockReader, err error) {
br = &blockReader{
lxz: countingReader{r: xz},
header: h,
headerLen: hlen,
hash: hash,
}
fr, err := c.newFilterReader(&br.lxz, h.filters)
if err != nil {
return nil, err
}
br.r = io.TeeReader(fr, br.hash)
return br, nil
}
// uncompressedSize returns the uncompressed size of the block.
func (br *blockReader) uncompressedSize() int64 {
return br.n
}
// compressedSize returns the compressed size of the block.
func (br *blockReader) compressedSize() int64 {
return br.lxz.n
}
// unpaddedSize computes the unpadded size for the block.
func (br *blockReader) unpaddedSize() int64 {
n := int64(br.headerLen)
n += br.compressedSize()
n += int64(br.hash.Size())
return n
}
// record returns the index record for the current block.
func (br *blockReader) record() record {
return record{br.unpaddedSize(), br.uncompressedSize()}
}
// errBlockSize indicates that the size of the block in the block header
// is wrong.
var errBlockSize = errors.New("xz: wrong uncompressed size for block")
// Read reads data from the block.
func (br *blockReader) Read(p []byte) (n int, err error) {
n, err = br.r.Read(p)
br.n += int64(n)
u := br.header.uncompressedSize
if u >= 0 && br.uncompressedSize() > u {
return n, errors.New("xz: wrong uncompressed size for block")
}
c := br.header.compressedSize
if c >= 0 && br.compressedSize() > c {
return n, errors.New("xz: wrong compressed size for block")
}
if err != io.EOF {
return n, err
}
if br.uncompressedSize() < u || br.compressedSize() < c {
return n, io.ErrUnexpectedEOF
}
s := br.hash.Size()
k := padLen(br.lxz.n)
q := make([]byte, k+s, k+2*s)
if _, err = io.ReadFull(br.lxz.r, q); err != nil {
if err == io.EOF {
err = io.ErrUnexpectedEOF
}
return n, err
}
if !allZeros(q[:k]) {
return n, errors.New("xz: non-zero block padding")
}
checkSum := q[k:]
computedSum := br.hash.Sum(checkSum[s:])
if !bytes.Equal(checkSum, computedSum) {
return n, errors.New("xz: checksum error for block")
}
return n, io.EOF
}
func (c *ReaderConfig) newFilterReader(r io.Reader, f []filter) (fr io.Reader,
err error) {
if err = verifyFilters(f); err != nil {
return nil, err
}
fr = r
for i := len(f) - 1; i >= 0; i-- {
fr, err = f[i].reader(fr, c)
if err != nil {
return nil, err
}
}
return fr, nil
}

View File

@ -1,386 +0,0 @@
// Copyright 2014-2017 Ulrich Kunitz. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
package xz
import (
"errors"
"hash"
"io"
"github.com/ulikunitz/xz/lzma"
)
// WriterConfig describe the parameters for an xz writer.
type WriterConfig struct {
Properties *lzma.Properties
DictCap int
BufSize int
BlockSize int64
// checksum method: CRC32, CRC64 or SHA256
CheckSum byte
// match algorithm
Matcher lzma.MatchAlgorithm
}
// fill replaces zero values with default values.
func (c *WriterConfig) fill() {
if c.Properties == nil {
c.Properties = &lzma.Properties{LC: 3, LP: 0, PB: 2}
}
if c.DictCap == 0 {
c.DictCap = 8 * 1024 * 1024
}
if c.BufSize == 0 {
c.BufSize = 4096
}
if c.BlockSize == 0 {
c.BlockSize = maxInt64
}
if c.CheckSum == 0 {
c.CheckSum = CRC64
}
}
// Verify checks the configuration for errors. Zero values will be
// replaced by default values.
func (c *WriterConfig) Verify() error {
if c == nil {
return errors.New("xz: writer configuration is nil")
}
c.fill()
lc := lzma.Writer2Config{
Properties: c.Properties,
DictCap: c.DictCap,
BufSize: c.BufSize,
Matcher: c.Matcher,
}
if err := lc.Verify(); err != nil {
return err
}
if c.BlockSize <= 0 {
return errors.New("xz: block size out of range")
}
if err := verifyFlags(c.CheckSum); err != nil {
return err
}
return nil
}
// filters creates the filter list for the given parameters.
func (c *WriterConfig) filters() []filter {
return []filter{&lzmaFilter{int64(c.DictCap)}}
}
// maxInt64 defines the maximum 64-bit signed integer.
const maxInt64 = 1<<63 - 1
// verifyFilters checks the filter list for the length and the right
// sequence of filters.
func verifyFilters(f []filter) error {
if len(f) == 0 {
return errors.New("xz: no filters")
}
if len(f) > 4 {
return errors.New("xz: more than four filters")
}
for _, g := range f[:len(f)-1] {
if g.last() {
return errors.New("xz: last filter is not last")
}
}
if !f[len(f)-1].last() {
return errors.New("xz: wrong last filter")
}
return nil
}
// newFilterWriteCloser converts a filter list into a WriteCloser that
// can be used by a blockWriter.
func (c *WriterConfig) newFilterWriteCloser(w io.Writer, f []filter) (fw io.WriteCloser, err error) {
if err = verifyFilters(f); err != nil {
return nil, err
}
fw = nopWriteCloser(w)
for i := len(f) - 1; i >= 0; i-- {
fw, err = f[i].writeCloser(fw, c)
if err != nil {
return nil, err
}
}
return fw, nil
}
// nopWCloser implements a WriteCloser with a Close method not doing
// anything.
type nopWCloser struct {
io.Writer
}
// Close returns nil and doesn't do anything else.
func (c nopWCloser) Close() error {
return nil
}
// nopWriteCloser converts the Writer into a WriteCloser with a Close
// function that does nothing beside returning nil.
func nopWriteCloser(w io.Writer) io.WriteCloser {
return nopWCloser{w}
}
// Writer compresses data written to it. It is an io.WriteCloser.
type Writer struct {
WriterConfig
xz io.Writer
bw *blockWriter
newHash func() hash.Hash
h header
index []record
closed bool
}
// newBlockWriter creates a new block writer writes the header out.
func (w *Writer) newBlockWriter() error {
var err error
w.bw, err = w.WriterConfig.newBlockWriter(w.xz, w.newHash())
if err != nil {
return err
}
if err = w.bw.writeHeader(w.xz); err != nil {
return err
}
return nil
}
// closeBlockWriter closes a block writer and records the sizes in the
// index.
func (w *Writer) closeBlockWriter() error {
var err error
if err = w.bw.Close(); err != nil {
return err
}
w.index = append(w.index, w.bw.record())
return nil
}
// NewWriter creates a new xz writer using default parameters.
func NewWriter(xz io.Writer) (w *Writer, err error) {
return WriterConfig{}.NewWriter(xz)
}
// NewWriter creates a new Writer using the given configuration parameters.
func (c WriterConfig) NewWriter(xz io.Writer) (w *Writer, err error) {
if err = c.Verify(); err != nil {
return nil, err
}
w = &Writer{
WriterConfig: c,
xz: xz,
h: header{c.CheckSum},
index: make([]record, 0, 4),
}
if w.newHash, err = newHashFunc(c.CheckSum); err != nil {
return nil, err
}
data, err := w.h.MarshalBinary()
if _, err = xz.Write(data); err != nil {
return nil, err
}
if err = w.newBlockWriter(); err != nil {
return nil, err
}
return w, nil
}
// Write compresses the uncompressed data provided.
func (w *Writer) Write(p []byte) (n int, err error) {
if w.closed {
return 0, errClosed
}
for {
k, err := w.bw.Write(p[n:])
n += k
if err != errNoSpace {
return n, err
}
if err = w.closeBlockWriter(); err != nil {
return n, err
}
if err = w.newBlockWriter(); err != nil {
return n, err
}
}
}
// Close closes the writer and adds the footer to the Writer. Close
// doesn't close the underlying writer.
func (w *Writer) Close() error {
if w.closed {
return errClosed
}
w.closed = true
var err error
if err = w.closeBlockWriter(); err != nil {
return err
}
f := footer{flags: w.h.flags}
if f.indexSize, err = writeIndex(w.xz, w.index); err != nil {
return err
}
data, err := f.MarshalBinary()
if err != nil {
return err
}
if _, err = w.xz.Write(data); err != nil {
return err
}
return nil
}
// countingWriter is a writer that counts all data written to it.
type countingWriter struct {
w io.Writer
n int64
}
// Write writes data to the countingWriter.
func (cw *countingWriter) Write(p []byte) (n int, err error) {
n, err = cw.w.Write(p)
cw.n += int64(n)
if err == nil && cw.n < 0 {
return n, errors.New("xz: counter overflow")
}
return
}
// blockWriter is writes a single block.
type blockWriter struct {
cxz countingWriter
// mw combines io.WriteCloser w and the hash.
mw io.Writer
w io.WriteCloser
n int64
blockSize int64
closed bool
headerLen int
filters []filter
hash hash.Hash
}
// newBlockWriter creates a new block writer.
func (c *WriterConfig) newBlockWriter(xz io.Writer, hash hash.Hash) (bw *blockWriter, err error) {
bw = &blockWriter{
cxz: countingWriter{w: xz},
blockSize: c.BlockSize,
filters: c.filters(),
hash: hash,
}
bw.w, err = c.newFilterWriteCloser(&bw.cxz, bw.filters)
if err != nil {
return nil, err
}
bw.mw = io.MultiWriter(bw.w, bw.hash)
return bw, nil
}
// writeHeader writes the header. If the function is called after Close
// the commpressedSize and uncompressedSize fields will be filled.
func (bw *blockWriter) writeHeader(w io.Writer) error {
h := blockHeader{
compressedSize: -1,
uncompressedSize: -1,
filters: bw.filters,
}
if bw.closed {
h.compressedSize = bw.compressedSize()
h.uncompressedSize = bw.uncompressedSize()
}
data, err := h.MarshalBinary()
if err != nil {
return err
}
if _, err = w.Write(data); err != nil {
return err
}
bw.headerLen = len(data)
return nil
}
// compressed size returns the amount of data written to the underlying
// stream.
func (bw *blockWriter) compressedSize() int64 {
return bw.cxz.n
}
// uncompressedSize returns the number of data written to the
// blockWriter
func (bw *blockWriter) uncompressedSize() int64 {
return bw.n
}
// unpaddedSize returns the sum of the header length, the uncompressed
// size of the block and the hash size.
func (bw *blockWriter) unpaddedSize() int64 {
if bw.headerLen <= 0 {
panic("xz: block header not written")
}
n := int64(bw.headerLen)
n += bw.compressedSize()
n += int64(bw.hash.Size())
return n
}
// record returns the record for the current stream. Call Close before
// calling this method.
func (bw *blockWriter) record() record {
return record{bw.unpaddedSize(), bw.uncompressedSize()}
}
var errClosed = errors.New("xz: writer already closed")
var errNoSpace = errors.New("xz: no space")
// Write writes uncompressed data to the block writer.
func (bw *blockWriter) Write(p []byte) (n int, err error) {
if bw.closed {
return 0, errClosed
}
t := bw.blockSize - bw.n
if int64(len(p)) > t {
err = errNoSpace
p = p[:t]
}
var werr error
n, werr = bw.mw.Write(p)
bw.n += int64(n)
if werr != nil {
return n, werr
}
return n, err
}
// Close closes the writer.
func (bw *blockWriter) Close() error {
if bw.closed {
return errClosed
}
bw.closed = true
if err := bw.w.Close(); err != nil {
return err
}
s := bw.hash.Size()
k := padLen(bw.cxz.n)
p := make([]byte, k+s)
bw.hash.Sum(p[k:k])
if _, err := bw.cxz.w.Write(p); err != nil {
return err
}
return nil
}

12
vendor/gopkg.in/yaml.v2/.travis.yml generated vendored
View File

@ -1,12 +0,0 @@
language: go
go:
- 1.4
- 1.5
- 1.6
- 1.7
- 1.8
- 1.9
- tip
go_import_path: gopkg.in/yaml.v2

201
vendor/gopkg.in/yaml.v2/LICENSE generated vendored
View File

@ -1,201 +0,0 @@
Apache License
Version 2.0, January 2004
http://www.apache.org/licenses/
TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
1. Definitions.
"License" shall mean the terms and conditions for use, reproduction,
and distribution as defined by Sections 1 through 9 of this document.
"Licensor" shall mean the copyright owner or entity authorized by
the copyright owner that is granting the License.
"Legal Entity" shall mean the union of the acting entity and all
other entities that control, are controlled by, or are under common
control with that entity. For the purposes of this definition,
"control" means (i) the power, direct or indirect, to cause the
direction or management of such entity, whether by contract or
otherwise, or (ii) ownership of fifty percent (50%) or more of the
outstanding shares, or (iii) beneficial ownership of such entity.
"You" (or "Your") shall mean an individual or Legal Entity
exercising permissions granted by this License.
"Source" form shall mean the preferred form for making modifications,
including but not limited to software source code, documentation
source, and configuration files.
"Object" form shall mean any form resulting from mechanical
transformation or translation of a Source form, including but
not limited to compiled object code, generated documentation,
and conversions to other media types.
"Work" shall mean the work of authorship, whether in Source or
Object form, made available under the License, as indicated by a
copyright notice that is included in or attached to the work
(an example is provided in the Appendix below).
"Derivative Works" shall mean any work, whether in Source or Object
form, that is based on (or derived from) the Work and for which the
editorial revisions, annotations, elaborations, or other modifications
represent, as a whole, an original work of authorship. For the purposes
of this License, Derivative Works shall not include works that remain
separable from, or merely link (or bind by name) to the interfaces of,
the Work and Derivative Works thereof.
"Contribution" shall mean any work of authorship, including
the original version of the Work and any modifications or additions
to that Work or Derivative Works thereof, that is intentionally
submitted to Licensor for inclusion in the Work by the copyright owner
or by an individual or Legal Entity authorized to submit on behalf of
the copyright owner. For the purposes of this definition, "submitted"
means any form of electronic, verbal, or written communication sent
to the Licensor or its representatives, including but not limited to
communication on electronic mailing lists, source code control systems,
and issue tracking systems that are managed by, or on behalf of, the
Licensor for the purpose of discussing and improving the Work, but
excluding communication that is conspicuously marked or otherwise
designated in writing by the copyright owner as "Not a Contribution."
"Contributor" shall mean Licensor and any individual or Legal Entity
on behalf of whom a Contribution has been received by Licensor and
subsequently incorporated within the Work.
2. Grant of Copyright License. Subject to the terms and conditions of
this License, each Contributor hereby grants to You a perpetual,
worldwide, non-exclusive, no-charge, royalty-free, irrevocable
copyright license to reproduce, prepare Derivative Works of,
publicly display, publicly perform, sublicense, and distribute the
Work and such Derivative Works in Source or Object form.
3. Grant of Patent License. Subject to the terms and conditions of
this License, each Contributor hereby grants to You a perpetual,
worldwide, non-exclusive, no-charge, royalty-free, irrevocable
(except as stated in this section) patent license to make, have made,
use, offer to sell, sell, import, and otherwise transfer the Work,
where such license applies only to those patent claims licensable
by such Contributor that are necessarily infringed by their
Contribution(s) alone or by combination of their Contribution(s)
with the Work to which such Contribution(s) was submitted. If You
institute patent litigation against any entity (including a
cross-claim or counterclaim in a lawsuit) alleging that the Work
or a Contribution incorporated within the Work constitutes direct
or contributory patent infringement, then any patent licenses
granted to You under this License for that Work shall terminate
as of the date such litigation is filed.
4. Redistribution. You may reproduce and distribute copies of the
Work or Derivative Works thereof in any medium, with or without
modifications, and in Source or Object form, provided that You
meet the following conditions:
(a) You must give any other recipients of the Work or
Derivative Works a copy of this License; and
(b) You must cause any modified files to carry prominent notices
stating that You changed the files; and
(c) You must retain, in the Source form of any Derivative Works
that You distribute, all copyright, patent, trademark, and
attribution notices from the Source form of the Work,
excluding those notices that do not pertain to any part of
the Derivative Works; and
(d) If the Work includes a "NOTICE" text file as part of its
distribution, then any Derivative Works that You distribute must
include a readable copy of the attribution notices contained
within such NOTICE file, excluding those notices that do not
pertain to any part of the Derivative Works, in at least one
of the following places: within a NOTICE text file distributed
as part of the Derivative Works; within the Source form or
documentation, if provided along with the Derivative Works; or,
within a display generated by the Derivative Works, if and
wherever such third-party notices normally appear. The contents
of the NOTICE file are for informational purposes only and
do not modify the License. You may add Your own attribution
notices within Derivative Works that You distribute, alongside
or as an addendum to the NOTICE text from the Work, provided
that such additional attribution notices cannot be construed
as modifying the License.
You may add Your own copyright statement to Your modifications and
may provide additional or different license terms and conditions
for use, reproduction, or distribution of Your modifications, or
for any such Derivative Works as a whole, provided Your use,
reproduction, and distribution of the Work otherwise complies with
the conditions stated in this License.
5. Submission of Contributions. Unless You explicitly state otherwise,
any Contribution intentionally submitted for inclusion in the Work
by You to the Licensor shall be under the terms and conditions of
this License, without any additional terms or conditions.
Notwithstanding the above, nothing herein shall supersede or modify
the terms of any separate license agreement you may have executed
with Licensor regarding such Contributions.
6. Trademarks. This License does not grant permission to use the trade
names, trademarks, service marks, or product names of the Licensor,
except as required for reasonable and customary use in describing the
origin of the Work and reproducing the content of the NOTICE file.
7. Disclaimer of Warranty. Unless required by applicable law or
agreed to in writing, Licensor provides the Work (and each
Contributor provides its Contributions) on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
implied, including, without limitation, any warranties or conditions
of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
PARTICULAR PURPOSE. You are solely responsible for determining the
appropriateness of using or redistributing the Work and assume any
risks associated with Your exercise of permissions under this License.
8. Limitation of Liability. In no event and under no legal theory,
whether in tort (including negligence), contract, or otherwise,
unless required by applicable law (such as deliberate and grossly
negligent acts) or agreed to in writing, shall any Contributor be
liable to You for damages, including any direct, indirect, special,
incidental, or consequential damages of any character arising as a
result of this License or out of the use or inability to use the
Work (including but not limited to damages for loss of goodwill,
work stoppage, computer failure or malfunction, or any and all
other commercial damages or losses), even if such Contributor
has been advised of the possibility of such damages.
9. Accepting Warranty or Additional Liability. While redistributing
the Work or Derivative Works thereof, You may choose to offer,
and charge a fee for, acceptance of support, warranty, indemnity,
or other liability obligations and/or rights consistent with this
License. However, in accepting such obligations, You may act only
on Your own behalf and on Your sole responsibility, not on behalf
of any other Contributor, and only if You agree to indemnify,
defend, and hold each Contributor harmless for any liability
incurred by, or claims asserted against, such Contributor by reason
of your accepting any such warranty or additional liability.
END OF TERMS AND CONDITIONS
APPENDIX: How to apply the Apache License to your work.
To apply the Apache License to your work, attach the following
boilerplate notice, with the fields enclosed by brackets "{}"
replaced with your own identifying information. (Don't include
the brackets!) The text should be enclosed in the appropriate
comment syntax for the file format. We also recommend that a
file or class name and description of purpose be included on the
same "printed page" as the copyright notice for easier
identification within third-party archives.
Copyright {yyyy} {name of copyright owner}
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.

