1 /*
2 ** 2002 April 25
3 **
4 ** The author disclaims copyright to this source code. In place of
5 ** a legal notice, here is a blessing:
6 **
7 ** May you do good and not evil.
8 ** May you find forgiveness for yourself and forgive others.
9 ** May you share freely, never taking more than you give.
10 **
11 *************************************************************************
12 ** This file contains helper routines used to translate binary data into
13 ** a null-terminated string (suitable for use in SQLite) and back again.
14 ** These are convenience routines for use by people who want to store binary
15 ** data in an SQLite database. The code in this file is not used by any other
16 ** part of the SQLite library.
17 **
18 ** $Id: encode.c,v 1.12 2004/03/17 18:44:46 drh Exp $
19 */
20 #include <string.h>
21 #include <assert.h>
22
23 /*
24 ** How This Encoder Works
25 **
26 ** The output is allowed to contain any character except 0x27 (') and
27 ** 0x00. This is accomplished by using an escape character to encode
28 ** 0x27 and 0x00 as a two-byte sequence. The escape character is always
29 ** 0x01. An 0x00 is encoded as the two byte sequence 0x01 0x01. The
30 ** 0x27 character is encoded as the two byte sequence 0x01 0x28. Finally,
31 ** the escape character itself is encoded as the two-character sequence
32 ** 0x01 0x02.
33 **
34 ** To summarize, the encoder works by using an escape sequences as follows:
35 **
36 ** 0x00 -> 0x01 0x01
37 ** 0x01 -> 0x01 0x02
38 ** 0x27 -> 0x01 0x28
39 **
40 ** If that were all the encoder did, it would work, but in certain cases
41 ** it could double the size of the encoded string. For example, to
42 ** encode a string of 100 0x27 characters would require 100 instances of
43 ** the 0x01 0x03 escape sequence resulting in a 200-character output.
44 ** We would prefer to keep the size of the encoded string smaller than
45 ** this.
46 **
47 ** To minimize the encoding size, we first add a fixed offset value to each
48 ** byte in the sequence. The addition is modulo 256. (That is to say, if
49 ** the sum of the original character value and the offset exceeds 256, then
50 ** the higher order bits are truncated.) The offset is chosen to minimize
51 ** the number of characters in the string that need to be escaped. For
52 ** example, in the case above where the string was composed of 100 0x27
53 ** characters, the offset might be 0x01. Each of the 0x27 characters would
54 ** then be converted into an 0x28 character which would not need to be
55 ** escaped at all and so the 100 character input string would be converted
56 ** into just 100 characters of output. Actually 101 characters of output -
57 ** we have to record the offset used as the first byte in the sequence so
58 ** that the string can be decoded. Since the offset value is stored as
59 ** part of the output string and the output string is not allowed to contain
60 ** characters 0x00 or 0x27, the offset cannot be 0x00 or 0x27.
61 **
62 ** Here, then, are the encoding steps:
63 **
64 ** (1) Choose an offset value and make it the first character of
65 ** output.
66 **
67 ** (2) Copy each input character into the output buffer, one by
68 ** one, adding the offset value as you copy.
69 **
70 ** (3) If the value of an input character plus offset is 0x00, replace
71 ** that one character by the two-character sequence 0x01 0x01.
72 ** If the sum is 0x01, replace it with 0x01 0x02. If the sum
73 ** is 0x27, replace it with 0x01 0x03.
74 **
75 ** (4) Put a 0x00 terminator at the end of the output.
76 **
77 ** Decoding is obvious:
78 **
79 ** (5) Copy encoded characters except the first into the decode
80 ** buffer. Set the first encoded character aside for use as
81 ** the offset in step 7 below.
82 **
83 ** (6) Convert each 0x01 0x01 sequence into a single character 0x00.
84 ** Convert 0x01 0x02 into 0x01. Convert 0x01 0x28 into 0x27.
85 **
86 ** (7) Subtract the offset value that was the first character of
87 ** the encoded buffer from all characters in the output buffer.
88 **
89 ** The only tricky part is step (1) - how to compute an offset value to
90 ** minimize the size of the output buffer. This is accomplished by testing
91 ** all offset values and picking the one that results in the fewest number
92 ** of escapes. To do that, we first scan the entire input and count the
93 ** number of occurances of each character value in the input. Suppose
94 ** the number of 0x00 characters is N(0), the number of occurances of 0x01
95 ** is N(1), and so forth up to the number of occurances of 0xff is N(255).
96 ** An offset of 0 is not allowed so we don't have to test it. The number
97 ** of escapes required for an offset of 1 is N(1)+N(2)+N(40). The number
98 ** of escapes required for an offset of 2 is N(2)+N(3)+N(41). And so forth.
99 ** In this way we find the offset that gives the minimum number of escapes,
100 ** and thus minimizes the length of the output string.
101 */
102
103 /*
104 ** Encode a binary buffer "in" of size n bytes so that it contains
105 ** no instances of characters '\'' or '\000'. The output is
106 ** null-terminated and can be used as a string value in an INSERT
107 ** or UPDATE statement. Use sqlite_decode_binary() to convert the
108 ** string back into its original binary.
109 **
110 ** The result is written into a preallocated output buffer "out".
