source: python/trunk/Modules/binascii.c@ 389

Last change on this file since 389 was 2, checked in by Yuri Dario, 15 years ago

Initial import for vendor code.
  • Property svn:eol-style set to native
File size: 42.5 KB
Line 
1/*
2** Routines to represent binary data in ASCII and vice-versa
3**
4** This module currently supports the following encodings:
5** uuencode:
6** each line encodes 45 bytes (except possibly the last)
7** First char encodes (binary) length, rest data
8** each char encodes 6 bits, as follows:
9** binary: 01234567 abcdefgh ijklmnop
10** ascii: 012345 67abcd efghij klmnop
11** ASCII encoding method is "excess-space": 000000 is encoded as ' ', etc.
12** short binary data is zero-extended (so the bits are always in the
13** right place), this does *not* reflect in the length.
14** base64:
15** Line breaks are insignificant, but lines are at most 76 chars
16** each char encodes 6 bits, in similar order as uucode/hqx. Encoding
17** is done via a table.
18** Short binary data is filled (in ASCII) with '='.
19** hqx:
20** File starts with introductory text, real data starts and ends
21** with colons.
22** Data consists of three similar parts: info, datafork, resourcefork.
23** Each part is protected (at the end) with a 16-bit crc
24** The binary data is run-length encoded, and then ascii-fied:
25** binary: 01234567 abcdefgh ijklmnop
26** ascii: 012345 67abcd efghij klmnop
27** ASCII encoding is table-driven, see the code.
28** Short binary data results in the runt ascii-byte being output with
29** the bits in the right place.
30**
31** While I was reading dozens of programs that encode or decode the formats
32** here (documentation? hihi:-) I have formulated Jansen's Observation:
33**
34** Programs that encode binary data in ASCII are written in
35** such a style that they are as unreadable as possible. Devices used
36** include unnecessary global variables, burying important tables
37** in unrelated sourcefiles, putting functions in include files,
38** using seemingly-descriptive variable names for different purposes,
39** calls to empty subroutines and a host of others.
40**
41** I have attempted to break with this tradition, but I guess that that
42** does make the performance sub-optimal. Oh well, too bad...
43**
44** Jack Jansen, CWI, July 1995.
45**
46** Added support for quoted-printable encoding, based on rfc 1521 et al
47** quoted-printable encoding specifies that non printable characters (anything
48** below 32 and above 126) be encoded as =XX where XX is the hexadecimal value
49** of the character. It also specifies some other behavior to enable 8bit data
50** in a mail message with little difficulty (maximum line sizes, protecting