View File

@ -1,31 +0,0 @@
The following files were ported to Go from C files of libyaml, and thus
are still covered by their original copyright and license:
apic.go
emitterc.go
parserc.go
readerc.go
scannerc.go
writerc.go
yamlh.go
yamlprivateh.go
Copyright (c) 2006 Kirill Simonov
Permission is hereby granted, free of charge, to any person obtaining a copy of
this software and associated documentation files (the "Software"), to deal in
the Software without restriction, including without limitation the rights to
use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies
of the Software, and to permit persons to whom the Software is furnished to do
so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.

13
vendor/gopkg.in/yaml.v2/NOTICE generated vendored
View File

@ -1,13 +0,0 @@
Copyright 2011-2016 Canonical Ltd.
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.

133
vendor/gopkg.in/yaml.v2/README.md generated vendored
View File

@ -1,133 +0,0 @@
# YAML support for the Go language
Introduction
------------
The yaml package enables Go programs to comfortably encode and decode YAML
values. It was developed within [Canonical](https://www.canonical.com) as
part of the [juju](https://juju.ubuntu.com) project, and is based on a
pure Go port of the well-known [libyaml](http://pyyaml.org/wiki/LibYAML)
C library to parse and generate YAML data quickly and reliably.
Compatibility
-------------
The yaml package supports most of YAML 1.1 and 1.2, including support for
anchors, tags, map merging, etc. Multi-document unmarshalling is not yet
implemented, and base-60 floats from YAML 1.1 are purposefully not
supported since they're a poor design and are gone in YAML 1.2.
Installation and usage
----------------------
The import path for the package is *gopkg.in/yaml.v2*.
To install it, run:
go get gopkg.in/yaml.v2
API documentation
-----------------
If opened in a browser, the import path itself leads to the API documentation:
* [https://gopkg.in/yaml.v2](https://gopkg.in/yaml.v2)
API stability
-------------
The package API for yaml v2 will remain stable as described in [gopkg.in](https://gopkg.in).
License
-------
The yaml package is licensed under the Apache License 2.0. Please see the LICENSE file for details.
Example
-------
```Go
package main
import (
"fmt"
"log"
"gopkg.in/yaml.v2"
)
var data = `
a: Easy!
b:
c: 2
d: [3, 4]
`
// Note: struct fields must be public in order for unmarshal to
// correctly populate the data.
type T struct {
A string
B struct {
RenamedC int `yaml:"c"`
D []int `yaml:",flow"`
}
}
func main() {
t := T{}
err := yaml.Unmarshal([]byte(data), &t)
if err != nil {
log.Fatalf("error: %v", err)
}
fmt.Printf("--- t:\n%v\n\n", t)
d, err := yaml.Marshal(&t)
if err != nil {
log.Fatalf("error: %v", err)
}
fmt.Printf("--- t dump:\n%s\n\n", string(d))
m := make(map[interface{}]interface{})
err = yaml.Unmarshal([]byte(data), &m)
if err != nil {
log.Fatalf("error: %v", err)
}
fmt.Printf("--- m:\n%v\n\n", m)
d, err = yaml.Marshal(&m)
if err != nil {
log.Fatalf("error: %v", err)
}
fmt.Printf("--- m dump:\n%s\n\n", string(d))
}
```
This example will generate the following output:
```
--- t:
{Easy! {2 [3 4]}}
--- t dump:
a: Easy!
b:
c: 2
d: [3, 4]
--- m:
map[a:Easy! b:map[c:2 d:[3 4]]]
--- m dump:
a: Easy!
b:
c: 2
d:
- 3
- 4
```

739
vendor/gopkg.in/yaml.v2/apic.go generated vendored
View File

@ -1,739 +0,0 @@
package yaml
import (
"io"
)
func yaml_insert_token(parser *yaml_parser_t, pos int, token *yaml_token_t) {
//fmt.Println("yaml_insert_token", "pos:", pos, "typ:", token.typ, "head:", parser.tokens_head, "len:", len(parser.tokens))
// Check if we can move the queue at the beginning of the buffer.
if parser.tokens_head > 0 && len(parser.tokens) == cap(parser.tokens) {
if parser.tokens_head != len(parser.tokens) {
copy(parser.tokens, parser.tokens[parser.tokens_head:])
}
parser.tokens = parser.tokens[:len(parser.tokens)-parser.tokens_head]
parser.tokens_head = 0
}
parser.tokens = append(parser.tokens, *token)
if pos < 0 {
return
}
copy(parser.tokens[parser.tokens_head+pos+1:], parser.tokens[parser.tokens_head+pos:])
parser.tokens[parser.tokens_head+pos] = *token
}
// Create a new parser object.
func yaml_parser_initialize(parser *yaml_parser_t) bool {
*parser = yaml_parser_t{
raw_buffer: make([]byte, 0, input_raw_buffer_size),
buffer: make([]byte, 0, input_buffer_size),
}
return true
}
// Destroy a parser object.
func yaml_parser_delete(parser *yaml_parser_t) {
*parser = yaml_parser_t{}
}
// String read handler.
func yaml_string_read_handler(parser *yaml_parser_t, buffer []byte) (n int, err error) {
if parser.input_pos == len(parser.input) {
return 0, io.EOF
}
n = copy(buffer, parser.input[parser.input_pos:])
parser.input_pos += n
return n, nil
}
// Reader read handler.
func yaml_reader_read_handler(parser *yaml_parser_t, buffer []byte) (n int, err error) {
return parser.input_reader.Read(buffer)
}
// Set a string input.
func yaml_parser_set_input_string(parser *yaml_parser_t, input []byte) {
if parser.read_handler != nil {
panic("must set the input source only once")
}
parser.read_handler = yaml_string_read_handler
parser.input = input
parser.input_pos = 0
}
// Set a file input.
func yaml_parser_set_input_reader(parser *yaml_parser_t, r io.Reader) {
if parser.read_handler != nil {
panic("must set the input source only once")
}
parser.read_handler = yaml_reader_read_handler
parser.input_reader = r
}
// Set the source encoding.
func yaml_parser_set_encoding(parser *yaml_parser_t, encoding yaml_encoding_t) {
if parser.encoding != yaml_ANY_ENCODING {
panic("must set the encoding only once")
}
parser.encoding = encoding
}
// Create a new emitter object.
func yaml_emitter_initialize(emitter *yaml_emitter_t) {
*emitter = yaml_emitter_t{
buffer: make([]byte, output_buffer_size),
raw_buffer: make([]byte, 0, output_raw_buffer_size),
states: make([]yaml_emitter_state_t, 0, initial_stack_size),
events: make([]yaml_event_t, 0, initial_queue_size),
}
}
// Destroy an emitter object.
func yaml_emitter_delete(emitter *yaml_emitter_t) {
*emitter = yaml_emitter_t{}
}
// String write handler.
func yaml_string_write_handler(emitter *yaml_emitter_t, buffer []byte) error {
*emitter.output_buffer = append(*emitter.output_buffer, buffer...)
return nil
}
// yaml_writer_write_handler uses emitter.output_writer to write the
// emitted text.
func yaml_writer_write_handler(emitter *yaml_emitter_t, buffer []byte) error {
_, err := emitter.output_writer.Write(buffer)
return err
}
// Set a string output.
func yaml_emitter_set_output_string(emitter *yaml_emitter_t, output_buffer *[]byte) {
if emitter.write_handler != nil {
panic("must set the output target only once")
}
emitter.write_handler = yaml_string_write_handler
emitter.output_buffer = output_buffer
}
// Set a file output.
func yaml_emitter_set_output_writer(emitter *yaml_emitter_t, w io.Writer) {
if emitter.write_handler != nil {
panic("must set the output target only once")
}
emitter.write_handler = yaml_writer_write_handler
emitter.output_writer = w
}
// Set the output encoding.
func yaml_emitter_set_encoding(emitter *yaml_emitter_t, encoding yaml_encoding_t) {
if emitter.encoding != yaml_ANY_ENCODING {
panic("must set the output encoding only once")
}
emitter.encoding = encoding
}
// Set the canonical output style.
func yaml_emitter_set_canonical(emitter *yaml_emitter_t, canonical bool) {
emitter.canonical = canonical
}
//// Set the indentation increment.
func yaml_emitter_set_indent(emitter *yaml_emitter_t, indent int) {
if indent < 2 || indent > 9 {
indent = 2
}
emitter.best_indent = indent
}
// Set the preferred line width.
func yaml_emitter_set_width(emitter *yaml_emitter_t, width int) {
if width < 0 {
width = -1
}
emitter.best_width = width
}
// Set if unescaped non-ASCII characters are allowed.
func yaml_emitter_set_unicode(emitter *yaml_emitter_t, unicode bool) {
emitter.unicode = unicode
}
// Set the preferred line break character.
func yaml_emitter_set_break(emitter *yaml_emitter_t, line_break yaml_break_t) {
emitter.line_break = line_break
}
///*
// * Destroy a token object.
// */
//
//YAML_DECLARE(void)
//yaml_token_delete(yaml_token_t *token)
//{
// assert(token); // Non-NULL token object expected.
//
// switch (token.type)
// {
// case YAML_TAG_DIRECTIVE_TOKEN:
// yaml_free(token.data.tag_directive.handle);
// yaml_free(token.data.tag_directive.prefix);
// break;
//
// case YAML_ALIAS_TOKEN:
// yaml_free(token.data.alias.value);
// break;
//
// case YAML_ANCHOR_TOKEN:
// yaml_free(token.data.anchor.value);
// break;
//
// case YAML_TAG_TOKEN:
// yaml_free(token.data.tag.handle);
// yaml_free(token.data.tag.suffix);
// break;
//
// case YAML_SCALAR_TOKEN:
// yaml_free(token.data.scalar.value);
// break;
//
// default:
// break;
// }
//
// memset(token, 0, sizeof(yaml_token_t));
//}
//
///*
// * Check if a string is a valid UTF-8 sequence.
// *
// * Check 'reader.c' for more details on UTF-8 encoding.
// */
//
//static int
//yaml_check_utf8(yaml_char_t *start, size_t length)
//{
// yaml_char_t *end = start+length;
// yaml_char_t *pointer = start;
//
// while (pointer < end) {
// unsigned char octet;
// unsigned int width;
// unsigned int value;
// size_t k;
//
// octet = pointer[0];
// width = (octet & 0x80) == 0x00 ? 1 :
// (octet & 0xE0) == 0xC0 ? 2 :
// (octet & 0xF0) == 0xE0 ? 3 :
// (octet & 0xF8) == 0xF0 ? 4 : 0;
// value = (octet & 0x80) == 0x00 ? octet & 0x7F :
// (octet & 0xE0) == 0xC0 ? octet & 0x1F :
// (octet & 0xF0) == 0xE0 ? octet & 0x0F :
// (octet & 0xF8) == 0xF0 ? octet & 0x07 : 0;
// if (!width) return 0;
// if (pointer+width > end) return 0;
// for (k = 1; k < width; k ++) {
// octet = pointer[k];
// if ((octet & 0xC0) != 0x80) return 0;
// value = (value << 6) + (octet & 0x3F);
// }
// if (!((width == 1) ||
// (width == 2 && value >= 0x80) ||
// (width == 3 && value >= 0x800) ||
// (width == 4 && value >= 0x10000))) return 0;
//
// pointer += width;
// }
//
// return 1;
//}
//
// Create STREAM-START.
func yaml_stream_start_event_initialize(event *yaml_event_t, encoding yaml_encoding_t) {
*event = yaml_event_t{
typ: yaml_STREAM_START_EVENT,
encoding: encoding,
}
}
// Create STREAM-END.
func yaml_stream_end_event_initialize(event *yaml_event_t) {
*event = yaml_event_t{
typ: yaml_STREAM_END_EVENT,
}
}
// Create DOCUMENT-START.
func yaml_document_start_event_initialize(
event *yaml_event_t,
version_directive *yaml_version_directive_t,
tag_directives []yaml_tag_directive_t,
implicit bool,
) {
*event = yaml_event_t{
typ: yaml_DOCUMENT_START_EVENT,
version_directive: version_directive,
tag_directives: tag_directives,
implicit: implicit,
}
}
// Create DOCUMENT-END.
func yaml_document_end_event_initialize(event *yaml_event_t, implicit bool) {
*event = yaml_event_t{
typ: yaml_DOCUMENT_END_EVENT,
implicit: implicit,
}
}
///*
// * Create ALIAS.
// */
//
//YAML_DECLARE(int)
//yaml_alias_event_initialize(event *yaml_event_t, anchor *yaml_char_t)
//{
// mark yaml_mark_t = { 0, 0, 0 }
// anchor_copy *yaml_char_t = NULL
//
// assert(event) // Non-NULL event object is expected.
// assert(anchor) // Non-NULL anchor is expected.
//
// if (!yaml_check_utf8(anchor, strlen((char *)anchor))) return 0
//
// anchor_copy = yaml_strdup(anchor)
// if (!anchor_copy)
// return 0
//
// ALIAS_EVENT_INIT(*event, anchor_copy, mark, mark)
//
// return 1
//}
// Create SCALAR.
func yaml_scalar_event_initialize(event *yaml_event_t, anchor, tag, value []byte, plain_implicit, quoted_implicit bool, style yaml_scalar_style_t) bool {
*event = yaml_event_t{
typ: yaml_SCALAR_EVENT,
anchor: anchor,
tag: tag,
value: value,
implicit: plain_implicit,
quoted_implicit: quoted_implicit,
style: yaml_style_t(style),
}
return true
}
// Create SEQUENCE-START.
func yaml_sequence_start_event_initialize(event *yaml_event_t, anchor, tag []byte, implicit bool, style yaml_sequence_style_t) bool {
*event = yaml_event_t{
typ: yaml_SEQUENCE_START_EVENT,
anchor: anchor,
tag: tag,
implicit: implicit,
style: yaml_style_t(style),
}
return true
}
// Create SEQUENCE-END.
func yaml_sequence_end_event_initialize(event *yaml_event_t) bool {
*event = yaml_event_t{
typ: yaml_SEQUENCE_END_EVENT,
}
return true
}
// Create MAPPING-START.
func yaml_mapping_start_event_initialize(event *yaml_event_t, anchor, tag []byte, implicit bool, style yaml_mapping_style_t) {
*event = yaml_event_t{
typ: yaml_MAPPING_START_EVENT,
anchor: anchor,
tag: tag,
implicit: implicit,
style: yaml_style_t(style),
}
}
// Create MAPPING-END.
func yaml_mapping_end_event_initialize(event *yaml_event_t) {
*event = yaml_event_t{
typ: yaml_MAPPING_END_EVENT,
}
}
// Destroy an event object.
func yaml_event_delete(event *yaml_event_t) {
*event = yaml_event_t{}
}
///*
// * Create a document object.
// */
//
//YAML_DECLARE(int)
//yaml_document_initialize(document *yaml_document_t,
// version_directive *yaml_version_directive_t,
// tag_directives_start *yaml_tag_directive_t,
// tag_directives_end *yaml_tag_directive_t,
// start_implicit int, end_implicit int)
//{
// struct {
// error yaml_error_type_t
// } context
// struct {
// start *yaml_node_t
// end *yaml_node_t
// top *yaml_node_t
// } nodes = { NULL, NULL, NULL }
// version_directive_copy *yaml_version_directive_t = NULL
// struct {
// start *yaml_tag_directive_t
// end *yaml_tag_directive_t
// top *yaml_tag_directive_t
// } tag_directives_copy = { NULL, NULL, NULL }
// value yaml_tag_directive_t = { NULL, NULL }
// mark yaml_mark_t = { 0, 0, 0 }
//
// assert(document) // Non-NULL document object is expected.
// assert((tag_directives_start && tag_directives_end) ||
// (tag_directives_start == tag_directives_end))
// // Valid tag directives are expected.
//
// if (!STACK_INIT(&context, nodes, INITIAL_STACK_SIZE)) goto error
//
// if (version_directive) {
// version_directive_copy = yaml_malloc(sizeof(yaml_version_directive_t))
// if (!version_directive_copy) goto error
// version_directive_copy.major = version_directive.major
// version_directive_copy.minor = version_directive.minor
// }
//
// if (tag_directives_start != tag_directives_end) {
// tag_directive *yaml_tag_directive_t
// if (!STACK_INIT(&context, tag_directives_copy, INITIAL_STACK_SIZE))
// goto error
// for (tag_directive = tag_directives_start
// tag_directive != tag_directives_end; tag_directive ++) {
// assert(tag_directive.handle)
// assert(tag_directive.prefix)
// if (!yaml_check_utf8(tag_directive.handle,
// strlen((char *)tag_directive.handle)))
// goto error
// if (!yaml_check_utf8(tag_directive.prefix,
// strlen((char *)tag_directive.prefix)))
// goto error
// value.handle = yaml_strdup(tag_directive.handle)
// value.prefix = yaml_strdup(tag_directive.prefix)
// if (!value.handle || !value.prefix) goto error
// if (!PUSH(&context, tag_directives_copy, value))
// goto error
// value.handle = NULL
// value.prefix = NULL
// }
// }
//
// DOCUMENT_INIT(*document, nodes.start, nodes.end, version_directive_copy,
// tag_directives_copy.start, tag_directives_copy.top,
// start_implicit, end_implicit, mark, mark)
//
// return 1
//
//error:
// STACK_DEL(&context, nodes)
// yaml_free(version_directive_copy)
// while (!STACK_EMPTY(&context, tag_directives_copy)) {
// value yaml_tag_directive_t = POP(&context, tag_directives_copy)
// yaml_free(value.handle)
// yaml_free(value.prefix)
// }
// STACK_DEL(&context, tag_directives_copy)
// yaml_free(value.handle)
// yaml_free(value.prefix)
//
// return 0
//}
//
///*
// * Destroy a document object.
// */
//
//YAML_DECLARE(void)
//yaml_document_delete(document *yaml_document_t)
//{
// struct {
// error yaml_error_type_t
// } context
// tag_directive *yaml_tag_directive_t
//
// context.error = YAML_NO_ERROR // Eliminate a compiler warning.
//
// assert(document) // Non-NULL document object is expected.
//
// while (!STACK_EMPTY(&context, document.nodes)) {
// node yaml_node_t = POP(&context, document.nodes)
// yaml_free(node.tag)
// switch (node.type) {
// case YAML_SCALAR_NODE:
// yaml_free(node.data.scalar.value)
// break
// case YAML_SEQUENCE_NODE:
// STACK_DEL(&context, node.data.sequence.items)
// break
// case YAML_MAPPING_NODE:
// STACK_DEL(&context, node.data.mapping.pairs)
// break
// default:
// assert(0) // Should not happen.
// }
// }
// STACK_DEL(&context, document.nodes)
//
// yaml_free(document.version_directive)
// for (tag_directive = document.tag_directives.start
// tag_directive != document.tag_directives.end
// tag_directive++) {
// yaml_free(tag_directive.handle)
// yaml_free(tag_directive.prefix)
// }
// yaml_free(document.tag_directives.start)
//
// memset(document, 0, sizeof(yaml_document_t))
//}
//
///**
// * Get a document node.
// */
//
//YAML_DECLARE(yaml_node_t *)
//yaml_document_get_node(document *yaml_document_t, index int)
//{
// assert(document) // Non-NULL document object is expected.
//
// if (index > 0 && document.nodes.start + index <= document.nodes.top) {
// return document.nodes.start + index - 1
// }
// return NULL
//}
//
///**
// * Get the root object.
// */
//
//YAML_DECLARE(yaml_node_t *)
//yaml_document_get_root_node(document *yaml_document_t)
//{
// assert(document) // Non-NULL document object is expected.
//
// if (document.nodes.top != document.nodes.start) {
// return document.nodes.start
// }
// return NULL
//}
//
///*
// * Add a scalar node to a document.
// */
//
//YAML_DECLARE(int)
//yaml_document_add_scalar(document *yaml_document_t,
// tag *yaml_char_t, value *yaml_char_t, length int,
// style yaml_scalar_style_t)
//{
// struct {
// error yaml_error_type_t
// } context
// mark yaml_mark_t = { 0, 0, 0 }
// tag_copy *yaml_char_t = NULL
// value_copy *yaml_char_t = NULL
// node yaml_node_t
//
// assert(document) // Non-NULL document object is expected.
// assert(value) // Non-NULL value is expected.
//
// if (!tag) {
// tag = (yaml_char_t *)YAML_DEFAULT_SCALAR_TAG
// }
//
// if (!yaml_check_utf8(tag, strlen((char *)tag))) goto error
// tag_copy = yaml_strdup(tag)
// if (!tag_copy) goto error
//
// if (length < 0) {
// length = strlen((char *)value)
// }
//
// if (!yaml_check_utf8(value, length)) goto error
// value_copy = yaml_malloc(length+1)
// if (!value_copy) goto error
// memcpy(value_copy, value, length)
// value_copy[length] = '\0'
//
// SCALAR_NODE_INIT(node, tag_copy, value_copy, length, style, mark, mark)
// if (!PUSH(&context, document.nodes, node)) goto error
//
// return document.nodes.top - document.nodes.start
//
//error:
// yaml_free(tag_copy)
// yaml_free(value_copy)
//
// return 0
//}
//
///*
// * Add a sequence node to a document.
// */
//
//YAML_DECLARE(int)
//yaml_document_add_sequence(document *yaml_document_t,
// tag *yaml_char_t, style yaml_sequence_style_t)
//{
// struct {
// error yaml_error_type_t
// } context
// mark yaml_mark_t = { 0, 0, 0 }
// tag_copy *yaml_char_t = NULL
// struct {
// start *yaml_node_item_t
// end *yaml_node_item_t
// top *yaml_node_item_t
// } items = { NULL, NULL, NULL }
// node yaml_node_t
//
// assert(document) // Non-NULL document object is expected.
//
// if (!tag) {
// tag = (yaml_char_t *)YAML_DEFAULT_SEQUENCE_TAG
// }
//
// if (!yaml_check_utf8(tag, strlen((char *)tag))) goto error
// tag_copy = yaml_strdup(tag)
// if (!tag_copy) goto error
//
// if (!STACK_INIT(&context, items, INITIAL_STACK_SIZE)) goto error
//
// SEQUENCE_NODE_INIT(node, tag_copy, items.start, items.end,
// style, mark, mark)
// if (!PUSH(&context, document.nodes, node)) goto error
//
// return document.nodes.top - document.nodes.start
//
//error:
// STACK_DEL(&context, items)
// yaml_free(tag_copy)
//
// return 0
//}
//
///*
// * Add a mapping node to a document.
// */
//
//YAML_DECLARE(int)
//yaml_document_add_mapping(document *yaml_document_t,
// tag *yaml_char_t, style yaml_mapping_style_t)
//{
// struct {
// error yaml_error_type_t
// } context
// mark yaml_mark_t = { 0, 0, 0 }
// tag_copy *yaml_char_t = NULL
// struct {
// start *yaml_node_pair_t
// end *yaml_node_pair_t
// top *yaml_node_pair_t
// } pairs = { NULL, NULL, NULL }
// node yaml_node_t
//
// assert(document) // Non-NULL document object is expected.
//
// if (!tag) {
// tag = (yaml_char_t *)YAML_DEFAULT_MAPPING_TAG
// }
//
// if (!yaml_check_utf8(tag, strlen((char *)tag))) goto error
// tag_copy = yaml_strdup(tag)
// if (!tag_copy) goto error
//
// if (!STACK_INIT(&context, pairs, INITIAL_STACK_SIZE)) goto error
//
// MAPPING_NODE_INIT(node, tag_copy, pairs.start, pairs.end,
// style, mark, mark)
// if (!PUSH(&context, document.nodes, node)) goto error
//
// return document.nodes.top - document.nodes.start
//
//error:
// STACK_DEL(&context, pairs)
// yaml_free(tag_copy)
//
// return 0
//}
//
///*
// * Append an item to a sequence node.
// */
//
//YAML_DECLARE(int)
//yaml_document_append_sequence_item(document *yaml_document_t,
// sequence int, item int)
//{
// struct {
// error yaml_error_type_t
// } context
//
// assert(document) // Non-NULL document is required.
// assert(sequence > 0
// && document.nodes.start + sequence <= document.nodes.top)
// // Valid sequence id is required.
// assert(document.nodes.start[sequence-1].type == YAML_SEQUENCE_NODE)
// // A sequence node is required.
// assert(item > 0 && document.nodes.start + item <= document.nodes.top)
// // Valid item id is required.
//
// if (!PUSH(&context,
// document.nodes.start[sequence-1].data.sequence.items, item))
// return 0
//
// return 1
//}
//
///*
// * Append a pair of a key and a value to a mapping node.
// */
//
//YAML_DECLARE(int)
//yaml_document_append_mapping_pair(document *yaml_document_t,
// mapping int, key int, value int)
//{
// struct {
// error yaml_error_type_t
// } context
//
// pair yaml_node_pair_t
//
// assert(document) // Non-NULL document is required.
// assert(mapping > 0
// && document.nodes.start + mapping <= document.nodes.top)
// // Valid mapping id is required.
// assert(document.nodes.start[mapping-1].type == YAML_MAPPING_NODE)
// // A mapping node is required.
// assert(key > 0 && document.nodes.start + key <= document.nodes.top)
// // Valid key id is required.
// assert(value > 0 && document.nodes.start + value <= document.nodes.top)
// // Valid value id is required.
//
// pair.key = key
// pair.value = value
//
// if (!PUSH(&context,
// document.nodes.start[mapping-1].data.mapping.pairs, pair))
// return 0
//
// return 1
//}
//
//

