Crypto++  8.8
Free C++ class library of cryptographic schemes
gzip.cpp
1 // gzip.cpp - originally written and placed in the public domain by Wei Dai
2 
3 #include "pch.h"
4 #include "gzip.h"
5 #include "argnames.h"
6 
7 NAMESPACE_BEGIN(CryptoPP)
8 
9 // Checks whether the character is valid for ISO/IEC 8859-1 as required by RFC 1952
10 static inline bool Is8859Character(char c) {
11  const unsigned char cc = static_cast<unsigned char>(c);
12  return (cc >= 32 && cc <= 126) || (cc >= 160);
13 }
14 
16 {
17  Deflator::IsolatedInitialize(parameters);
18 
20  if (parameters.GetValue(Name::FileName(), v))
21  m_filename.assign(reinterpret_cast<const char*>(v.begin()), v.size());
22  if (parameters.GetValue(Name::Comment(), v))
23  m_comment.assign(reinterpret_cast<const char*>(v.begin()), v.size());
24  m_filetime = parameters.GetIntValueWithDefault(Name::FileTime(), 0);
25 }
26 
27 void Gzip::WritePrestreamHeader()
28 {
29  m_totalLen = 0;
30  m_crc.Restart();
31 
32  int flags = 0;
33  if(!m_filename.empty())
34  flags |= FILENAME;
35  if(!m_comment.empty())
36  flags |= COMMENTS;
37 
38  AttachedTransformation()->Put(MAGIC1);
39  AttachedTransformation()->Put(MAGIC2);
40  AttachedTransformation()->Put(DEFLATED);
41  AttachedTransformation()->Put((byte)flags); // general flag
42  AttachedTransformation()->PutWord32(m_filetime, LITTLE_ENDIAN_ORDER); // time stamp
43  byte extra = static_cast<byte>((GetDeflateLevel() == 1) ?
44  FAST : ((GetDeflateLevel() == 9) ? SLOW : 0));
45  AttachedTransformation()->Put(extra);
46  AttachedTransformation()->Put(GZIP_OS_CODE);
47 
48  // Filename is NULL terminated, hence the +1
49  if(!m_filename.empty())
50  AttachedTransformation()->Put((const unsigned char*)m_filename.data(), m_filename.size() +1);
51 
52  // Comment is NULL terminated, hence the +1
53  if(!m_comment.empty())
54  AttachedTransformation()->Put((const unsigned char*)m_comment.data(), m_comment.size() +1);
55 }
56 
57 void Gzip::ProcessUncompressedData(const byte *inString, size_t length)
58 {
59  m_crc.Update(inString, length);
60  m_totalLen += (word32)length;
61 }
62 
63 void Gzip::WritePoststreamTail()
64 {
65  SecByteBlock crc(4);
66  m_crc.Final(crc);
67  AttachedTransformation()->Put(crc, 4);
69 
70  m_filetime = 0;
71  m_filename.clear();
72  m_comment.clear();
73 }
74 
75 void Gzip::SetComment(const std::string& comment, bool throwOnEncodingError)
76 {
77  if(throwOnEncodingError)
78  {
79  for(size_t i = 0; i < comment.length(); i++) {
80  const char c = comment[i];
81  if(!Is8859Character(c))
82  throw InvalidDataFormat("The comment is not ISO/IEC 8859-1 encoded");
83  }
84  }
85 
86  m_comment = comment;
87 }
88 
89 void Gzip::SetFilename(const std::string& filename, bool throwOnEncodingError)
90 {
91  if(throwOnEncodingError)
92  {
93  for(size_t i = 0; i < filename.length(); i++) {
94  const char c = filename[i];
95  if(!Is8859Character(c))
96  throw InvalidDataFormat("The filename is not ISO/IEC 8859-1 encoded");
97  }
98  }
99 
100  m_filename = filename;
101 }
102 
103 // *************************************************************
104 
105 Gunzip::Gunzip(BufferedTransformation *attachment, bool repeat, int propagation)
106  : Inflator(attachment, repeat, propagation), m_length(0), m_filetime(0)
107 {
108 }
109 
110 void Gunzip::ProcessPrestreamHeader()
111 {
112  m_length = 0;
113  m_crc.Restart();
114 
115  m_filetime = 0;
116  m_filename.clear();
117  m_comment.clear();
118 
119  byte buf[6];
120  byte b, flags;
121 
122  if (m_inQueue.Get(buf, 2)!=2) throw HeaderErr();
123  if (buf[0] != MAGIC1 || buf[1] != MAGIC2) throw HeaderErr();
124  if (!m_inQueue.Get(b) || (b != DEFLATED)) throw HeaderErr(); // skip CM flag
125  if (!m_inQueue.Get(flags)) throw HeaderErr();
126  if (flags & (ENCRYPTED | CONTINUED)) throw HeaderErr();
127  if (m_inQueue.GetWord32(m_filetime, LITTLE_ENDIAN_ORDER) != 4) throw HeaderErr();
128  if (m_inQueue.Skip(2)!=2) throw HeaderErr(); // Skip extra flags and OS type
129 
130  if (flags & EXTRA_FIELDS) // skip extra fields
131  {
132  word16 length;
133  if (m_inQueue.GetWord16(length, LITTLE_ENDIAN_ORDER) != 2) throw HeaderErr();
134  if (m_inQueue.Skip(length)!=length) throw HeaderErr();
135  }
136 
137  if (flags & FILENAME) // extract filename
138  {
139  do
140  {
141  if(!m_inQueue.Get(b)) throw HeaderErr();
142  if(b) m_filename.append( 1, (char)b );
143  }
144  while (b);
145  }
146 
147  if (flags & COMMENTS) // extract comments
148  {
149  do
150  {
151  if(!m_inQueue.Get(b)) throw HeaderErr();