51** some cases of whitespace, etc).
52**
53** Brandon Long, September 2001.
54*/
55
56#define PY_SSIZE_T_CLEAN
57
58#include "Python.h"
59#ifdef USE_ZLIB_CRC32
60#include "zlib.h"
61#endif
62
63static PyObject *Error;
64static PyObject *Incomplete;
65
66/*
67** hqx lookup table, ascii->binary.
68*/
69
70#define RUNCHAR 0x90
71
72#define DONE 0x7F
73#define SKIP 0x7E
74#define FAIL 0x7D
75
76static unsigned char table_a2b_hqx[256] = {
77/* ^@ ^A ^B ^C ^D ^E ^F ^G */
78/* 0*/ FAIL, FAIL, FAIL, FAIL, FAIL, FAIL, FAIL, FAIL,
79/* \b \t \n ^K ^L \r ^N ^O */
80/* 1*/ FAIL, FAIL, SKIP, FAIL, FAIL, SKIP, FAIL, FAIL,
81/* ^P ^Q ^R ^S ^T ^U ^V ^W */
82/* 2*/ FAIL, FAIL, FAIL, FAIL, FAIL, FAIL, FAIL, FAIL,
83/* ^X ^Y ^Z ^[ ^\ ^] ^^ ^_ */
84/* 3*/ FAIL, FAIL, FAIL, FAIL, FAIL, FAIL, FAIL, FAIL,
85/* ! " # $ % & ' */
86/* 4*/ FAIL, 0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06,
87/* ( ) * + , - . / */
88/* 5*/ 0x07, 0x08, 0x09, 0x0A, 0x0B, 0x0C, FAIL, FAIL,
89/* 0 1 2 3 4 5 6 7 */
90/* 6*/ 0x0D, 0x0E, 0x0F, 0x10, 0x11, 0x12, 0x13, FAIL,
91/* 8 9 : ; < = > ? */
92/* 7*/ 0x14, 0x15, DONE, FAIL, FAIL, FAIL, FAIL, FAIL,
93/* @ A B C D E F G */
94/* 8*/ 0x16, 0x17, 0x18, 0x19, 0x1A, 0x1B, 0x1C, 0x1D,
95/* H I J K L M N O */
96/* 9*/ 0x1E, 0x1F, 0x20, 0x21, 0x22, 0x23, 0x24, FAIL,
97/* P Q R S T U V W */
98/*10*/ 0x25, 0x26, 0x27, 0x28, 0x29, 0x2A, 0x2B, FAIL,
99/* X Y Z [ \ ] ^ _ */
100/*11*/ 0x2C, 0x2D, 0x2E, 0x2F, FAIL, FAIL, FAIL, FAIL,
101/* ` a b c d e f g */
102/*12*/ 0x30, 0x31, 0x32, 0x33, 0x34, 0x35, 0x36, FAIL,
103/* h i j k l m n o */
104/*13*/ 0x37, 0x38, 0x39, 0x3A, 0x3B, 0x3C, FAIL, FAIL,
105/* p q r s t u v w */
106/*14*/ 0x3D, 0x3E, 0x3F, FAIL, FAIL, FAIL, FAIL, FAIL,
107/* x y z { | } ~ ^? */
108/*15*/ FAIL, FAIL, FAIL, FAIL, FAIL, FAIL, FAIL, FAIL,
109/*16*/ FAIL, FAIL, FAIL, FAIL, FAIL, FAIL, FAIL, FAIL,
110 FAIL, FAIL, FAIL, FAIL, FAIL, FAIL, FAIL, FAIL,
111 FAIL, FAIL, FAIL, FAIL, FAIL, FAIL, FAIL, FAIL,
112 FAIL, FAIL, FAIL, FAIL, FAIL, FAIL, FAIL, FAIL,
113 FAIL, FAIL, FAIL, FAIL, FAIL, FAIL, FAIL, FAIL,
114 FAIL, FAIL, FAIL, FAIL, FAIL, FAIL, FAIL, FAIL,
115 FAIL, FAIL, FAIL, FAIL, FAIL, FAIL, FAIL, FAIL,
116 FAIL, FAIL, FAIL, FAIL, FAIL, FAIL, FAIL, FAIL,
117 FAIL, FAIL, FAIL, FAIL, FAIL, FAIL, FAIL, FAIL,
118 FAIL, FAIL, FAIL, FAIL, FAIL, FAIL, FAIL, FAIL,
119 FAIL, FAIL, FAIL, FAIL, FAIL, FAIL, FAIL, FAIL,
120 FAIL, FAIL, FAIL, FAIL, FAIL, FAIL, FAIL, FAIL,
121 FAIL, FAIL, FAIL, FAIL, FAIL, FAIL, FAIL, FAIL,
122 FAIL, FAIL, FAIL, FAIL, FAIL, FAIL, FAIL, FAIL,
123 FAIL, FAIL, FAIL, FAIL, FAIL, FAIL, FAIL, FAIL,
124 FAIL, FAIL, FAIL, FAIL, FAIL, FAIL, FAIL, FAIL,
125};
126
127static unsigned char table_b2a_hqx[] =
128"!\"#$%&'()*+,-012345689@ABCDEFGHIJKLMNPQRSTUVXYZ[`abcdefhijklmpqr";