775
vendor/gopkg.in/yaml.v2/decode.go generated vendored
View File

@ -1,775 +0,0 @@
package yaml
import (
"encoding"
"encoding/base64"
"fmt"
"io"
"math"
"reflect"
"strconv"
"time"
)
const (
documentNode = 1 << iota
mappingNode
sequenceNode
scalarNode
aliasNode
)
type node struct {
kind int
line, column int
tag string
// For an alias node, alias holds the resolved alias.
alias *node
value string
implicit bool
children []*node
anchors map[string]*node
}
// ----------------------------------------------------------------------------
// Parser, produces a node tree out of a libyaml event stream.
type parser struct {
parser yaml_parser_t
event yaml_event_t
doc *node
doneInit bool
}
func newParser(b []byte) *parser {
p := parser{}
if !yaml_parser_initialize(&p.parser) {
panic("failed to initialize YAML emitter")
}
if len(b) == 0 {
b = []byte{'\n'}
}
yaml_parser_set_input_string(&p.parser, b)
return &p
}
func newParserFromReader(r io.Reader) *parser {
p := parser{}
if !yaml_parser_initialize(&p.parser) {
panic("failed to initialize YAML emitter")
}
yaml_parser_set_input_reader(&p.parser, r)
return &p
}
func (p *parser) init() {
if p.doneInit {
return
}
p.expect(yaml_STREAM_START_EVENT)
p.doneInit = true
}
func (p *parser) destroy() {
if p.event.typ != yaml_NO_EVENT {
yaml_event_delete(&p.event)
}
yaml_parser_delete(&p.parser)
}
// expect consumes an event from the event stream and
// checks that it's of the expected type.
func (p *parser) expect(e yaml_event_type_t) {
if p.event.typ == yaml_NO_EVENT {
if !yaml_parser_parse(&p.parser, &p.event) {
p.fail()
}
}
if p.event.typ == yaml_STREAM_END_EVENT {
failf("attempted to go past the end of stream; corrupted value?")
}
if p.event.typ != e {
p.parser.problem = fmt.Sprintf("expected %s event but got %s", e, p.event.typ)
p.fail()
}
yaml_event_delete(&p.event)
p.event.typ = yaml_NO_EVENT
}
// peek peeks at the next event in the event stream,
// puts the results into p.event and returns the event type.
func (p *parser) peek() yaml_event_type_t {
if p.event.typ != yaml_NO_EVENT {
return p.event.typ
}
if !yaml_parser_parse(&p.parser, &p.event) {
p.fail()
}
return p.event.typ
}
func (p *parser) fail() {
var where string
var line int
if p.parser.problem_mark.line != 0 {
line = p.parser.problem_mark.line
// Scanner errors don't iterate line before returning error
if p.parser.error == yaml_SCANNER_ERROR {
line++
}
} else if p.parser.context_mark.line != 0 {
line = p.parser.context_mark.line
}
if line != 0 {
where = "line " + strconv.Itoa(line) + ": "
}
var msg string
if len(p.parser.problem) > 0 {
msg = p.parser.problem
} else {
msg = "unknown problem parsing YAML content"
}
failf("%s%s", where, msg)
}
func (p *parser) anchor(n *node, anchor []byte) {
if anchor != nil {
p.doc.anchors[string(anchor)] = n
}
}
func (p *parser) parse() *node {
p.init()
switch p.peek() {
case yaml_SCALAR_EVENT:
return p.scalar()
case yaml_ALIAS_EVENT:
return p.alias()
case yaml_MAPPING_START_EVENT:
return p.mapping()
case yaml_SEQUENCE_START_EVENT:
return p.sequence()
case yaml_DOCUMENT_START_EVENT:
return p.document()
case yaml_STREAM_END_EVENT:
// Happens when attempting to decode an empty buffer.
return nil
default:
panic("attempted to parse unknown event: " + p.event.typ.String())
}
}
func (p *parser) node(kind int) *node {
return &node{
kind: kind,
line: p.event.start_mark.line,
column: p.event.start_mark.column,
}
}
func (p *parser) document() *node {
n := p.node(documentNode)
n.anchors = make(map[string]*node)
p.doc = n
p.expect(yaml_DOCUMENT_START_EVENT)
n.children = append(n.children, p.parse())
p.expect(yaml_DOCUMENT_END_EVENT)
return n
}
func (p *parser) alias() *node {
n := p.node(aliasNode)
n.value = string(p.event.anchor)
n.alias = p.doc.anchors[n.value]
if n.alias == nil {
failf("unknown anchor '%s' referenced", n.value)
}
p.expect(yaml_ALIAS_EVENT)
return n
}
func (p *parser) scalar() *node {
n := p.node(scalarNode)
n.value = string(p.event.value)
n.tag = string(p.event.tag)
n.implicit = p.event.implicit
p.anchor(n, p.event.anchor)
p.expect(yaml_SCALAR_EVENT)
return n
}
func (p *parser) sequence() *node {
n := p.node(sequenceNode)
p.anchor(n, p.event.anchor)
p.expect(yaml_SEQUENCE_START_EVENT)
for p.peek() != yaml_SEQUENCE_END_EVENT {
n.children = append(n.children, p.parse())
}
p.expect(yaml_SEQUENCE_END_EVENT)
return n
}
func (p *parser) mapping() *node {
n := p.node(mappingNode)
p.anchor(n, p.event.anchor)
p.expect(yaml_MAPPING_START_EVENT)
for p.peek() != yaml_MAPPING_END_EVENT {
n.children = append(n.children, p.parse(), p.parse())
}
p.expect(yaml_MAPPING_END_EVENT)
return n
}
// ----------------------------------------------------------------------------
// Decoder, unmarshals a node into a provided value.
type decoder struct {
doc *node
aliases map[*node]bool
mapType reflect.Type
terrors []string
strict bool
}
var (
mapItemType = reflect.TypeOf(MapItem{})
durationType = reflect.TypeOf(time.Duration(0))
defaultMapType = reflect.TypeOf(map[interface{}]interface{}{})
ifaceType = defaultMapType.Elem()
timeType = reflect.TypeOf(time.Time{})
ptrTimeType = reflect.TypeOf(&time.Time{})
)
func newDecoder(strict bool) *decoder {
d := &decoder{mapType: defaultMapType, strict: strict}
d.aliases = make(map[*node]bool)
return d
}
func (d *decoder) terror(n *node, tag string, out reflect.Value) {
if n.tag != "" {
tag = n.tag
}
value := n.value
if tag != yaml_SEQ_TAG && tag != yaml_MAP_TAG {
if len(value) > 10 {
value = " `" + value[:7] + "...`"
} else {
value = " `" + value + "`"
}
}
d.terrors = append(d.terrors, fmt.Sprintf("line %d: cannot unmarshal %s%s into %s", n.line+1, shortTag(tag), value, out.Type()))
}
func (d *decoder) callUnmarshaler(n *node, u Unmarshaler) (good bool) {
terrlen := len(d.terrors)
err := u.UnmarshalYAML(func(v interface{}) (err error) {
defer handleErr(&err)
d.unmarshal(n, reflect.ValueOf(v))
if len(d.terrors) > terrlen {
issues := d.terrors[terrlen:]
d.terrors = d.terrors[:terrlen]
return &TypeError{issues}
}
return nil
})
if e, ok := err.(*TypeError); ok {
d.terrors = append(d.terrors, e.Errors...)
return false
}
if err != nil {
fail(err)
}
return true
}
// d.prepare initializes and dereferences pointers and calls UnmarshalYAML
// if a value is found to implement it.
// It returns the initialized and dereferenced out value, whether
// unmarshalling was already done by UnmarshalYAML, and if so whether
// its types unmarshalled appropriately.
//
// If n holds a null value, prepare returns before doing anything.
func (d *decoder) prepare(n *node, out reflect.Value) (newout reflect.Value, unmarshaled, good bool) {
if n.tag == yaml_NULL_TAG || n.kind == scalarNode && n.tag == "" && (n.value == "null" || n.value == "~" || n.value == "" && n.implicit) {
return out, false, false
}
again := true
for again {
again = false
if out.Kind() == reflect.Ptr {
if out.IsNil() {
out.Set(reflect.New(out.Type().Elem()))
}
out = out.Elem()
again = true
}
if out.CanAddr() {
if u, ok := out.Addr().Interface().(Unmarshaler); ok {
good = d.callUnmarshaler(n, u)
return out, true, good
}
}
}
return out, false, false
}
func (d *decoder) unmarshal(n *node, out reflect.Value) (good bool) {
switch n.kind {
case documentNode:
return d.document(n, out)
case aliasNode:
return d.alias(n, out)
}
out, unmarshaled, good := d.prepare(n, out)
if unmarshaled {
return good
}
switch n.kind {
case scalarNode:
good = d.scalar(n, out)
case mappingNode:
good = d.mapping(n, out)
case sequenceNode:
good = d.sequence(n, out)
default:
panic("internal error: unknown node kind: " + strconv.Itoa(n.kind))
}
return good
}
func (d *decoder) document(n *node, out reflect.Value) (good bool) {
if len(n.children) == 1 {
d.doc = n
d.unmarshal(n.children[0], out)
return true
}
return false
}
func (d *decoder) alias(n *node, out reflect.Value) (good bool) {
if d.aliases[n] {
// TODO this could actually be allowed in some circumstances.
failf("anchor '%s' value contains itself", n.value)
}
d.aliases[n] = true
good = d.unmarshal(n.alias, out)
delete(d.aliases, n)
return good
}
var zeroValue reflect.Value
func resetMap(out reflect.Value) {
for _, k := range out.MapKeys() {
out.SetMapIndex(k, zeroValue)
}
}
func (d *decoder) scalar(n *node, out reflect.Value) bool {
var tag string
var resolved interface{}
if n.tag == "" && !n.implicit {
tag = yaml_STR_TAG
resolved = n.value
} else {
tag, resolved = resolve(n.tag, n.value)
if tag == yaml_BINARY_TAG {
data, err := base64.StdEncoding.DecodeString(resolved.(string))
if err != nil {
failf("!!binary value contains invalid base64 data")
}
resolved = string(data)
}
}
if resolved == nil {
if out.Kind() == reflect.Map && !out.CanAddr() {
resetMap(out)
} else {
out.Set(reflect.Zero(out.Type()))
}
return true
}
if resolvedv := reflect.ValueOf(resolved); out.Type() == resolvedv.Type() {
// We've resolved to exactly the type we want, so use that.
out.Set(resolvedv)
return true
}
// Perhaps we can use the value as a TextUnmarshaler to
// set its value.
if out.CanAddr() {
u, ok := out.Addr().Interface().(encoding.TextUnmarshaler)
if ok {
var text []byte
if tag == yaml_BINARY_TAG {
text = []byte(resolved.(string))
} else {
// We let any value be unmarshaled into TextUnmarshaler.
// That might be more lax than we'd like, but the
// TextUnmarshaler itself should bowl out any dubious values.
text = []byte(n.value)
}
err := u.UnmarshalText(text)
if err != nil {
fail(err)
}
return true
}
}
switch out.Kind() {
case reflect.String:
if tag == yaml_BINARY_TAG {
out.SetString(resolved.(string))
return true
}
if resolved != nil {
out.SetString(n.value)
return true
}
case reflect.Interface:
if resolved == nil {
out.Set(reflect.Zero(out.Type()))
} else if tag == yaml_TIMESTAMP_TAG {
// It looks like a timestamp but for backward compatibility
// reasons we set it as a string, so that code that unmarshals
// timestamp-like values into interface{} will continue to
// see a string and not a time.Time.
// TODO(v3) Drop this.
out.Set(reflect.ValueOf(n.value))
} else {
out.Set(reflect.ValueOf(resolved))
}
return true
case reflect.Int, reflect.Int8, reflect.Int16, reflect.Int32, reflect.Int64:
switch resolved := resolved.(type) {
case int:
if !out.OverflowInt(int64(resolved)) {
out.SetInt(int64(resolved))
return true
}
case int64:
if !out.OverflowInt(resolved) {
out.SetInt(resolved)
return true
}
case uint64:
if resolved <= math.MaxInt64 && !out.OverflowInt(int64(resolved)) {
out.SetInt(int64(resolved))
return true
}
case float64:
if resolved <= math.MaxInt64 && !out.OverflowInt(int64(resolved)) {
out.SetInt(int64(resolved))
return true
}
case string:
if out.Type() == durationType {
d, err := time.ParseDuration(resolved)
if err == nil {
out.SetInt(int64(d))
return true
}
}
}
case reflect.Uint, reflect.Uint8, reflect.Uint16, reflect.Uint32, reflect.Uint64, reflect.Uintptr:
switch resolved := resolved.(type) {
case int:
if resolved >= 0 && !out.OverflowUint(uint64(resolved)) {
out.SetUint(uint64(resolved))
return true
}
case int64:
if resolved >= 0 && !out.OverflowUint(uint64(resolved)) {
out.SetUint(uint64(resolved))
return true
}
case uint64:
if !out.OverflowUint(uint64(resolved)) {
out.SetUint(uint64(resolved))
return true
}
case float64:
if resolved <= math.MaxUint64 && !out.OverflowUint(uint64(resolved)) {
out.SetUint(uint64(resolved))
return true
}
}
case reflect.Bool:
switch resolved := resolved.(type) {
case bool:
out.SetBool(resolved)
return true
}
case reflect.Float32, reflect.Float64:
switch resolved := resolved.(type) {
case int:
out.SetFloat(float64(resolved))
return true
case int64:
out.SetFloat(float64(resolved))
return true
case uint64:
out.SetFloat(float64(resolved))
return true
case float64:
out.SetFloat(resolved)
return true
}
case reflect.Struct:
if resolvedv := reflect.ValueOf(resolved); out.Type() == resolvedv.Type() {
out.Set(resolvedv)
return true
}
case reflect.Ptr:
if out.Type().Elem() == reflect.TypeOf(resolved) {
// TODO DOes this make sense? When is out a Ptr except when decoding a nil value?
elem := reflect.New(out.Type().Elem())
elem.Elem().Set(reflect.ValueOf(resolved))
out.Set(elem)
return true
}
}
d.terror(n, tag, out)
return false
}
func settableValueOf(i interface{}) reflect.Value {
v := reflect.ValueOf(i)
sv := reflect.New(v.Type()).Elem()
sv.Set(v)
return sv
}
func (d *decoder) sequence(n *node, out reflect.Value) (good bool) {
l := len(n.children)
var iface reflect.Value
switch out.Kind() {
case reflect.Slice:
out.Set(reflect.MakeSlice(out.Type(), l, l))
case reflect.Array:
if l != out.Len() {
failf("invalid array: want %d elements but got %d", out.Len(), l)
}
case reflect.Interface:
// No type hints. Will have to use a generic sequence.
iface = out
out = settableValueOf(make([]interface{}, l))
default:
d.terror(n, yaml_SEQ_TAG, out)
return false
}
et := out.Type().Elem()
j := 0
for i := 0; i < l; i++ {
e := reflect.New(et).Elem()
if ok := d.unmarshal(n.children[i], e); ok {
out.Index(j).Set(e)
j++
}
}
if out.Kind() != reflect.Array {
out.Set(out.Slice(0, j))
}
if iface.IsValid() {
iface.Set(out)
}
return true
}
func (d *decoder) mapping(n *node, out reflect.Value) (good bool) {
switch out.Kind() {
case reflect.Struct:
return d.mappingStruct(n, out)
case reflect.Slice:
return d.mappingSlice(n, out)
case reflect.Map:
// okay
case reflect.Interface:
if d.mapType.Kind() == reflect.Map {
iface := out
out = reflect.MakeMap(d.mapType)
iface.Set(out)
} else {
slicev := reflect.New(d.mapType).Elem()
if !d.mappingSlice(n, slicev) {
return false
}
out.Set(slicev)
return true
}
default:
d.terror(n, yaml_MAP_TAG, out)
return false
}
outt := out.Type()
kt := outt.Key()
et := outt.Elem()
mapType := d.mapType
if outt.Key() == ifaceType && outt.Elem() == ifaceType {
d.mapType = outt
}
if out.IsNil() {
out.Set(reflect.MakeMap(outt))
}
l := len(n.children)
for i := 0; i < l; i += 2 {
if isMerge(n.children[i]) {
d.merge(n.children[i+1], out)
continue
}
k := reflect.New(kt).Elem()
if d.unmarshal(n.children[i], k) {
kkind := k.Kind()
if kkind == reflect.Interface {
kkind = k.Elem().Kind()
}
if kkind == reflect.Map || kkind == reflect.Slice {
failf("invalid map key: %#v", k.Interface())
}
e := reflect.New(et).Elem()
if d.unmarshal(n.children[i+1], e) {
d.setMapIndex(n.children[i+1], out, k, e)
}
}
}
d.mapType = mapType
return true
}
func (d *decoder) setMapIndex(n *node, out, k, v reflect.Value) {
if d.strict && out.MapIndex(k) != zeroValue {
d.terrors = append(d.terrors, fmt.Sprintf("line %d: key %#v already set in map", n.line+1, k.Interface()))
return
}
out.SetMapIndex(k, v)
}
func (d *decoder) mappingSlice(n *node, out reflect.Value) (good bool) {
outt := out.Type()
if outt.Elem() != mapItemType {
d.terror(n, yaml_MAP_TAG, out)
return false
}
mapType := d.mapType
d.mapType = outt
var slice []MapItem
var l = len(n.children)
for i := 0; i < l; i += 2 {
if isMerge(n.children[i]) {
d.merge(n.children[i+1], out)
continue
}
item := MapItem{}
k := reflect.ValueOf(&item.Key).Elem()
if d.unmarshal(n.children[i], k) {
v := reflect.ValueOf(&item.Value).Elem()
if d.unmarshal(n.children[i+1], v) {
slice = append(slice, item)
}
}
}
out.Set(reflect.ValueOf(slice))
d.mapType = mapType
return true
}
func (d *decoder) mappingStruct(n *node, out reflect.Value) (good bool) {
sinfo, err := getStructInfo(out.Type())
if err != nil {
panic(err)
}
name := settableValueOf("")
l := len(n.children)
var inlineMap reflect.Value
var elemType reflect.Type
if sinfo.InlineMap != -1 {
inlineMap = out.Field(sinfo.InlineMap)
inlineMap.Set(reflect.New(inlineMap.Type()).Elem())
elemType = inlineMap.Type().Elem()
}
var doneFields []bool
if d.strict {
doneFields = make([]bool, len(sinfo.FieldsList))
}
for i := 0; i < l; i += 2 {
ni := n.children[i]
if isMerge(ni) {
d.merge(n.children[i+1], out)
continue
}
if !d.unmarshal(ni, name) {
continue
}
if info, ok := sinfo.FieldsMap[name.String()]; ok {
if d.strict {
if doneFields[info.Id] {
d.terrors = append(d.terrors, fmt.Sprintf("line %d: field %s already set in type %s", ni.line+1, name.String(), out.Type()))
continue
}
doneFields[info.Id] = true
}
var field reflect.Value
if info.Inline == nil {
field = out.Field(info.Num)
} else {
field = out.FieldByIndex(info.Inline)
}
d.unmarshal(n.children[i+1], field)
} else if sinfo.InlineMap != -1 {
if inlineMap.IsNil() {
inlineMap.Set(reflect.MakeMap(inlineMap.Type()))
}
value := reflect.New(elemType).Elem()
d.unmarshal(n.children[i+1], value)
d.setMapIndex(n.children[i+1], inlineMap, name, value)
} else if d.strict {
d.terrors = append(d.terrors, fmt.Sprintf("line %d: field %s not found in type %s", ni.line+1, name.String(), out.Type()))
}
}
return true
}
func failWantMap() {
failf("map merge requires map or sequence of maps as the value")
}
func (d *decoder) merge(n *node, out reflect.Value) {
switch n.kind {
case mappingNode:
d.unmarshal(n, out)
case aliasNode:
an, ok := d.doc.anchors[n.value]
if ok && an.kind != mappingNode {
failWantMap()
}
d.unmarshal(n, out)
case sequenceNode:
// Step backwards as earlier nodes take precedence.
for i := len(n.children) - 1; i >= 0; i-- {
ni := n.children[i]
if ni.kind == aliasNode {
an, ok := d.doc.anchors[ni.value]
if ok && an.kind != mappingNode {
failWantMap()
}
} else if ni.kind != mappingNode {
failWantMap()
}
d.unmarshal(ni, out)
}
default:
failWantMap()
}
}
func isMerge(n *node) bool {
return n.kind == scalarNode && n.value == "<<" && (n.implicit == true || n.tag == yaml_MERGE_TAG)
}