111 ** "out" must be able to hold at least 2 +(257*n)/254 bytes.
112 ** In other words, the output will be expanded by as much as 3
113 ** bytes for every 254 bytes of input plus 2 bytes of fixed overhead.
114 ** (This is approximately 2 + 1.0118*n or about a 1.2% size increase.)
115 **
116 ** The return value is the number of characters in the encoded
117 ** string, excluding the "\000" terminator.
118 **
119 ** If out==NULL then no output is generated but the routine still returns
120 ** the number of characters that would have been generated if out had
121 ** not been NULL.
122 */
sqlite_encode_binary(const unsigned char * in,int n,unsigned char * out)123 int sqlite_encode_binary(const unsigned char *in, int n, unsigned char *out){
124 int i, j, e, m;
125 unsigned char x;
126 int cnt[256];
127 if( n<=0 ){
128 if( out ){
129 out[0] = 'x';
130 out[1] = 0;
131 }
132 return 1;
133 }
134 memset(cnt, 0, sizeof(cnt));
135 for(i=n-1; i>=0; i--){ cnt[in[i]]++; }
136 m = n;
137 for(i=1; i<256; i++){
138 int sum;
139 if( i=='\'' ) continue;
140 sum = cnt[i] + cnt[(i+1)&0xff] + cnt[(i+'\'')&0xff];
141 if( sum<m ){
142 m = sum;
143 e = i;
144 if( m==0 ) break;
145 }
146 }
147 if( out==0 ){
148 return n+m+1;
149 }
150 out[0] = e;
151 j = 1;
152 for(i=0; i<n; i++){
153 x = in[i] - e;
154 if( x==0 || x==1 || x=='\''){
155 out[j++] = 1;
156 x++;
157 }
158 out[j++] = x;
159 }
160 out[j] = 0;
161 assert( j==n+m+1 );
162 return j;
163 }
164
165 /*
166 ** Decode the string "in" into binary data and write it into "out".
167 ** This routine reverses the encoding created by sqlite_encode_binary().
168 ** The output will always be a few bytes less than the input. The number
169 ** of bytes of output is returned. If the input is not a well-formed
170 ** encoding, -1 is returned.
171 **
172 ** The "in" and "out" parameters may point to the same buffer in order
173 ** to decode a string in place.
174 */
sqlite_decode_binary(const unsigned char * in,unsigned char * out)175 int sqlite_decode_binary(const unsigned char *in, unsigned char *out){
176 int i, e;
177 unsigned char c;
178 e = *(in++);
179 i = 0;
180 while( (c = *(in++))!=0 ){
181 if( c==1 ){
182 c = *(in++) - 1;
183 }
184 out[i++] = c + e;
185 }
186 return i;
187 }
188
189 #ifdef ENCODER_TEST
190 #include <stdio.h>
191 /*
192 ** The subroutines above are not tested by the usual test suite. To test
193 ** these routines, compile just this one file with a -DENCODER_TEST=1 option
194 ** and run the result.
195 */
main(int argc,char ** argv)196 int main(int argc, char **argv){
197 int i, j, n, m, nOut, nByteIn, nByteOut;
198 unsigned char in[30000];
199 unsigned char out[33000];
200
201 nByteIn = nByteOut = 0;
202 for(i=0; i<sizeof(in); i++){
203 printf("Test %d: ", i+1);
204 n = rand() % (i+1);
205 if( i%100==0 ){
206 int k;
207 for(j=k=0; j<n; j++){
208 /* if( k==0 || k=='\'' ) k++; */
209 in[j] = k;
210 k = (k+1)&0xff;
211 }
212 }else{
213 for(j=0; j<n; j++) in[j] = rand() & 0xff;
214 }
215 nByteIn += n;
216 nOut = sqlite_encode_binary(in, n, out);
217 nByteOut += nOut;
218 if( nOut!=strlen(out) ){
219 printf(" ERROR return value is %d instead of %d\n", nOut, strlen(out));
220 exit(1);
221 }
222 if( nOut!=sqlite_encode_binary(in, n, 0) ){
223 printf(" ERROR actual output size disagrees with predicted size\n");
224 exit(1);
225 }
226 m = (256*n + 1262)/253;
227 printf("size %d->%d (max %d)", n, strlen(out)+1, m);
228 if( strlen(out)+1>m ){
229 printf(" ERROR output too big\n");
230 exit(1);
231 }
232 for(j=0; out[j]; j++){
233 if( out[j]=='\'' ){
234 printf(" ERROR contains (')\n");
235 exit(1);
236 }
237 }
238 j = sqlite_decode_binary(out, out);
239 if( j!=n ){
240 printf(" ERROR decode size %d\n", j);
241 exit(1);
242 }
243 if( memcmp(in, out, n)!=0 ){
244 printf(" ERROR decode mismatch\n");
245 exit(1);
246 }
247 printf(" OK\n");
248 }
249 fprintf(stderr,"Finished. Total encoding: %d->%d bytes\n",
250 nByteIn, nByteOut);
251 fprintf(stderr,"Avg size increase: %.3f%%\n",
252 (nByteOut-nByteIn)*100.0/(double)nByteIn);
253 }
254 #endif /* ENCODER_TEST */
255