152  if(b) m_comment.append( 1, (char)b );
153  }
154  while (b);
155  }
156 }
157 
158 void Gunzip::ProcessDecompressedData(const byte *inString, size_t length)
159 {
160  AttachedTransformation()->Put(inString, length);
161  m_crc.Update(inString, length);
162  m_length += (word32)length;
163 }
164 
165 void Gunzip::ProcessPoststreamTail()
166 {
167  SecByteBlock crc(4);
168  if (m_inQueue.Get(crc, 4) != 4)
169  throw TailErr();
170  if (!m_crc.Verify(crc))
171  throw CrcErr();
172 
173  word32 lengthCheck;
174  if (m_inQueue.GetWord32(lengthCheck, LITTLE_ENDIAN_ORDER) != 4)
175  throw TailErr();
176  if (lengthCheck != m_length)
177  throw LengthErr();
178 }
179 
180 const std::string& Gunzip::GetComment(bool throwOnEncodingError) const
181 {
182  if(throwOnEncodingError)
183  {
184  for(size_t i = 0; i < m_comment.length(); i++) {
185  const char c = m_comment[i];
186  if(!Is8859Character(c))
187  throw InvalidDataFormat("The comment is not ISO/IEC 8859-1 encoded");
188  }
189  }
190 
191  return m_comment;
192 }
193 
194 const std::string& Gunzip::GetFilename(bool throwOnEncodingError) const
195 {
196  if(throwOnEncodingError)
197  {
198  for(size_t i = 0; i < m_filename.length(); i++) {
199  const char c = m_filename[i];
200  if(!Is8859Character(c))
201  throw InvalidDataFormat("The filename is not ISO/IEC 8859-1 encoded");
202  }
203  }
204 
205  return m_filename;
206 }
207 
208 NAMESPACE_END
Standard names for retrieving values by name when working with NameValuePairs.
Interface for buffered transformations.
Definition: cryptlib.h:1657
size_t Put(byte inByte, bool blocking=true)
Input a byte for processing.
Definition: cryptlib.h:1678
size_t PutWord32(word32 value, ByteOrder order=BIG_ENDIAN_ORDER, bool blocking=true)
Input a 32-bit word for processing.
void Update(const byte *input, size_t length)
Updates a hash with additional input.
Definition: crc.cpp:147
Used to pass byte array input as part of a NameValuePairs object.
Definition: algparam.h:25
size_t size() const
Length of the memory block.
Definition: algparam.h:88
const byte * begin() const
Pointer to the first byte in the memory block.
Definition: algparam.h:84
void IsolatedInitialize(const NameValuePairs &parameters)
Initialize or reinitialize this object, without signal propagation.
Definition: zdeflate.cpp:265
int GetDeflateLevel() const
Retrieves the deflation level.
Definition: zdeflate.h:118
BufferedTransformation * AttachedTransformation()
Retrieve attached transformation.
Gunzip(BufferedTransformation *attachment=NULL, bool repeat=false, int autoSignalPropagation=-1)
Construct a Gunzip decompressor.
Definition: gzip.cpp:105
const std::string & GetComment(bool throwOnEncodingError=false) const
Definition: gzip.cpp:180
const std::string & GetFilename(bool throwOnEncodingError=false) const
Definition: gzip.cpp:194
void SetComment(const std::string &comment, bool throwOnEncodingError=false)
Definition: gzip.cpp:75
void IsolatedInitialize(const NameValuePairs &parameters)
Initialize or reinitialize this object, without signal propagation.
Definition: gzip.cpp:15
void SetFilename(const std::string &filename, bool throwOnEncodingError=false)
Definition: gzip.cpp:89
virtual bool Verify(const byte *digest)
Verifies the hash of the current message.
Definition: cryptlib.h:1205
virtual void Restart()
Restart the hash.
Definition: cryptlib.h:1152
virtual void Final(byte *digest)
Computes the hash of the current message.
Definition: cryptlib.h:1147
DEFLATE decompressor (RFC 1951)
Definition: zinflate.h:95
Input data was received that did not conform to expected format.
Definition: cryptlib.h:218
Interface for retrieving values given their names.
Definition: cryptlib.h:327
bool GetValue(const char *name, T &value) const
Get a named value.
Definition: cryptlib.h:384
CRYPTOPP_DLL int GetIntValueWithDefault(const char *name, int defaultValue) const
Get a named value with type int, with default.
Definition: cryptlib.h:429
SecBlock<byte> typedef.
Definition: secblock.h:1226
unsigned int word32
32-bit unsigned datatype
Definition: config_int.h:72
unsigned short word16
16-bit unsigned datatype
Definition: config_int.h:69
@ LITTLE_ENDIAN_ORDER
byte order is little-endian
Definition: cryptlib.h:150
GZIP compression and decompression (RFC 1952)
Crypto++ library namespace.
const char * FileTime()
int
Definition: argnames.h:92
const char * FileName()
const char *
Definition: argnames.h:91
const char * Comment()
const char *
Definition: argnames.h:93
Precompiled header file.