1685
vendor/gopkg.in/yaml.v2/emitterc.go generated vendored

File diff suppressed because it is too large Load Diff

390
vendor/gopkg.in/yaml.v2/encode.go generated vendored
View File

@ -1,390 +0,0 @@
package yaml
import (
"encoding"
"fmt"
"io"
"reflect"
"regexp"
"sort"
"strconv"
"strings"
"time"
"unicode/utf8"
)
// jsonNumber is the interface of the encoding/json.Number datatype.
// Repeating the interface here avoids a dependency on encoding/json, and also
// supports other libraries like jsoniter, which use a similar datatype with
// the same interface. Detecting this interface is useful when dealing with
// structures containing json.Number, which is a string under the hood. The
// encoder should prefer the use of Int64(), Float64() and string(), in that
// order, when encoding this type.
type jsonNumber interface {
Float64() (float64, error)
Int64() (int64, error)
String() string
}
type encoder struct {
emitter yaml_emitter_t
event yaml_event_t
out []byte
flow bool
// doneInit holds whether the initial stream_start_event has been
// emitted.
doneInit bool
}
func newEncoder() *encoder {
e := &encoder{}
yaml_emitter_initialize(&e.emitter)
yaml_emitter_set_output_string(&e.emitter, &e.out)
yaml_emitter_set_unicode(&e.emitter, true)
return e
}
func newEncoderWithWriter(w io.Writer) *encoder {
e := &encoder{}
yaml_emitter_initialize(&e.emitter)
yaml_emitter_set_output_writer(&e.emitter, w)
yaml_emitter_set_unicode(&e.emitter, true)
return e
}
func (e *encoder) init() {
if e.doneInit {
return
}
yaml_stream_start_event_initialize(&e.event, yaml_UTF8_ENCODING)
e.emit()
e.doneInit = true
}
func (e *encoder) finish() {
e.emitter.open_ended = false
yaml_stream_end_event_initialize(&e.event)
e.emit()
}
func (e *encoder) destroy() {
yaml_emitter_delete(&e.emitter)
}
func (e *encoder) emit() {
// This will internally delete the e.event value.
e.must(yaml_emitter_emit(&e.emitter, &e.event))
}
func (e *encoder) must(ok bool) {
if !ok {
msg := e.emitter.problem
if msg == "" {
msg = "unknown problem generating YAML content"
}
failf("%s", msg)
}
}
func (e *encoder) marshalDoc(tag string, in reflect.Value) {
e.init()
yaml_document_start_event_initialize(&e.event, nil, nil, true)
e.emit()
e.marshal(tag, in)
yaml_document_end_event_initialize(&e.event, true)
e.emit()
}
func (e *encoder) marshal(tag string, in reflect.Value) {
if !in.IsValid() || in.Kind() == reflect.Ptr && in.IsNil() {
e.nilv()
return
}
iface := in.Interface()
switch m := iface.(type) {
case jsonNumber:
integer, err := m.Int64()
if err == nil {
// In this case the json.Number is a valid int64
in = reflect.ValueOf(integer)
break
}
float, err := m.Float64()
if err == nil {
// In this case the json.Number is a valid float64
in = reflect.ValueOf(float)
break
}
// fallback case - no number could be obtained
in = reflect.ValueOf(m.String())
case time.Time, *time.Time:
// Although time.Time implements TextMarshaler,
// we don't want to treat it as a string for YAML
// purposes because YAML has special support for
// timestamps.
case Marshaler:
v, err := m.MarshalYAML()
if err != nil {
fail(err)
}
if v == nil {
e.nilv()
return
}
in = reflect.ValueOf(v)
case encoding.TextMarshaler:
text, err := m.MarshalText()
if err != nil {
fail(err)
}
in = reflect.ValueOf(string(text))
case nil:
e.nilv()
return
}
switch in.Kind() {
case reflect.Interface:
e.marshal(tag, in.Elem())
case reflect.Map:
e.mapv(tag, in)
case reflect.Ptr:
if in.Type() == ptrTimeType {
e.timev(tag, in.Elem())
} else {
e.marshal(tag, in.Elem())
}
case reflect.Struct:
if in.Type() == timeType {
e.timev(tag, in)
} else {
e.structv(tag, in)
}
case reflect.Slice, reflect.Array:
if in.Type().Elem() == mapItemType {
e.itemsv(tag, in)
} else {
e.slicev(tag, in)
}
case reflect.String:
e.stringv(tag, in)
case reflect.Int, reflect.Int8, reflect.Int16, reflect.Int32, reflect.Int64:
if in.Type() == durationType {
e.stringv(tag, reflect.ValueOf(iface.(time.Duration).String()))
} else {
e.intv(tag, in)
}
case reflect.Uint, reflect.Uint8, reflect.Uint16, reflect.Uint32, reflect.Uint64, reflect.Uintptr:
e.uintv(tag, in)
case reflect.Float32, reflect.Float64:
e.floatv(tag, in)
case reflect.Bool:
e.boolv(tag, in)
default:
panic("cannot marshal type: " + in.Type().String())
}
}
func (e *encoder) mapv(tag string, in reflect.Value) {
e.mappingv(tag, func() {
keys := keyList(in.MapKeys())
sort.Sort(keys)
for _, k := range keys {
e.marshal("", k)
e.marshal("", in.MapIndex(k))
}
})
}
func (e *encoder) itemsv(tag string, in reflect.Value) {
e.mappingv(tag, func() {
slice := in.Convert(reflect.TypeOf([]MapItem{})).Interface().([]MapItem)
for _, item := range slice {
e.marshal("", reflect.ValueOf(item.Key))
e.marshal("", reflect.ValueOf(item.Value))
}
})
}
func (e *encoder) structv(tag string, in reflect.Value) {
sinfo, err := getStructInfo(in.Type())
if err != nil {
panic(err)
}
e.mappingv(tag, func() {
for _, info := range sinfo.FieldsList {
var value reflect.Value
if info.Inline == nil {
value = in.Field(info.Num)
} else {
value = in.FieldByIndex(info.Inline)
}
if info.OmitEmpty && isZero(value) {
continue
}
e.marshal("", reflect.ValueOf(info.Key))
e.flow = info.Flow
e.marshal("", value)
}
if sinfo.InlineMap >= 0 {
m := in.Field(sinfo.InlineMap)
if m.Len() > 0 {
e.flow = false
keys := keyList(m.MapKeys())
sort.Sort(keys)
for _, k := range keys {
if _, found := sinfo.FieldsMap[k.String()]; found {
panic(fmt.Sprintf("Can't have key %q in inlined map; conflicts with struct field", k.String()))
}
e.marshal("", k)
e.flow = false
e.marshal("", m.MapIndex(k))
}
}
}
})
}
func (e *encoder) mappingv(tag string, f func()) {
implicit := tag == ""
style := yaml_BLOCK_MAPPING_STYLE
if e.flow {
e.flow = false
style = yaml_FLOW_MAPPING_STYLE
}
yaml_mapping_start_event_initialize(&e.event, nil, []byte(tag), implicit, style)
e.emit()
f()
yaml_mapping_end_event_initialize(&e.event)
e.emit()
}
func (e *encoder) slicev(tag string, in reflect.Value) {
implicit := tag == ""
style := yaml_BLOCK_SEQUENCE_STYLE
if e.flow {
e.flow = false
style = yaml_FLOW_SEQUENCE_STYLE
}
e.must(yaml_sequence_start_event_initialize(&e.event, nil, []byte(tag), implicit, style))
e.emit()
n := in.Len()
for i := 0; i < n; i++ {
e.marshal("", in.Index(i))
}
e.must(yaml_sequence_end_event_initialize(&e.event))
e.emit()
}
// isBase60 returns whether s is in base 60 notation as defined in YAML 1.1.
//
// The base 60 float notation in YAML 1.1 is a terrible idea and is unsupported
// in YAML 1.2 and by this package, but these should be marshalled quoted for
// the time being for compatibility with other parsers.
func isBase60Float(s string) (result bool) {
// Fast path.
if s == "" {
return false
}
c := s[0]
if !(c == '+' || c == '-' || c >= '0' && c <= '9') || strings.IndexByte(s, ':') < 0 {
return false
}
// Do the full match.
return base60float.MatchString(s)
}
// From http://yaml.org/type/float.html, except the regular expression there
// is bogus. In practice parsers do not enforce the "\.[0-9_]*" suffix.
var base60float = regexp.MustCompile(`^[-+]?[0-9][0-9_]*(?::[0-5]?[0-9])+(?:\.[0-9_]*)?$`)
func (e *encoder) stringv(tag string, in reflect.Value) {
var style yaml_scalar_style_t
s := in.String()
canUsePlain := true
switch {
case !utf8.ValidString(s):
if tag == yaml_BINARY_TAG {
failf("explicitly tagged !!binary data must be base64-encoded")
}
if tag != "" {
failf("cannot marshal invalid UTF-8 data as %s", shortTag(tag))
}
// It can't be encoded directly as YAML so use a binary tag
// and encode it as base64.
tag = yaml_BINARY_TAG
s = encodeBase64(s)
case tag == "":
// Check to see if it would resolve to a specific
// tag when encoded unquoted. If it doesn't,
// there's no need to quote it.
rtag, _ := resolve("", s)
canUsePlain = rtag == yaml_STR_TAG && !isBase60Float(s)
}
// Note: it's possible for user code to emit invalid YAML
// if they explicitly specify a tag and a string containing
// text that's incompatible with that tag.
switch {
case strings.Contains(s, "\n"):
style = yaml_LITERAL_SCALAR_STYLE
case canUsePlain:
style = yaml_PLAIN_SCALAR_STYLE
default:
style = yaml_DOUBLE_QUOTED_SCALAR_STYLE
}
e.emitScalar(s, "", tag, style)
}
func (e *encoder) boolv(tag string, in reflect.Value) {
var s string
if in.Bool() {
s = "true"
} else {
s = "false"
}
e.emitScalar(s, "", tag, yaml_PLAIN_SCALAR_STYLE)
}
func (e *encoder) intv(tag string, in reflect.Value) {
s := strconv.FormatInt(in.Int(), 10)
e.emitScalar(s, "", tag, yaml_PLAIN_SCALAR_STYLE)
}
func (e *encoder) uintv(tag string, in reflect.Value) {
s := strconv.FormatUint(in.Uint(), 10)
e.emitScalar(s, "", tag, yaml_PLAIN_SCALAR_STYLE)
}
func (e *encoder) timev(tag string, in reflect.Value) {
t := in.Interface().(time.Time)
s := t.Format(time.RFC3339Nano)
e.emitScalar(s, "", tag, yaml_PLAIN_SCALAR_STYLE)
}
func (e *encoder) floatv(tag string, in reflect.Value) {
// Issue #352: When formatting, use the precision of the underlying value
precision := 64
if in.Kind() == reflect.Float32 {
precision = 32
}
s := strconv.FormatFloat(in.Float(), 'g', -1, precision)
switch s {
case "+Inf":
s = ".inf"
case "-Inf":
s = "-.inf"
case "NaN":
s = ".nan"
}
e.emitScalar(s, "", tag, yaml_PLAIN_SCALAR_STYLE)
}
func (e *encoder) nilv() {
e.emitScalar("null", "", "", yaml_PLAIN_SCALAR_STYLE)
}
func (e *encoder) emitScalar(value, anchor, tag string, style yaml_scalar_style_t) {
implicit := tag == ""
e.must(yaml_scalar_event_initialize(&e.event, []byte(anchor), []byte(tag), []byte(value), implicit, implicit, style))
e.emit()
}

5
vendor/gopkg.in/yaml.v2/go.mod generated vendored
View File

@ -1,5 +0,0 @@
module "gopkg.in/yaml.v2"
require (
"gopkg.in/check.v1" v0.0.0-20161208181325-20d25e280405
)

1095
vendor/gopkg.in/yaml.v2/parserc.go generated vendored

File diff suppressed because it is too large Load Diff

412
vendor/gopkg.in/yaml.v2/readerc.go generated vendored
View File

@ -1,412 +0,0 @@
package yaml
import (
"io"
)
// Set the reader error and return 0.
func yaml_parser_set_reader_error(parser *yaml_parser_t, problem string, offset int, value int) bool {
parser.error = yaml_READER_ERROR
parser.problem = problem
parser.problem_offset = offset
parser.problem_value = value
return false
}
// Byte order marks.
const (
bom_UTF8 = "\xef\xbb\xbf"
bom_UTF16LE = "\xff\xfe"
bom_UTF16BE = "\xfe\xff"
)
// Determine the input stream encoding by checking the BOM symbol. If no BOM is
// found, the UTF-8 encoding is assumed. Return 1 on success, 0 on failure.
func yaml_parser_determine_encoding(parser *yaml_parser_t) bool {
// Ensure that we had enough bytes in the raw buffer.
for !parser.eof && len(parser.raw_buffer)-parser.raw_buffer_pos < 3 {
if !yaml_parser_update_raw_buffer(parser) {
return false
}
}
// Determine the encoding.
buf := parser.raw_buffer
pos := parser.raw_buffer_pos
avail := len(buf) - pos
if avail >= 2 && buf[pos] == bom_UTF16LE[0] && buf[pos+1] == bom_UTF16LE[1] {
parser.encoding = yaml_UTF16LE_ENCODING
parser.raw_buffer_pos += 2
parser.offset += 2
} else if avail >= 2 && buf[pos] == bom_UTF16BE[0] && buf[pos+1] == bom_UTF16BE[1] {
parser.encoding = yaml_UTF16BE_ENCODING
parser.raw_buffer_pos += 2
parser.offset += 2
} else if avail >= 3 && buf[pos] == bom_UTF8[0] && buf[pos+1] == bom_UTF8[1] && buf[pos+2] == bom_UTF8[2] {
parser.encoding = yaml_UTF8_ENCODING
parser.raw_buffer_pos += 3
parser.offset += 3
} else {
parser.encoding = yaml_UTF8_ENCODING
}
return true
}
// Update the raw buffer.
func yaml_parser_update_raw_buffer(parser *yaml_parser_t) bool {
size_read := 0
// Return if the raw buffer is full.
if parser.raw_buffer_pos == 0 && len(parser.raw_buffer) == cap(parser.raw_buffer) {
return true
}
// Return on EOF.
if parser.eof {
return true
}
// Move the remaining bytes in the raw buffer to the beginning.
if parser.raw_buffer_pos > 0 && parser.raw_buffer_pos < len(parser.raw_buffer) {
copy(parser.raw_buffer, parser.raw_buffer[parser.raw_buffer_pos:])
}
parser.raw_buffer = parser.raw_buffer[:len(parser.raw_buffer)-parser.raw_buffer_pos]
parser.raw_buffer_pos = 0
// Call the read handler to fill the buffer.
size_read, err := parser.read_handler(parser, parser.raw_buffer[len(parser.raw_buffer):cap(parser.raw_buffer)])
parser.raw_buffer = parser.raw_buffer[:len(parser.raw_buffer)+size_read]
if err == io.EOF {
parser.eof = true
} else if err != nil {
return yaml_parser_set_reader_error(parser, "input error: "+err.Error(), parser.offset, -1)
}
return true
}
// Ensure that the buffer contains at least `length` characters.
// Return true on success, false on failure.
//
// The length is supposed to be significantly less that the buffer size.
func yaml_parser_update_buffer(parser *yaml_parser_t, length int) bool {
if parser.read_handler == nil {
panic("read handler must be set")
}
// [Go] This function was changed to guarantee the requested length size at EOF.
// The fact we need to do this is pretty awful, but the description above implies
// for that to be the case, and there are tests
// If the EOF flag is set and the raw buffer is empty, do nothing.
if parser.eof && parser.raw_buffer_pos == len(parser.raw_buffer) {
// [Go] ACTUALLY! Read the documentation of this function above.
// This is just broken. To return true, we need to have the
// given length in the buffer. Not doing that means every single
// check that calls this function to make sure the buffer has a
// given length is Go) panicking; or C) accessing invalid memory.
//return true
}
// Return if the buffer contains enough characters.
if parser.unread >= length {
return true
}
// Determine the input encoding if it is not known yet.
if parser.encoding == yaml_ANY_ENCODING {
if !yaml_parser_determine_encoding(parser) {
return false
}
}
// Move the unread characters to the beginning of the buffer.
buffer_len := len(parser.buffer)
if parser.buffer_pos > 0 && parser.buffer_pos < buffer_len {
copy(parser.buffer, parser.buffer[parser.buffer_pos:])
buffer_len -= parser.buffer_pos
parser.buffer_pos = 0
} else if parser.buffer_pos == buffer_len {
buffer_len = 0
parser.buffer_pos = 0
}
// Open the whole buffer for writing, and cut it before returning.
parser.buffer = parser.buffer[:cap(parser.buffer)]
// Fill the buffer until it has enough characters.
first := true
for parser.unread < length {
// Fill the raw buffer if necessary.
if !first || parser.raw_buffer_pos == len(parser.raw_buffer) {
if !yaml_parser_update_raw_buffer(parser) {
parser.buffer = parser.buffer[:buffer_len]
return false
}
}
first = false
// Decode the raw buffer.
inner:
for parser.raw_buffer_pos != len(parser.raw_buffer) {
var value rune
var width int
raw_unread := len(parser.raw_buffer) - parser.raw_buffer_pos
// Decode the next character.
switch parser.encoding {
case yaml_UTF8_ENCODING:
// Decode a UTF-8 character. Check RFC 3629
// (http://www.ietf.org/rfc/rfc3629.txt) for more details.
//
// The following table (taken from the RFC) is used for
// decoding.
//
// Char. number range | UTF-8 octet sequence
// (hexadecimal) | (binary)
// --------------------+------------------------------------
// 0000 0000-0000 007F | 0xxxxxxx
// 0000 0080-0000 07FF | 110xxxxx 10xxxxxx
// 0000 0800-0000 FFFF | 1110xxxx 10xxxxxx 10xxxxxx
// 0001 0000-0010 FFFF | 11110xxx 10xxxxxx 10xxxxxx 10xxxxxx
//
// Additionally, the characters in the range 0xD800-0xDFFF
// are prohibited as they are reserved for use with UTF-16
// surrogate pairs.
// Determine the length of the UTF-8 sequence.
octet := parser.raw_buffer[parser.raw_buffer_pos]
switch {
case octet&0x80 == 0x00:
width = 1
case octet&0xE0 == 0xC0:
width = 2
case octet&0xF0 == 0xE0:
width = 3
case octet&0xF8 == 0xF0:
width = 4
default:
// The leading octet is invalid.
return yaml_parser_set_reader_error(parser,
"invalid leading UTF-8 octet",
parser.offset, int(octet))
}
// Check if the raw buffer contains an incomplete character.
if width > raw_unread {
if parser.eof {
return yaml_parser_set_reader_error(parser,
"incomplete UTF-8 octet sequence",
parser.offset, -1)
}
break inner
}
// Decode the leading octet.
switch {
case octet&0x80 == 0x00:
value = rune(octet & 0x7F)
case octet&0xE0 == 0xC0:
value = rune(octet & 0x1F)
case octet&0xF0 == 0xE0:
value = rune(octet & 0x0F)
case octet&0xF8 == 0xF0:
value = rune(octet & 0x07)
default:
value = 0
}
// Check and decode the trailing octets.
for k := 1; k < width; k++ {
octet = parser.raw_buffer[parser.raw_buffer_pos+k]
// Check if the octet is valid.
if (octet & 0xC0) != 0x80 {
return yaml_parser_set_reader_error(parser,
"invalid trailing UTF-8 octet",
parser.offset+k, int(octet))
}
// Decode the octet.
value = (value << 6) + rune(octet&0x3F)
}
// Check the length of the sequence against the value.
switch {
case width == 1:
case width == 2 && value >= 0x80:
case width == 3 && value >= 0x800:
case width == 4 && value >= 0x10000:
default:
return yaml_parser_set_reader_error(parser,
"invalid length of a UTF-8 sequence",
parser.offset, -1)
}
// Check the range of the value.
if value >= 0xD800 && value <= 0xDFFF || value > 0x10FFFF {
return yaml_parser_set_reader_error(parser,
"invalid Unicode character",
parser.offset, int(value))
}
case yaml_UTF16LE_ENCODING, yaml_UTF16BE_ENCODING:
var low, high int
if parser.encoding == yaml_UTF16LE_ENCODING {
low, high = 0, 1
} else {
low, high = 1, 0
}
// The UTF-16 encoding is not as simple as one might
// naively think. Check RFC 2781
// (http://www.ietf.org/rfc/rfc2781.txt).
//
// Normally, two subsequent bytes describe a Unicode
// character. However a special technique (called a
// surrogate pair) is used for specifying character
// values larger than 0xFFFF.
//
// A surrogate pair consists of two pseudo-characters:
// high surrogate area (0xD800-0xDBFF)
// low surrogate area (0xDC00-0xDFFF)
//
// The following formulas are used for decoding
// and encoding characters using surrogate pairs:
//
// U = U' + 0x10000 (0x01 00 00 <= U <= 0x10 FF FF)
// U' = yyyyyyyyyyxxxxxxxxxx (0 <= U' <= 0x0F FF FF)
// W1 = 110110yyyyyyyyyy
// W2 = 110111xxxxxxxxxx
//
// where U is the character value, W1 is the high surrogate
// area, W2 is the low surrogate area.
// Check for incomplete UTF-16 character.
if raw_unread < 2 {
if parser.eof {
return yaml_parser_set_reader_error(parser,
"incomplete UTF-16 character",
parser.offset, -1)
}
break inner
}
// Get the character.
value = rune(parser.raw_buffer[parser.raw_buffer_pos+low]) +
(rune(parser.raw_buffer[parser.raw_buffer_pos+high]) << 8)
// Check for unexpected low surrogate area.
if value&0xFC00 == 0xDC00 {
return yaml_parser_set_reader_error(parser,
"unexpected low surrogate area",
parser.offset, int(value))
}
// Check for a high surrogate area.
if value&0xFC00 == 0xD800 {
width = 4
// Check for incomplete surrogate pair.
if raw_unread < 4 {
if parser.eof {
return yaml_parser_set_reader_error(parser,
"incomplete UTF-16 surrogate pair",
parser.offset, -1)
}
break inner
}
// Get the next character.
value2 := rune(parser.raw_buffer[parser.raw_buffer_pos+low+2]) +
(rune(parser.raw_buffer[parser.raw_buffer_pos+high+2]) << 8)
// Check for a low surrogate area.
if value2&0xFC00 != 0xDC00 {
return yaml_parser_set_reader_error(parser,
"expected low surrogate area",
parser.offset+2, int(value2))
}
// Generate the value of the surrogate pair.
value = 0x10000 + ((value & 0x3FF) << 10) + (value2 & 0x3FF)
} else {
width = 2
}
default:
panic("impossible")
}
// Check if the character is in the allowed range:
// #x9 | #xA | #xD | [#x20-#x7E] (8 bit)
// | #x85 | [#xA0-#xD7FF] | [#xE000-#xFFFD] (16 bit)
// | [#x10000-#x10FFFF] (32 bit)
switch {
case value == 0x09:
case value == 0x0A:
case value == 0x0D:
case value >= 0x20 && value <= 0x7E:
case value == 0x85:
case value >= 0xA0 && value <= 0xD7FF:
case value >= 0xE000 && value <= 0xFFFD:
case value >= 0x10000 && value <= 0x10FFFF:
default:
return yaml_parser_set_reader_error(parser,
"control characters are not allowed",
parser.offset, int(value))
}
// Move the raw pointers.
parser.raw_buffer_pos += width
parser.offset += width
// Finally put the character into the buffer.
if value <= 0x7F {
// 0000 0000-0000 007F . 0xxxxxxx
parser.buffer[buffer_len+0] = byte(value)
buffer_len += 1
} else if value <= 0x7FF {
// 0000 0080-0000 07FF . 110xxxxx 10xxxxxx
parser.buffer[buffer_len+0] = byte(0xC0 + (value >> 6))
parser.buffer[buffer_len+1] = byte(0x80 + (value & 0x3F))
buffer_len += 2
} else if value <= 0xFFFF {
// 0000 0800-0000 FFFF . 1110xxxx 10xxxxxx 10xxxxxx
parser.buffer[buffer_len+0] = byte(0xE0 + (value >> 12))
parser.buffer[buffer_len+1] = byte(0x80 + ((value >> 6) & 0x3F))
parser.buffer[buffer_len+2] = byte(0x80 + (value & 0x3F))
buffer_len += 3
} else {
// 0001 0000-0010 FFFF . 11110xxx 10xxxxxx 10xxxxxx 10xxxxxx
parser.buffer[buffer_len+0] = byte(0xF0 + (value >> 18))
parser.buffer[buffer_len+1] = byte(0x80 + ((value >> 12) & 0x3F))
parser.buffer[buffer_len+2] = byte(0x80 + ((value >> 6) & 0x3F))
parser.buffer[buffer_len+3] = byte(0x80 + (value & 0x3F))
buffer_len += 4
}
parser.unread++
}
// On EOF, put NUL into the buffer and return.
if parser.eof {
parser.buffer[buffer_len] = 0
buffer_len++
parser.unread++
break
}
}
// [Go] Read the documentation of this function above. To return true,
// we need to have the given length in the buffer. Not doing that means
// every single check that calls this function to make sure the buffer
// has a given length is Go) panicking; or C) accessing invalid memory.
// This happens here due to the EOF above breaking early.
for buffer_len < length {
parser.buffer[buffer_len] = 0
buffer_len++
}
parser.buffer = parser.buffer[:buffer_len]
return true
}

258
vendor/gopkg.in/yaml.v2/resolve.go generated vendored
View File

@ -1,258 +0,0 @@
package yaml
import (
"encoding/base64"
"math"
"regexp"
"strconv"
"strings"
"time"
)
type resolveMapItem struct {
value interface{}
tag string
}
var resolveTable = make([]byte, 256)
var resolveMap = make(map[string]resolveMapItem)
func init() {
t := resolveTable
t[int('+')] = 'S' // Sign
t[int('-')] = 'S'
for _, c := range "0123456789" {
t[int(c)] = 'D' // Digit
}
for _, c := range "yYnNtTfFoO~" {
t[int(c)] = 'M' // In map
}
t[int('.')] = '.' // Float (potentially in map)
var resolveMapList = []struct {
v interface{}
tag string
l []string
}{
{true, yaml_BOOL_TAG, []string{"y", "Y", "yes", "Yes", "YES"}},
{true, yaml_BOOL_TAG, []string{"true", "True", "TRUE"}},
{true, yaml_BOOL_TAG, []string{"on", "On", "ON"}},
{false, yaml_BOOL_TAG, []string{"n", "N", "no", "No", "NO"}},
{false, yaml_BOOL_TAG, []string{"false", "False", "FALSE"}},
{false, yaml_BOOL_TAG, []string{"off", "Off", "OFF"}},
{nil, yaml_NULL_TAG, []string{"", "~", "null", "Null", "NULL"}},
{math.NaN(), yaml_FLOAT_TAG, []string{".nan", ".NaN", ".NAN"}},
{math.Inf(+1), yaml_FLOAT_TAG, []string{".inf", ".Inf", ".INF"}},
{math.Inf(+1), yaml_FLOAT_TAG, []string{"+.inf", "+.Inf", "+.INF"}},
{math.Inf(-1), yaml_FLOAT_TAG, []string{"-.inf", "-.Inf", "-.INF"}},
{"<<", yaml_MERGE_TAG, []string{"<<"}},
}
m := resolveMap
for _, item := range resolveMapList {
for _, s := range item.l {
m[s] = resolveMapItem{item.v, item.tag}
}
}
}
const longTagPrefix = "tag:yaml.org,2002:"
func shortTag(tag string) string {
// TODO This can easily be made faster and produce less garbage.
if strings.HasPrefix(tag, longTagPrefix) {
return "!!" + tag[len(longTagPrefix):]
}
return tag
}
func longTag(tag string) string {
if strings.HasPrefix(tag, "!!") {
return longTagPrefix + tag[2:]
}
return tag
}
func resolvableTag(tag string) bool {
switch tag {
case "", yaml_STR_TAG, yaml_BOOL_TAG, yaml_INT_TAG, yaml_FLOAT_TAG, yaml_NULL_TAG, yaml_TIMESTAMP_TAG:
return true
}
return false
}
var yamlStyleFloat = regexp.MustCompile(`^[-+]?[0-9]*\.?[0-9]+([eE][-+][0-9]+)?$`)
func resolve(tag string, in string) (rtag string, out interface{}) {
if !resolvableTag(tag) {
return tag, in
}
defer func() {
switch tag {
case "", rtag, yaml_STR_TAG, yaml_BINARY_TAG:
return
case yaml_FLOAT_TAG:
if rtag == yaml_INT_TAG {
switch v := out.(type) {
case int64:
rtag = yaml_FLOAT_TAG
out = float64(v)
return
case int:
rtag = yaml_FLOAT_TAG
out = float64(v)
return
}
}
}
failf("cannot decode %s `%s` as a %s", shortTag(rtag), in, shortTag(tag))
}()
// Any data is accepted as a !!str or !!binary.
// Otherwise, the prefix is enough of a hint about what it might be.
hint := byte('N')
if in != "" {
hint = resolveTable[in[0]]
}
if hint != 0 && tag != yaml_STR_TAG && tag != yaml_BINARY_TAG {
// Handle things we can lookup in a map.
if item, ok := resolveMap[in]; ok {
return item.tag, item.value
}
// Base 60 floats are a bad idea, were dropped in YAML 1.2, and
// are purposefully unsupported here. They're still quoted on
// the way out for compatibility with other parser, though.
switch hint {
case 'M':
// We've already checked the map above.
case '.':
// Not in the map, so maybe a normal float.
floatv, err := strconv.ParseFloat(in, 64)
if err == nil {
return yaml_FLOAT_TAG, floatv
}
case 'D', 'S':
// Int, float, or timestamp.
// Only try values as a timestamp if the value is unquoted or there's an explicit
// !!timestamp tag.
if tag == "" || tag == yaml_TIMESTAMP_TAG {
t, ok := parseTimestamp(in)
if ok {
return yaml_TIMESTAMP_TAG, t
}
}
plain := strings.Replace(in, "_", "", -1)
intv, err := strconv.ParseInt(plain, 0, 64)
if err == nil {
if intv == int64(int(intv)) {
return yaml_INT_TAG, int(intv)
} else {
return yaml_INT_TAG, intv
}
}
uintv, err := strconv.ParseUint(plain, 0, 64)
if err == nil {
return yaml_INT_TAG, uintv
}
if yamlStyleFloat.MatchString(plain) {
floatv, err := strconv.ParseFloat(plain, 64)
if err == nil {
return yaml_FLOAT_TAG, floatv
}
}
if strings.HasPrefix(plain, "0b") {
intv, err := strconv.ParseInt(plain[2:], 2, 64)
if err == nil {
if intv == int64(int(intv)) {
return yaml_INT_TAG, int(intv)
} else {
return yaml_INT_TAG, intv
}
}
uintv, err := strconv.ParseUint(plain[2:], 2, 64)
if err == nil {
return yaml_INT_TAG, uintv
}
} else if strings.HasPrefix(plain, "-0b") {
intv, err := strconv.ParseInt("-" + plain[3:], 2, 64)
if err == nil {
if true || intv == int64(int(intv)) {
return yaml_INT_TAG, int(intv)
} else {
return yaml_INT_TAG, intv
}
}
}
default:
panic("resolveTable item not yet handled: " + string(rune(hint)) + " (with " + in + ")")
}
}
return yaml_STR_TAG, in
}
// encodeBase64 encodes s as base64 that is broken up into multiple lines
// as appropriate for the resulting length.
func encodeBase64(s string) string {
const lineLen = 70
encLen := base64.StdEncoding.EncodedLen(len(s))
lines := encLen/lineLen + 1
buf := make([]byte, encLen*2+lines)
in := buf[0:encLen]
out := buf[encLen:]
base64.StdEncoding.Encode(in, []byte(s))
k := 0
for i := 0; i < len(in); i += lineLen {
j := i + lineLen
if j > len(in) {
j = len(in)
}
k += copy(out[k:], in[i:j])
if lines > 1 {
out[k] = '\n'
k++
}
}
return string(out[:k])
}
// This is a subset of the formats allowed by the regular expression
// defined at http://yaml.org/type/timestamp.html.
var allowedTimestampFormats = []string{
"2006-1-2T15:4:5.999999999Z07:00", // RCF3339Nano with short date fields.
"2006-1-2t15:4:5.999999999Z07:00", // RFC3339Nano with short date fields and lower-case "t".
"2006-1-2 15:4:5.999999999", // space separated with no time zone
"2006-1-2", // date only
// Notable exception: time.Parse cannot handle: "2001-12-14 21:59:43.10 -5"
// from the set of examples.
}
// parseTimestamp parses s as a timestamp string and
// returns the timestamp and reports whether it succeeded.
// Timestamp formats are defined at http://yaml.org/type/timestamp.html
func parseTimestamp(s string) (time.Time, bool) {
// TODO write code to check all the formats supported by
// http://yaml.org/type/timestamp.html instead of using time.Parse.
// Quick check: all date formats start with YYYY-.
i := 0
for ; i < len(s); i++ {
if c := s[i]; c < '0' || c > '9' {
break
}
}
if i != 4 || i == len(s) || s[i] != '-' {
return time.Time{}, false
}
for _, format := range allowedTimestampFormats {
if t, err := time.Parse(format, s); err == nil {
return t, true
}
}
return time.Time{}, false
}

2696
vendor/gopkg.in/yaml.v2/scannerc.go generated vendored

File diff suppressed because it is too large Load Diff

113
vendor/gopkg.in/yaml.v2/sorter.go generated vendored
View File

@ -1,113 +0,0 @@
package yaml
import (
"reflect"
"unicode"
)
type keyList []reflect.Value
func (l keyList) Len() int { return len(l) }
func (l keyList) Swap(i, j int) { l[i], l[j] = l[j], l[i] }
func (l keyList) Less(i, j int) bool {
a := l[i]
b := l[j]
ak := a.Kind()
bk := b.Kind()
for (ak == reflect.Interface || ak == reflect.Ptr) && !a.IsNil() {
a = a.Elem()
ak = a.Kind()
}
for (bk == reflect.Interface || bk == reflect.Ptr) && !b.IsNil() {
b = b.Elem()
bk = b.Kind()
}
af, aok := keyFloat(a)
bf, bok := keyFloat(b)
if aok && bok {
if af != bf {
return af < bf
}
if ak != bk {
return ak < bk
}
return numLess(a, b)
}
if ak != reflect.String || bk != reflect.String {
return ak < bk
}
ar, br := []rune(a.String()), []rune(b.String())
for i := 0; i < len(ar) && i < len(br); i++ {
if ar[i] == br[i] {
continue
}
al := unicode.IsLetter(ar[i])
bl := unicode.IsLetter(br[i])
if al && bl {
return ar[i] < br[i]
}
if al || bl {
return bl
}
var ai, bi int
var an, bn int64
if ar[i] == '0' || br[i] == '0' {
for j := i-1; j >= 0 && unicode.IsDigit(ar[j]); j-- {
if ar[j] != '0' {
an = 1
bn = 1
break
}
}
}
for ai = i; ai < len(ar) && unicode.IsDigit(ar[ai]); ai++ {
an = an*10 + int64(ar[ai]-'0')
}
for bi = i; bi < len(br) && unicode.IsDigit(br[bi]); bi++ {
bn = bn*10 + int64(br[bi]-'0')
}
if an != bn {
return an < bn
}
if ai != bi {
return ai < bi
}
return ar[i] < br[i]
}
return len(ar) < len(br)
}
// keyFloat returns a float value for v if it is a number/bool
// and whether it is a number/bool or not.
func keyFloat(v reflect.Value) (f float64, ok bool) {
switch v.Kind() {
case reflect.Int, reflect.Int8, reflect.Int16, reflect.Int32, reflect.Int64:
return float64(v.Int()), true
case reflect.Float32, reflect.Float64:
return v.Float(), true
case reflect.Uint, reflect.Uint8, reflect.Uint16, reflect.Uint32, reflect.Uint64, reflect.Uintptr:
return float64(v.Uint()), true
case reflect.Bool:
if v.Bool() {
return 1, true
}
return 0, true
}
return 0, false
}
// numLess returns whether a < b.
// a and b must necessarily have the same kind.
func numLess(a, b reflect.Value) bool {
switch a.Kind() {
case reflect.Int, reflect.Int8, reflect.Int16, reflect.Int32, reflect.Int64:
return a.Int() < b.Int()
case reflect.Float32, reflect.Float64:
return a.Float() < b.Float()
case reflect.Uint, reflect.Uint8, reflect.Uint16, reflect.Uint32, reflect.Uint64, reflect.Uintptr:
return a.Uint() < b.Uint()
case reflect.Bool:
return !a.Bool() && b.Bool()
}
panic("not a number")
}

26
vendor/gopkg.in/yaml.v2/writerc.go generated vendored
View File

@ -1,26 +0,0 @@
package yaml
// Set the writer error and return false.
func yaml_emitter_set_writer_error(emitter *yaml_emitter_t, problem string) bool {
emitter.error = yaml_WRITER_ERROR
emitter.problem = problem
return false
}
// Flush the output buffer.
func yaml_emitter_flush(emitter *yaml_emitter_t) bool {
if emitter.write_handler == nil {
panic("write handler not set")
}
// Check if the buffer is empty.
if emitter.buffer_pos == 0 {
return true
}
if err := emitter.write_handler(emitter, emitter.buffer[:emitter.buffer_pos]); err != nil {
return yaml_emitter_set_writer_error(emitter, "write error: "+err.Error())
}
emitter.buffer_pos = 0
return true
}

466
vendor/gopkg.in/yaml.v2/yaml.go generated vendored
View File

@ -1,466 +0,0 @@
// Package yaml implements YAML support for the Go language.
//
// Source code and other details for the project are available at GitHub:
//
// https://github.com/go-yaml/yaml
//
package yaml
import (
"errors"
"fmt"
"io"
"reflect"
"strings"
"sync"
)
// MapSlice encodes and decodes as a YAML map.
// The order of keys is preserved when encoding and decoding.
type MapSlice []MapItem
// MapItem is an item in a MapSlice.
type MapItem struct {
Key, Value interface{}
}
// The Unmarshaler interface may be implemented by types to customize their
// behavior when being unmarshaled from a YAML document. The UnmarshalYAML
// method receives a function that may be called to unmarshal the original
// YAML value into a field or variable. It is safe to call the unmarshal
// function parameter more than once if necessary.
type Unmarshaler interface {
UnmarshalYAML(unmarshal func(interface{}) error) error
}
// The Marshaler interface may be implemented by types to customize their
// behavior when being marshaled into a YAML document. The returned value
// is marshaled in place of the original value implementing Marshaler.
//
// If an error is returned by MarshalYAML, the marshaling procedure stops
// and returns with the provided error.
type Marshaler interface {
MarshalYAML() (interface{}, error)
}
// Unmarshal decodes the first document found within the in byte slice
// and assigns decoded values into the out value.
//
// Maps and pointers (to a struct, string, int, etc) are accepted as out
// values. If an internal pointer within a struct is not initialized,
// the yaml package will initialize it if necessary for unmarshalling
// the provided data. The out parameter must not be nil.
//
// The type of the decoded values should be compatible with the respective
// values in out. If one or more values cannot be decoded due to a type
// mismatches, decoding continues partially until the end of the YAML
// content, and a *yaml.TypeError is returned with details for all
// missed values.
//
// Struct fields are only unmarshalled if they are exported (have an
// upper case first letter), and are unmarshalled using the field name
// lowercased as the default key. Custom keys may be defined via the
// "yaml" name in the field tag: the content preceding the first comma
// is used as the key, and the following comma-separated options are
// used to tweak the marshalling process (see Marshal).
// Conflicting names result in a runtime error.
//
// For example:
//
// type T struct {
// F int `yaml:"a,omitempty"`
// B int
// }
// var t T
// yaml.Unmarshal([]byte("a: 1\nb: 2"), &t)
//
// See the documentation of Marshal for the format of tags and a list of
// supported tag options.
//
func Unmarshal(in []byte, out interface{}) (err error) {
return unmarshal(in, out, false)
}
// UnmarshalStrict is like Unmarshal except that any fields that are found
// in the data that do not have corresponding struct members, or mapping
// keys that are duplicates, will result in
// an error.
func UnmarshalStrict(in []byte, out interface{}) (err error) {
return unmarshal(in, out, true)
}
// A Decorder reads and decodes YAML values from an input stream.
type Decoder struct {
strict bool
parser *parser
}
// NewDecoder returns a new decoder that reads from r.
//
// The decoder introduces its own buffering and may read
// data from r beyond the YAML values requested.
func NewDecoder(r io.Reader) *Decoder {
return &Decoder{
parser: newParserFromReader(r),
}
}
// SetStrict sets whether strict decoding behaviour is enabled when
// decoding items in the data (see UnmarshalStrict). By default, decoding is not strict.
func (dec *Decoder) SetStrict(strict bool) {
dec.strict = strict
}
// Decode reads the next YAML-encoded value from its input
// and stores it in the value pointed to by v.
//
// See the documentation for Unmarshal for details about the
// conversion of YAML into a Go value.
func (dec *Decoder) Decode(v interface{}) (err error) {
d := newDecoder(dec.strict)
defer handleErr(&err)
node := dec.parser.parse()
if node == nil {
return io.EOF
}
out := reflect.ValueOf(v)
if out.Kind() == reflect.Ptr && !out.IsNil() {
out = out.Elem()
}
d.unmarshal(node, out)
if len(d.terrors) > 0 {
return &TypeError{d.terrors}
}
return nil
}
func unmarshal(in []byte, out interface{}, strict bool) (err error) {
defer handleErr(&err)
d := newDecoder(strict)
p := newParser(in)
defer p.destroy()
node := p.parse()
if node != nil {
v := reflect.ValueOf(out)
if v.Kind() == reflect.Ptr && !v.IsNil() {
v = v.Elem()
}
d.unmarshal(node, v)
}
if len(d.terrors) > 0 {
return &TypeError{d.terrors}
}
return nil
}
// Marshal serializes the value provided into a YAML document. The structure
// of the generated document will reflect the structure of the value itself.
// Maps and pointers (to struct, string, int, etc) are accepted as the in value.
//
// Struct fields are only marshalled if they are exported (have an upper case
// first letter), and are marshalled using the field name lowercased as the
// default key. Custom keys may be defined via the "yaml" name in the field
// tag: the content preceding the first comma is used as the key, and the
// following comma-separated options are used to tweak the marshalling process.
// Conflicting names result in a runtime error.
//
// The field tag format accepted is:
//
// `(...) yaml:"[<key>][,<flag1>[,<flag2>]]" (...)`
//
// The following flags are currently supported:
//
// omitempty Only include the field if it's not set to the zero
// value for the type or to empty slices or maps.
// Zero valued structs will be omitted if all their public
// fields are zero, unless they implement an IsZero
// method (see the IsZeroer interface type), in which
// case the field will be included if that method returns true.
//
// flow Marshal using a flow style (useful for structs,
// sequences and maps).
//
// inline Inline the field, which must be a struct or a map,
// causing all of its fields or keys to be processed as if
// they were part of the outer struct. For maps, keys must
// not conflict with the yaml keys of other struct fields.
//
// In addition, if the key is "-", the field is ignored.
//
// For example:
//
// type T struct {
// F int `yaml:"a,omitempty"`
// B int
// }
// yaml.Marshal(&T{B: 2}) // Returns "b: 2\n"
// yaml.Marshal(&T{F: 1}} // Returns "a: 1\nb: 0\n"
//
func Marshal(in interface{}) (out []byte, err error) {
defer handleErr(&err)
e := newEncoder()
defer e.destroy()
e.marshalDoc("", reflect.ValueOf(in))
e.finish()
out = e.out
return
}
// An Encoder writes YAML values to an output stream.
type Encoder struct {
encoder *encoder
}
// NewEncoder returns a new encoder that writes to w.
// The Encoder should be closed after use to flush all data
// to w.
func NewEncoder(w io.Writer) *Encoder {
return &Encoder{
encoder: newEncoderWithWriter(w),
}
}
// Encode writes the YAML encoding of v to the stream.
// If multiple items are encoded to the stream, the
// second and subsequent document will be preceded
// with a "---" document separator, but the first will not.
//
// See the documentation for Marshal for details about the conversion of Go
// values to YAML.
func (e *Encoder) Encode(v interface{}) (err error) {
defer handleErr(&err)
e.encoder.marshalDoc("", reflect.ValueOf(v))
return nil
}
// Close closes the encoder by writing any remaining data.
// It does not write a stream terminating string "...".
func (e *Encoder) Close() (err error) {
defer handleErr(&err)
e.encoder.finish()
return nil
}
func handleErr(err *error) {
if v := recover(); v != nil {
if e, ok := v.(yamlError); ok {
*err = e.err
} else {
panic(v)
}
}
}
type yamlError struct {
err error
}
func fail(err error) {
panic(yamlError{err})
}
func failf(format string, args ...interface{}) {
panic(yamlError{fmt.Errorf("yaml: "+format, args...)})
}
// A TypeError is returned by Unmarshal when one or more fields in
// the YAML document cannot be properly decoded into the requested
// types. When this error is returned, the value is still
// unmarshaled partially.
type TypeError struct {
Errors []string
}
func (e *TypeError) Error() string {
return fmt.Sprintf("yaml: unmarshal errors:\n %s", strings.Join(e.Errors, "\n "))
}
// --------------------------------------------------------------------------
// Maintain a mapping of keys to structure field indexes
// The code in this section was copied from mgo/bson.
// structInfo holds details for the serialization of fields of
// a given struct.
type structInfo struct {
FieldsMap map[string]fieldInfo
FieldsList []fieldInfo
// InlineMap is the number of the field in the struct that
// contains an ,inline map, or -1 if there's none.
InlineMap int
}
type fieldInfo struct {
Key string
Num int
OmitEmpty bool
Flow bool
// Id holds the unique field identifier, so we can cheaply
// check for field duplicates without maintaining an extra map.
Id int
// Inline holds the field index if the field is part of an inlined struct.
Inline []int
}
var structMap = make(map[reflect.Type]*structInfo)
var fieldMapMutex sync.RWMutex
func getStructInfo(st reflect.Type) (*structInfo, error) {
fieldMapMutex.RLock()
sinfo, found := structMap[st]
fieldMapMutex.RUnlock()
if found {
return sinfo, nil
}
n := st.NumField()
fieldsMap := make(map[string]fieldInfo)
fieldsList := make([]fieldInfo, 0, n)
inlineMap := -1
for i := 0; i != n; i++ {
field := st.Field(i)
if field.PkgPath != "" && !field.Anonymous {
continue // Private field
}
info := fieldInfo{Num: i}
tag := field.Tag.Get("yaml")
if tag == "" && strings.Index(string(field.Tag), ":") < 0 {
tag = string(field.Tag)
}
if tag == "-" {
continue
}
inline := false
fields := strings.Split(tag, ",")
if len(fields) > 1 {
for _, flag := range fields[1:] {
switch flag {
case "omitempty":
info.OmitEmpty = true
case "flow":
info.Flow = true
case "inline":
inline = true
default:
return nil, errors.New(fmt.Sprintf("Unsupported flag %q in tag %q of type %s", flag, tag, st))
}
}
tag = fields[0]
}
if inline {
switch field.Type.Kind() {
case reflect.Map:
if inlineMap >= 0 {
return nil, errors.New("Multiple ,inline maps in struct " + st.String())
}
if field.Type.Key() != reflect.TypeOf("") {
return nil, errors.New("Option ,inline needs a map with string keys in struct " + st.String())
}
inlineMap = info.Num
case reflect.Struct:
sinfo, err := getStructInfo(field.Type)
if err != nil {
return nil, err
}
for _, finfo := range sinfo.FieldsList {
if _, found := fieldsMap[finfo.Key]; found {
msg := "Duplicated key '" + finfo.Key + "' in struct " + st.String()
return nil, errors.New(msg)
}
if finfo.Inline == nil {
finfo.Inline = []int{i, finfo.Num}
} else {
finfo.Inline = append([]int{i}, finfo.Inline...)
}
finfo.Id = len(fieldsList)
fieldsMap[finfo.Key] = finfo
fieldsList = append(fieldsList, finfo)
}
default:
//return nil, errors.New("Option ,inline needs a struct value or map field")
return nil, errors.New("Option ,inline needs a struct value field")
}
continue
}
if tag != "" {
info.Key = tag
} else {
info.Key = strings.ToLower(field.Name)
}
if _, found = fieldsMap[info.Key]; found {
msg := "Duplicated key '" + info.Key + "' in struct " + st.String()
return nil, errors.New(msg)
}
info.Id = len(fieldsList)
fieldsList = append(fieldsList, info)
fieldsMap[info.Key] = info
}
sinfo = &structInfo{
FieldsMap: fieldsMap,
FieldsList: fieldsList,
InlineMap: inlineMap,
}
fieldMapMutex.Lock()
structMap[st] = sinfo
fieldMapMutex.Unlock()
return sinfo, nil
}
// IsZeroer is used to check whether an object is zero to
// determine whether it should be omitted when marshaling
// with the omitempty flag. One notable implementation
// is time.Time.
type IsZeroer interface {
IsZero() bool
}
func isZero(v reflect.Value) bool {
kind := v.Kind()
if z, ok := v.Interface().(IsZeroer); ok {
if (kind == reflect.Ptr || kind == reflect.Interface) && v.IsNil() {
return true
}
return z.IsZero()
}
switch kind {
case reflect.String:
return len(v.String()) == 0
case reflect.Interface, reflect.Ptr:
return v.IsNil()
case reflect.Slice:
return v.Len() == 0
case reflect.Map:
return v.Len() == 0
case reflect.Int, reflect.Int8, reflect.Int16, reflect.Int32, reflect.Int64:
return v.Int() == 0
case reflect.Float32, reflect.Float64:
return v.Float() == 0
case reflect.Uint, reflect.Uint8, reflect.Uint16, reflect.Uint32, reflect.Uint64, reflect.Uintptr:
return v.Uint() == 0
case reflect.Bool:
return !v.Bool()
case reflect.Struct:
vt := v.Type()
for i := v.NumField() - 1; i >= 0; i-- {
if vt.Field(i).PkgPath != "" {
continue // Private field
}
if !isZero(v.Field(i)) {
return false
}
}
return true
}
return false
}

738
vendor/gopkg.in/yaml.v2/yamlh.go generated vendored
View File

@ -1,738 +0,0 @@
package yaml
import (
"fmt"
"io"
)
// The version directive data.
type yaml_version_directive_t struct {
major int8 // The major version number.
minor int8 // The minor version number.
}
// The tag directive data.
type yaml_tag_directive_t struct {
handle []byte // The tag handle.
prefix []byte // The tag prefix.
}
type yaml_encoding_t int
// The stream encoding.
const (
// Let the parser choose the encoding.
yaml_ANY_ENCODING yaml_encoding_t = iota
yaml_UTF8_ENCODING // The default UTF-8 encoding.
yaml_UTF16LE_ENCODING // The UTF-16-LE encoding with BOM.
yaml_UTF16BE_ENCODING // The UTF-16-BE encoding with BOM.
)
type yaml_break_t int
// Line break types.
const (
// Let the parser choose the break type.
yaml_ANY_BREAK yaml_break_t = iota
yaml_CR_BREAK // Use CR for line breaks (Mac style).
yaml_LN_BREAK // Use LN for line breaks (Unix style).
yaml_CRLN_BREAK // Use CR LN for line breaks (DOS style).
)
type yaml_error_type_t int
// Many bad things could happen with the parser and emitter.
const (
// No error is produced.
yaml_NO_ERROR yaml_error_type_t = iota
yaml_MEMORY_ERROR // Cannot allocate or reallocate a block of memory.
yaml_READER_ERROR // Cannot read or decode the input stream.
yaml_SCANNER_ERROR // Cannot scan the input stream.
yaml_PARSER_ERROR // Cannot parse the input stream.
yaml_COMPOSER_ERROR // Cannot compose a YAML document.
yaml_WRITER_ERROR // Cannot write to the output stream.
yaml_EMITTER_ERROR // Cannot emit a YAML stream.
)
// The pointer position.
type yaml_mark_t struct {
index int // The position index.
line int // The position line.
column int // The position column.
}
// Node Styles
type yaml_style_t int8
type yaml_scalar_style_t yaml_style_t
// Scalar styles.
const (
// Let the emitter choose the style.
yaml_ANY_SCALAR_STYLE yaml_scalar_style_t = iota
yaml_PLAIN_SCALAR_STYLE // The plain scalar style.
yaml_SINGLE_QUOTED_SCALAR_STYLE // The single-quoted scalar style.
yaml_DOUBLE_QUOTED_SCALAR_STYLE // The double-quoted scalar style.
yaml_LITERAL_SCALAR_STYLE // The literal scalar style.
yaml_FOLDED_SCALAR_STYLE // The folded scalar style.
)
type yaml_sequence_style_t yaml_style_t
// Sequence styles.
const (
// Let the emitter choose the style.
yaml_ANY_SEQUENCE_STYLE yaml_sequence_style_t = iota
yaml_BLOCK_SEQUENCE_STYLE // The block sequence style.
yaml_FLOW_SEQUENCE_STYLE // The flow sequence style.
)
type yaml_mapping_style_t yaml_style_t
// Mapping styles.
const (
// Let the emitter choose the style.
yaml_ANY_MAPPING_STYLE yaml_mapping_style_t = iota
yaml_BLOCK_MAPPING_STYLE // The block mapping style.
yaml_FLOW_MAPPING_STYLE // The flow mapping style.
)
// Tokens
type yaml_token_type_t int
// Token types.
const (
// An empty token.
yaml_NO_TOKEN yaml_token_type_t = iota
yaml_STREAM_START_TOKEN // A STREAM-START token.
yaml_STREAM_END_TOKEN // A STREAM-END token.
yaml_VERSION_DIRECTIVE_TOKEN // A VERSION-DIRECTIVE token.
yaml_TAG_DIRECTIVE_TOKEN // A TAG-DIRECTIVE token.
yaml_DOCUMENT_START_TOKEN // A DOCUMENT-START token.
yaml_DOCUMENT_END_TOKEN // A DOCUMENT-END token.
yaml_BLOCK_SEQUENCE_START_TOKEN // A BLOCK-SEQUENCE-START token.
yaml_BLOCK_MAPPING_START_TOKEN // A BLOCK-SEQUENCE-END token.
yaml_BLOCK_END_TOKEN // A BLOCK-END token.
yaml_FLOW_SEQUENCE_START_TOKEN // A FLOW-SEQUENCE-START token.
yaml_FLOW_SEQUENCE_END_TOKEN // A FLOW-SEQUENCE-END token.
yaml_FLOW_MAPPING_START_TOKEN // A FLOW-MAPPING-START token.
yaml_FLOW_MAPPING_END_TOKEN // A FLOW-MAPPING-END token.
yaml_BLOCK_ENTRY_TOKEN // A BLOCK-ENTRY token.
yaml_FLOW_ENTRY_TOKEN // A FLOW-ENTRY token.
yaml_KEY_TOKEN // A KEY token.
yaml_VALUE_TOKEN // A VALUE token.
yaml_ALIAS_TOKEN // An ALIAS token.
yaml_ANCHOR_TOKEN // An ANCHOR token.
yaml_TAG_TOKEN // A TAG token.
yaml_SCALAR_TOKEN // A SCALAR token.
)
func (tt yaml_token_type_t) String() string {
switch tt {
case yaml_NO_TOKEN:
return "yaml_NO_TOKEN"
case yaml_STREAM_START_TOKEN:
return "yaml_STREAM_START_TOKEN"
case yaml_STREAM_END_TOKEN:
return "yaml_STREAM_END_TOKEN"
case yaml_VERSION_DIRECTIVE_TOKEN:
return "yaml_VERSION_DIRECTIVE_TOKEN"
case yaml_TAG_DIRECTIVE_TOKEN:
return "yaml_TAG_DIRECTIVE_TOKEN"
case yaml_DOCUMENT_START_TOKEN:
return "yaml_DOCUMENT_START_TOKEN"
case yaml_DOCUMENT_END_TOKEN:
return "yaml_DOCUMENT_END_TOKEN"
case yaml_BLOCK_SEQUENCE_START_TOKEN:
return "yaml_BLOCK_SEQUENCE_START_TOKEN"
case yaml_BLOCK_MAPPING_START_TOKEN:
return "yaml_BLOCK_MAPPING_START_TOKEN"
case yaml_BLOCK_END_TOKEN:
return "yaml_BLOCK_END_TOKEN"
case yaml_FLOW_SEQUENCE_START_TOKEN:
return "yaml_FLOW_SEQUENCE_START_TOKEN"
case yaml_FLOW_SEQUENCE_END_TOKEN:
return "yaml_FLOW_SEQUENCE_END_TOKEN"
case yaml_FLOW_MAPPING_START_TOKEN:
return "yaml_FLOW_MAPPING_START_TOKEN"
case yaml_FLOW_MAPPING_END_TOKEN:
return "yaml_FLOW_MAPPING_END_TOKEN"
case yaml_BLOCK_ENTRY_TOKEN:
return "yaml_BLOCK_ENTRY_TOKEN"
case yaml_FLOW_ENTRY_TOKEN:
return "yaml_FLOW_ENTRY_TOKEN"
case yaml_KEY_TOKEN:
return "yaml_KEY_TOKEN"
case yaml_VALUE_TOKEN:
return "yaml_VALUE_TOKEN"
case yaml_ALIAS_TOKEN:
return "yaml_ALIAS_TOKEN"
case yaml_ANCHOR_TOKEN:
return "yaml_ANCHOR_TOKEN"
case yaml_TAG_TOKEN:
return "yaml_TAG_TOKEN"
case yaml_SCALAR_TOKEN:
return "yaml_SCALAR_TOKEN"
}
return "<unknown token>"
}
// The token structure.
type yaml_token_t struct {
// The token type.
typ yaml_token_type_t
// The start/end of the token.
start_mark, end_mark yaml_mark_t
// The stream encoding (for yaml_STREAM_START_TOKEN).
encoding yaml_encoding_t
// The alias/anchor/scalar value or tag/tag directive handle
// (for yaml_ALIAS_TOKEN, yaml_ANCHOR_TOKEN, yaml_SCALAR_TOKEN, yaml_TAG_TOKEN, yaml_TAG_DIRECTIVE_TOKEN).
value []byte
// The tag suffix (for yaml_TAG_TOKEN).
suffix []byte
// The tag directive prefix (for yaml_TAG_DIRECTIVE_TOKEN).
prefix []byte
// The scalar style (for yaml_SCALAR_TOKEN).
style yaml_scalar_style_t
// The version directive major/minor (for yaml_VERSION_DIRECTIVE_TOKEN).
major, minor int8
}
// Events
type yaml_event_type_t int8
// Event types.
const (
// An empty event.
yaml_NO_EVENT yaml_event_type_t = iota
yaml_STREAM_START_EVENT // A STREAM-START event.
yaml_STREAM_END_EVENT // A STREAM-END event.
yaml_DOCUMENT_START_EVENT // A DOCUMENT-START event.
yaml_DOCUMENT_END_EVENT // A DOCUMENT-END event.
yaml_ALIAS_EVENT // An ALIAS event.
yaml_SCALAR_EVENT // A SCALAR event.
yaml_SEQUENCE_START_EVENT // A SEQUENCE-START event.
yaml_SEQUENCE_END_EVENT // A SEQUENCE-END event.
yaml_MAPPING_START_EVENT // A MAPPING-START event.
yaml_MAPPING_END_EVENT // A MAPPING-END event.
)
var eventStrings = []string{
yaml_NO_EVENT: "none",
yaml_STREAM_START_EVENT: "stream start",
yaml_STREAM_END_EVENT: "stream end",
yaml_DOCUMENT_START_EVENT: "document start",
yaml_DOCUMENT_END_EVENT: "document end",
yaml_ALIAS_EVENT: "alias",
yaml_SCALAR_EVENT: "scalar",
yaml_SEQUENCE_START_EVENT: "sequence start",
yaml_SEQUENCE_END_EVENT: "sequence end",
yaml_MAPPING_START_EVENT: "mapping start",
yaml_MAPPING_END_EVENT: "mapping end",
}
func (e yaml_event_type_t) String() string {
if e < 0 || int(e) >= len(eventStrings) {
return fmt.Sprintf("unknown event %d", e)
}
return eventStrings[e]
}
// The event structure.
type yaml_event_t struct {
// The event type.
typ yaml_event_type_t
// The start and end of the event.
start_mark, end_mark yaml_mark_t
// The document encoding (for yaml_STREAM_START_EVENT).
encoding yaml_encoding_t
// The version directive (for yaml_DOCUMENT_START_EVENT).
version_directive *yaml_version_directive_t
// The list of tag directives (for yaml_DOCUMENT_START_EVENT).
tag_directives []yaml_tag_directive_t
// The anchor (for yaml_SCALAR_EVENT, yaml_SEQUENCE_START_EVENT, yaml_MAPPING_START_EVENT, yaml_ALIAS_EVENT).
anchor []byte
// The tag (for yaml_SCALAR_EVENT, yaml_SEQUENCE_START_EVENT, yaml_MAPPING_START_EVENT).
tag []byte
// The scalar value (for yaml_SCALAR_EVENT).
value []byte
// Is the document start/end indicator implicit, or the tag optional?
// (for yaml_DOCUMENT_START_EVENT, yaml_DOCUMENT_END_EVENT, yaml_SEQUENCE_START_EVENT, yaml_MAPPING_START_EVENT, yaml_SCALAR_EVENT).
implicit bool
// Is the tag optional for any non-plain style? (for yaml_SCALAR_EVENT).
quoted_implicit bool
// The style (for yaml_SCALAR_EVENT, yaml_SEQUENCE_START_EVENT, yaml_MAPPING_START_EVENT).
style yaml_style_t
}
func (e *yaml_event_t) scalar_style() yaml_scalar_style_t { return yaml_scalar_style_t(e.style) }
func (e *yaml_event_t) sequence_style() yaml_sequence_style_t { return yaml_sequence_style_t(e.style) }
func (e *yaml_event_t) mapping_style() yaml_mapping_style_t { return yaml_mapping_style_t(e.style) }
// Nodes
const (
yaml_NULL_TAG = "tag:yaml.org,2002:null" // The tag !!null with the only possible value: null.
yaml_BOOL_TAG = "tag:yaml.org,2002:bool" // The tag !!bool with the values: true and false.
yaml_STR_TAG = "tag:yaml.org,2002:str" // The tag !!str for string values.
yaml_INT_TAG = "tag:yaml.org,2002:int" // The tag !!int for integer values.
yaml_FLOAT_TAG = "tag:yaml.org,2002:float" // The tag !!float for float values.
yaml_TIMESTAMP_TAG = "tag:yaml.org,2002:timestamp" // The tag !!timestamp for date and time values.
yaml_SEQ_TAG = "tag:yaml.org,2002:seq" // The tag !!seq is used to denote sequences.
yaml_MAP_TAG = "tag:yaml.org,2002:map" // The tag !!map is used to denote mapping.
// Not in original libyaml.
yaml_BINARY_TAG = "tag:yaml.org,2002:binary"
yaml_MERGE_TAG = "tag:yaml.org,2002:merge"
yaml_DEFAULT_SCALAR_TAG = yaml_STR_TAG // The default scalar tag is !!str.
yaml_DEFAULT_SEQUENCE_TAG = yaml_SEQ_TAG // The default sequence tag is !!seq.
yaml_DEFAULT_MAPPING_TAG = yaml_MAP_TAG // The default mapping tag is !!map.
)
type yaml_node_type_t int
// Node types.
const (
// An empty node.
yaml_NO_NODE yaml_node_type_t = iota
yaml_SCALAR_NODE // A scalar node.
yaml_SEQUENCE_NODE // A sequence node.
yaml_MAPPING_NODE // A mapping node.
)
// An element of a sequence node.
type yaml_node_item_t int
// An element of a mapping node.
type yaml_node_pair_t struct {
key int // The key of the element.
value int // The value of the element.
}
// The node structure.
type yaml_node_t struct {
typ yaml_node_type_t // The node type.
tag []byte // The node tag.
// The node data.
// The scalar parameters (for yaml_SCALAR_NODE).
scalar struct {
value []byte // The scalar value.
length int // The length of the scalar value.
style yaml_scalar_style_t // The scalar style.
}
// The sequence parameters (for YAML_SEQUENCE_NODE).
sequence struct {
items_data []yaml_node_item_t // The stack of sequence items.
style yaml_sequence_style_t // The sequence style.
}
// The mapping parameters (for yaml_MAPPING_NODE).
mapping struct {
pairs_data []yaml_node_pair_t // The stack of mapping pairs (key, value).
pairs_start *yaml_node_pair_t // The beginning of the stack.
pairs_end *yaml_node_pair_t // The end of the stack.
pairs_top *yaml_node_pair_t // The top of the stack.
style yaml_mapping_style_t // The mapping style.
}
start_mark yaml_mark_t // The beginning of the node.
end_mark yaml_mark_t // The end of the node.
}
// The document structure.
type yaml_document_t struct {
// The document nodes.
nodes []yaml_node_t
// The version directive.
version_directive *yaml_version_directive_t
// The list of tag directives.
tag_directives_data []yaml_tag_directive_t
tag_directives_start int // The beginning of the tag directives list.
tag_directives_end int // The end of the tag directives list.
start_implicit int // Is the document start indicator implicit?
end_implicit int // Is the document end indicator implicit?
// The start/end of the document.
start_mark, end_mark yaml_mark_t
}
// The prototype of a read handler.
//
// The read handler is called when the parser needs to read more bytes from the
// source. The handler should write not more than size bytes to the buffer.
// The number of written bytes should be set to the size_read variable.
//
// [in,out] data A pointer to an application data specified by
// yaml_parser_set_input().
// [out] buffer The buffer to write the data from the source.
// [in] size The size of the buffer.
// [out] size_read The actual number of bytes read from the source.
//
// On success, the handler should return 1. If the handler failed,
// the returned value should be 0. On EOF, the handler should set the
// size_read to 0 and return 1.
type yaml_read_handler_t func(parser *yaml_parser_t, buffer []byte) (n int, err error)
// This structure holds information about a potential simple key.
type yaml_simple_key_t struct {
possible bool // Is a simple key possible?
required bool // Is a simple key required?
token_number int // The number of the token.
mark yaml_mark_t // The position mark.
}
// The states of the parser.
type yaml_parser_state_t int
const (
yaml_PARSE_STREAM_START_STATE yaml_parser_state_t = iota
yaml_PARSE_IMPLICIT_DOCUMENT_START_STATE // Expect the beginning of an implicit document.
yaml_PARSE_DOCUMENT_START_STATE // Expect DOCUMENT-START.
yaml_PARSE_DOCUMENT_CONTENT_STATE // Expect the content of a document.
yaml_PARSE_DOCUMENT_END_STATE // Expect DOCUMENT-END.
yaml_PARSE_BLOCK_NODE_STATE // Expect a block node.
yaml_PARSE_BLOCK_NODE_OR_INDENTLESS_SEQUENCE_STATE // Expect a block node or indentless sequence.
yaml_PARSE_FLOW_NODE_STATE // Expect a flow node.
yaml_PARSE_BLOCK_SEQUENCE_FIRST_ENTRY_STATE // Expect the first entry of a block sequence.
yaml_PARSE_BLOCK_SEQUENCE_ENTRY_STATE // Expect an entry of a block sequence.
yaml_PARSE_INDENTLESS_SEQUENCE_ENTRY_STATE // Expect an entry of an indentless sequence.
yaml_PARSE_BLOCK_MAPPING_FIRST_KEY_STATE // Expect the first key of a block mapping.
yaml_PARSE_BLOCK_MAPPING_KEY_STATE // Expect a block mapping key.
yaml_PARSE_BLOCK_MAPPING_VALUE_STATE // Expect a block mapping value.
yaml_PARSE_FLOW_SEQUENCE_FIRST_ENTRY_STATE // Expect the first entry of a flow sequence.
yaml_PARSE_FLOW_SEQUENCE_ENTRY_STATE // Expect an entry of a flow sequence.
yaml_PARSE_FLOW_SEQUENCE_ENTRY_MAPPING_KEY_STATE // Expect a key of an ordered mapping.
yaml_PARSE_FLOW_SEQUENCE_ENTRY_MAPPING_VALUE_STATE // Expect a value of an ordered mapping.
yaml_PARSE_FLOW_SEQUENCE_ENTRY_MAPPING_END_STATE // Expect the and of an ordered mapping entry.
yaml_PARSE_FLOW_MAPPING_FIRST_KEY_STATE // Expect the first key of a flow mapping.
yaml_PARSE_FLOW_MAPPING_KEY_STATE // Expect a key of a flow mapping.
yaml_PARSE_FLOW_MAPPING_VALUE_STATE // Expect a value of a flow mapping.
yaml_PARSE_FLOW_MAPPING_EMPTY_VALUE_STATE // Expect an empty value of a flow mapping.
yaml_PARSE_END_STATE // Expect nothing.
)
func (ps yaml_parser_state_t) String() string {
switch ps {
case yaml_PARSE_STREAM_START_STATE:
return "yaml_PARSE_STREAM_START_STATE"
case yaml_PARSE_IMPLICIT_DOCUMENT_START_STATE:
return "yaml_PARSE_IMPLICIT_DOCUMENT_START_STATE"
case yaml_PARSE_DOCUMENT_START_STATE:
return "yaml_PARSE_DOCUMENT_START_STATE"
case yaml_PARSE_DOCUMENT_CONTENT_STATE:
return "yaml_PARSE_DOCUMENT_CONTENT_STATE"
case yaml_PARSE_DOCUMENT_END_STATE:
return "yaml_PARSE_DOCUMENT_END_STATE"
case yaml_PARSE_BLOCK_NODE_STATE:
return "yaml_PARSE_BLOCK_NODE_STATE"
case yaml_PARSE_BLOCK_NODE_OR_INDENTLESS_SEQUENCE_STATE:
return "yaml_PARSE_BLOCK_NODE_OR_INDENTLESS_SEQUENCE_STATE"
case yaml_PARSE_FLOW_NODE_STATE:
return "yaml_PARSE_FLOW_NODE_STATE"
case yaml_PARSE_BLOCK_SEQUENCE_FIRST_ENTRY_STATE:
return "yaml_PARSE_BLOCK_SEQUENCE_FIRST_ENTRY_STATE"
case yaml_PARSE_BLOCK_SEQUENCE_ENTRY_STATE:
return "yaml_PARSE_BLOCK_SEQUENCE_ENTRY_STATE"
case yaml_PARSE_INDENTLESS_SEQUENCE_ENTRY_STATE:
return "yaml_PARSE_INDENTLESS_SEQUENCE_ENTRY_STATE"
case yaml_PARSE_BLOCK_MAPPING_FIRST_KEY_STATE:
return "yaml_PARSE_BLOCK_MAPPING_FIRST_KEY_STATE"
case yaml_PARSE_BLOCK_MAPPING_KEY_STATE:
return "yaml_PARSE_BLOCK_MAPPING_KEY_STATE"
case yaml_PARSE_BLOCK_MAPPING_VALUE_STATE:
return "yaml_PARSE_BLOCK_MAPPING_VALUE_STATE"
case yaml_PARSE_FLOW_SEQUENCE_FIRST_ENTRY_STATE:
return "yaml_PARSE_FLOW_SEQUENCE_FIRST_ENTRY_STATE"
case yaml_PARSE_FLOW_SEQUENCE_ENTRY_STATE:
return "yaml_PARSE_FLOW_SEQUENCE_ENTRY_STATE"
case yaml_PARSE_FLOW_SEQUENCE_ENTRY_MAPPING_KEY_STATE:
return "yaml_PARSE_FLOW_SEQUENCE_ENTRY_MAPPING_KEY_STATE"
case yaml_PARSE_FLOW_SEQUENCE_ENTRY_MAPPING_VALUE_STATE:
return "yaml_PARSE_FLOW_SEQUENCE_ENTRY_MAPPING_VALUE_STATE"
case yaml_PARSE_FLOW_SEQUENCE_ENTRY_MAPPING_END_STATE:
return "yaml_PARSE_FLOW_SEQUENCE_ENTRY_MAPPING_END_STATE"
case yaml_PARSE_FLOW_MAPPING_FIRST_KEY_STATE:
return "yaml_PARSE_FLOW_MAPPING_FIRST_KEY_STATE"
case yaml_PARSE_FLOW_MAPPING_KEY_STATE:
return "yaml_PARSE_FLOW_MAPPING_KEY_STATE"
case yaml_PARSE_FLOW_MAPPING_VALUE_STATE:
return "yaml_PARSE_FLOW_MAPPING_VALUE_STATE"
case yaml_PARSE_FLOW_MAPPING_EMPTY_VALUE_STATE:
return "yaml_PARSE_FLOW_MAPPING_EMPTY_VALUE_STATE"
case yaml_PARSE_END_STATE:
return "yaml_PARSE_END_STATE"
}
return "<unknown parser state>"
}
// This structure holds aliases data.
type yaml_alias_data_t struct {
anchor []byte // The anchor.
index int // The node id.
mark yaml_mark_t // The anchor mark.
}
// The parser structure.
//
// All members are internal. Manage the structure using the
// yaml_parser_ family of functions.
type yaml_parser_t struct {
// Error handling
error yaml_error_type_t // Error type.
problem string // Error description.
// The byte about which the problem occurred.
problem_offset int
problem_value int
problem_mark yaml_mark_t
// The error context.
context string
context_mark yaml_mark_t
// Reader stuff
read_handler yaml_read_handler_t // Read handler.
input_reader io.Reader // File input data.
input []byte // String input data.
input_pos int
eof bool // EOF flag
buffer []byte // The working buffer.
buffer_pos int // The current position of the buffer.
unread int // The number of unread characters in the buffer.
raw_buffer []byte // The raw buffer.
raw_buffer_pos int // The current position of the buffer.
encoding yaml_encoding_t // The input encoding.
offset int // The offset of the current position (in bytes).
mark yaml_mark_t // The mark of the current position.
// Scanner stuff
stream_start_produced bool // Have we started to scan the input stream?
stream_end_produced bool // Have we reached the end of the input stream?
flow_level int // The number of unclosed '[' and '{' indicators.
tokens []yaml_token_t // The tokens queue.
tokens_head int // The head of the tokens queue.
tokens_parsed int // The number of tokens fetched from the queue.
token_available bool // Does the tokens queue contain a token ready for dequeueing.
indent int // The current indentation level.
indents []int // The indentation levels stack.
simple_key_allowed bool // May a simple key occur at the current position?
simple_keys []yaml_simple_key_t // The stack of simple keys.
// Parser stuff
state yaml_parser_state_t // The current parser state.
states []yaml_parser_state_t // The parser states stack.
marks []yaml_mark_t // The stack of marks.
tag_directives []yaml_tag_directive_t // The list of TAG directives.
// Dumper stuff
aliases []yaml_alias_data_t // The alias data.
document *yaml_document_t // The currently parsed document.
}
// Emitter Definitions
// The prototype of a write handler.
//
// The write handler is called when the emitter needs to flush the accumulated
// characters to the output. The handler should write @a size bytes of the
// @a buffer to the output.
//
// @param[in,out] data A pointer to an application data specified by
// yaml_emitter_set_output().
// @param[in] buffer The buffer with bytes to be written.
// @param[in] size The size of the buffer.
//
// @returns On success, the handler should return @c 1. If the handler failed,
// the returned value should be @c 0.
//
type yaml_write_handler_t func(emitter *yaml_emitter_t, buffer []byte) error
type yaml_emitter_state_t int
// The emitter states.
const (
// Expect STREAM-START.
yaml_EMIT_STREAM_START_STATE yaml_emitter_state_t = iota
yaml_EMIT_FIRST_DOCUMENT_START_STATE // Expect the first DOCUMENT-START or STREAM-END.
yaml_EMIT_DOCUMENT_START_STATE // Expect DOCUMENT-START or STREAM-END.
yaml_EMIT_DOCUMENT_CONTENT_STATE // Expect the content of a document.
yaml_EMIT_DOCUMENT_END_STATE // Expect DOCUMENT-END.
yaml_EMIT_FLOW_SEQUENCE_FIRST_ITEM_STATE // Expect the first item of a flow sequence.
yaml_EMIT_FLOW_SEQUENCE_ITEM_STATE // Expect an item of a flow sequence.
yaml_EMIT_FLOW_MAPPING_FIRST_KEY_STATE // Expect the first key of a flow mapping.
yaml_EMIT_FLOW_MAPPING_KEY_STATE // Expect a key of a flow mapping.
yaml_EMIT_FLOW_MAPPING_SIMPLE_VALUE_STATE // Expect a value for a simple key of a flow mapping.
yaml_EMIT_FLOW_MAPPING_VALUE_STATE // Expect a value of a flow mapping.
yaml_EMIT_BLOCK_SEQUENCE_FIRST_ITEM_STATE // Expect the first item of a block sequence.
yaml_EMIT_BLOCK_SEQUENCE_ITEM_STATE // Expect an item of a block sequence.
yaml_EMIT_BLOCK_MAPPING_FIRST_KEY_STATE // Expect the first key of a block mapping.
yaml_EMIT_BLOCK_MAPPING_KEY_STATE // Expect the key of a block mapping.
yaml_EMIT_BLOCK_MAPPING_SIMPLE_VALUE_STATE // Expect a value for a simple key of a block mapping.
yaml_EMIT_BLOCK_MAPPING_VALUE_STATE // Expect a value of a block mapping.
yaml_EMIT_END_STATE // Expect nothing.
)
// The emitter structure.
//
// All members are internal. Manage the structure using the @c yaml_emitter_
// family of functions.
type yaml_emitter_t struct {
// Error handling
error yaml_error_type_t // Error type.
problem string // Error description.
// Writer stuff
write_handler yaml_write_handler_t // Write handler.
output_buffer *[]byte // String output data.
output_writer io.Writer // File output data.
buffer []byte // The working buffer.
buffer_pos int // The current position of the buffer.
raw_buffer []byte // The raw buffer.
raw_buffer_pos int // The current position of the buffer.
encoding yaml_encoding_t // The stream encoding.
// Emitter stuff
canonical bool // If the output is in the canonical style?
best_indent int // The number of indentation spaces.
best_width int // The preferred width of the output lines.
unicode bool // Allow unescaped non-ASCII characters?
line_break yaml_break_t // The preferred line break.
state yaml_emitter_state_t // The current emitter state.
states []yaml_emitter_state_t // The stack of states.
events []yaml_event_t // The event queue.
events_head int // The head of the event queue.
indents []int // The stack of indentation levels.
tag_directives []yaml_tag_directive_t // The list of tag directives.
indent int // The current indentation level.
flow_level int // The current flow level.
root_context bool // Is it the document root context?
sequence_context bool // Is it a sequence context?
mapping_context bool // Is it a mapping context?
simple_key_context bool // Is it a simple mapping key context?
line int // The current line.
column int // The current column.
whitespace bool // If the last character was a whitespace?
indention bool // If the last character was an indentation character (' ', '-', '?', ':')?
open_ended bool // If an explicit document end is required?
// Anchor analysis.
anchor_data struct {
anchor []byte // The anchor value.
alias bool // Is it an alias?
}
// Tag analysis.
tag_data struct {
handle []byte // The tag handle.
suffix []byte // The tag suffix.
}
// Scalar analysis.
scalar_data struct {
value []byte // The scalar value.
multiline bool // Does the scalar contain line breaks?
flow_plain_allowed bool // Can the scalar be expessed in the flow plain style?
block_plain_allowed bool // Can the scalar be expressed in the block plain style?
single_quoted_allowed bool // Can the scalar be expressed in the single quoted style?
block_allowed bool // Can the scalar be expressed in the literal or folded styles?
style yaml_scalar_style_t // The output style.
}
// Dumper stuff
opened bool // If the stream was already opened?
closed bool // If the stream was already closed?
// The information associated with the document nodes.
anchors *struct {
references int // The number of references.
anchor int // The anchor id.
serialized bool // If the node has been emitted?
}
last_anchor_id int // The last assigned anchor id.
document *yaml_document_t // The currently emitted document.
}

View File

@ -1,173 +0,0 @@
package yaml
const (
// The size of the input raw buffer.
input_raw_buffer_size = 512
// The size of the input buffer.
// It should be possible to decode the whole raw buffer.
input_buffer_size = input_raw_buffer_size * 3
// The size of the output buffer.
output_buffer_size = 128
// The size of the output raw buffer.
// It should be possible to encode the whole output buffer.
output_raw_buffer_size = (output_buffer_size*2 + 2)
// The size of other stacks and queues.
initial_stack_size = 16
initial_queue_size = 16
initial_string_size = 16
)
// Check if the character at the specified position is an alphabetical
// character, a digit, '_', or '-'.
func is_alpha(b []byte, i int) bool {
return b[i] >= '0' && b[i] <= '9' || b[i] >= 'A' && b[i] <= 'Z' || b[i] >= 'a' && b[i] <= 'z' || b[i] == '_' || b[i] == '-'
}
// Check if the character at the specified position is a digit.
func is_digit(b []byte, i int) bool {
return b[i] >= '0' && b[i] <= '9'
}
// Get the value of a digit.
func as_digit(b []byte, i int) int {
return int(b[i]) - '0'
}
// Check if the character at the specified position is a hex-digit.
func is_hex(b []byte, i int) bool {
return b[i] >= '0' && b[i] <= '9' || b[i] >= 'A' && b[i] <= 'F' || b[i] >= 'a' && b[i] <= 'f'
}
// Get the value of a hex-digit.
func as_hex(b []byte, i int) int {
bi := b[i]
if bi >= 'A' && bi <= 'F' {
return int(bi) - 'A' + 10
}
if bi >= 'a' && bi <= 'f' {
return int(bi) - 'a' + 10
}
return int(bi) - '0'
}
// Check if the character is ASCII.
func is_ascii(b []byte, i int) bool {
return b[i] <= 0x7F
}
// Check if the character at the start of the buffer can be printed unescaped.
func is_printable(b []byte, i int) bool {
return ((b[i] == 0x0A) || // . == #x0A
(b[i] >= 0x20 && b[i] <= 0x7E) || // #x20 <= . <= #x7E
(b[i] == 0xC2 && b[i+1] >= 0xA0) || // #0xA0 <= . <= #xD7FF
(b[i] > 0xC2 && b[i] < 0xED) ||
(b[i] == 0xED && b[i+1] < 0xA0) ||
(b[i] == 0xEE) ||
(b[i] == 0xEF && // #xE000 <= . <= #xFFFD
!(b[i+1] == 0xBB && b[i+2] == 0xBF) && // && . != #xFEFF
!(b[i+1] == 0xBF && (b[i+2] == 0xBE || b[i+2] == 0xBF))))
}
// Check if the character at the specified position is NUL.
func is_z(b []byte, i int) bool {
return b[i] == 0x00
}
// Check if the beginning of the buffer is a BOM.
func is_bom(b []byte, i int) bool {
return b[0] == 0xEF && b[1] == 0xBB && b[2] == 0xBF
}
// Check if the character at the specified position is space.
func is_space(b []byte, i int) bool {
return b[i] == ' '
}
// Check if the character at the specified position is tab.
func is_tab(b []byte, i int) bool {
return b[i] == '\t'
}
// Check if the character at the specified position is blank (space or tab).
func is_blank(b []byte, i int) bool {
//return is_space(b, i) || is_tab(b, i)
return b[i] == ' ' || b[i] == '\t'
}
// Check if the character at the specified position is a line break.
func is_break(b []byte, i int) bool {
return (b[i] == '\r' || // CR (#xD)
b[i] == '\n' || // LF (#xA)
b[i] == 0xC2 && b[i+1] == 0x85 || // NEL (#x85)
b[i] == 0xE2 && b[i+1] == 0x80 && b[i+2] == 0xA8 || // LS (#x2028)
b[i] == 0xE2 && b[i+1] == 0x80 && b[i+2] == 0xA9) // PS (#x2029)
}
func is_crlf(b []byte, i int) bool {
return b[i] == '\r' && b[i+1] == '\n'
}
// Check if the character is a line break or NUL.
func is_breakz(b []byte, i int) bool {
//return is_break(b, i) || is_z(b, i)
return ( // is_break:
b[i] == '\r' || // CR (#xD)
b[i] == '\n' || // LF (#xA)
b[i] == 0xC2 && b[i+1] == 0x85 || // NEL (#x85)
b[i] == 0xE2 && b[i+1] == 0x80 && b[i+2] == 0xA8 || // LS (#x2028)
b[i] == 0xE2 && b[i+1] == 0x80 && b[i+2] == 0xA9 || // PS (#x2029)
// is_z:
b[i] == 0)
}
// Check if the character is a line break, space, or NUL.
func is_spacez(b []byte, i int) bool {
//return is_space(b, i) || is_breakz(b, i)
return ( // is_space:
b[i] == ' ' ||
// is_breakz:
b[i] == '\r' || // CR (#xD)
b[i] == '\n' || // LF (#xA)
b[i] == 0xC2 && b[i+1] == 0x85 || // NEL (#x85)
b[i] == 0xE2 && b[i+1] == 0x80 && b[i+2] == 0xA8 || // LS (#x2028)
b[i] == 0xE2 && b[i+1] == 0x80 && b[i+2] == 0xA9 || // PS (#x2029)
b[i] == 0)
}
// Check if the character is a line break, space, tab, or NUL.
func is_blankz(b []byte, i int) bool {
//return is_blank(b, i) || is_breakz(b, i)
return ( // is_blank:
b[i] == ' ' || b[i] == '\t' ||
// is_breakz:
b[i] == '\r' || // CR (#xD)
b[i] == '\n' || // LF (#xA)
b[i] == 0xC2 && b[i+1] == 0x85 || // NEL (#x85)
b[i] == 0xE2 && b[i+1] == 0x80 && b[i+2] == 0xA8 || // LS (#x2028)
b[i] == 0xE2 && b[i+1] == 0x80 && b[i+2] == 0xA9 || // PS (#x2029)
b[i] == 0)
}
// Determine the width of the character.
func width(b byte) int {
// Don't replace these by a switch without first
// confirming that it is being inlined.
if b&0x80 == 0x00 {
return 1
}
if b&0xE0 == 0xC0 {
return 2
}
if b&0xF0 == 0xE0 {
return 3
}
if b&0xF8 == 0xF0 {
return 4
}
return 0
}

7
vendor/modules.txt vendored
View File

@ -1,7 +0,0 @@
# github.com/ulikunitz/xz v0.5.5
github.com/ulikunitz/xz
github.com/ulikunitz/xz/internal/xlog
github.com/ulikunitz/xz/lzma
github.com/ulikunitz/xz/internal/hash
# gopkg.in/yaml.v2 v2.2.2
gopkg.in/yaml.v2