129
130static char table_a2b_base64[] = {
131 -1,-1,-1,-1, -1,-1,-1,-1, -1,-1,-1,-1, -1,-1,-1,-1,
132 -1,-1,-1,-1, -1,-1,-1,-1, -1,-1,-1,-1, -1,-1,-1,-1,
133 -1,-1,-1,-1, -1,-1,-1,-1, -1,-1,-1,62, -1,-1,-1,63,
134 52,53,54,55, 56,57,58,59, 60,61,-1,-1, -1, 0,-1,-1, /* Note PAD->0 */
135 -1, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9,10, 11,12,13,14,
136 15,16,17,18, 19,20,21,22, 23,24,25,-1, -1,-1,-1,-1,
137 -1,26,27,28, 29,30,31,32, 33,34,35,36, 37,38,39,40,
138 41,42,43,44, 45,46,47,48, 49,50,51,-1, -1,-1,-1,-1
139};
140
141#define BASE64_PAD '='
142
143/* Max binary chunk size; limited only by available memory */
144#define BASE64_MAXBIN (PY_SSIZE_T_MAX/2 - sizeof(PyStringObject) - 3)
145
146static unsigned char table_b2a_base64[] =
147"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/";
148
149
150
151static unsigned short crctab_hqx[256] = {
152 0x0000, 0x1021, 0x2042, 0x3063, 0x4084, 0x50a5, 0x60c6, 0x70e7,
153 0x8108, 0x9129, 0xa14a, 0xb16b, 0xc18c, 0xd1ad, 0xe1ce, 0xf1ef,
154 0x1231, 0x0210, 0x3273, 0x2252, 0x52b5, 0x4294, 0x72f7, 0x62d6,
155 0x9339, 0x8318, 0xb37b, 0xa35a, 0xd3bd, 0xc39c, 0xf3ff, 0xe3de,
156 0x2462, 0x3443, 0x0420, 0x1401, 0x64e6, 0x74c7, 0x44a4, 0x5485,
157 0xa56a, 0xb54b, 0x8528, 0x9509, 0xe5ee, 0xf5cf, 0xc5ac, 0xd58d,
158 0x3653, 0x2672, 0x1611, 0x0630, 0x76d7, 0x66f6, 0x5695, 0x46b4,
159 0xb75b, 0xa77a, 0x9719, 0x8738, 0xf7df, 0xe7fe, 0xd79d, 0xc7bc,
160 0x48c4, 0x58e5, 0x6886, 0x78a7, 0x0840, 0x1861, 0x2802, 0x3823,
161 0xc9cc, 0xd9ed, 0xe98e, 0xf9af, 0x8948, 0x9969, 0xa90a, 0xb92b,
162 0x5af5, 0x4ad4, 0x7ab7, 0x6a96, 0x1a71, 0x0a50, 0x3a33, 0x2a12,
163 0xdbfd, 0xcbdc, 0xfbbf, 0xeb9e, 0x9b79, 0x8b58, 0xbb3b, 0xab1a,
164 0x6ca6, 0x7c87, 0x4ce4, 0x5cc5, 0x2c22, 0x3c03, 0x0c60, 0x1c41,
165 0xedae, 0xfd8f, 0xcdec, 0xddcd, 0xad2a, 0xbd0b, 0x8d68, 0x9d49,
166 0x7e97, 0x6eb6, 0x5ed5, 0x4ef4, 0x3e13, 0x2e32, 0x1e51, 0x0e70,
167 0xff9f, 0xefbe, 0xdfdd, 0xcffc, 0xbf1b, 0xaf3a, 0x9f59, 0x8f78,
168 0x9188, 0x81a9, 0xb1ca, 0xa1eb, 0xd10c, 0xc12d, 0xf14e, 0xe16f,
169 0x1080, 0x00a1, 0x30c2, 0x20e3, 0x5004, 0x4025, 0x7046, 0x6067,
170 0x83b9, 0x9398, 0xa3fb, 0xb3da, 0xc33d, 0xd31c, 0xe37f, 0xf35e,
171 0x02b1, 0x1290, 0x22f3, 0x32d2, 0x4235, 0x5214, 0x6277, 0x7256,
172 0xb5ea, 0xa5cb, 0x95a8, 0x8589, 0xf56e, 0xe54f, 0xd52c, 0xc50d,
173 0x34e2, 0x24c3, 0x14a0, 0x0481, 0x7466, 0x6447, 0x5424, 0x4405,
174 0xa7db, 0xb7fa, 0x8799, 0x97b8, 0xe75f, 0xf77e, 0xc71d, 0xd73c,
175 0x26d3, 0x36f2, 0x0691, 0x16b0, 0x6657, 0x7676, 0x4615, 0x5634,
176 0xd94c, 0xc96d, 0xf90e, 0xe92f, 0x99c8, 0x89e9, 0xb98a, 0xa9ab,
177 0x5844, 0x4865, 0x7806, 0x6827, 0x18c0, 0x08e1, 0x3882, 0x28a3,
178 0xcb7d, 0xdb5c, 0xeb3f, 0xfb1e, 0x8bf9, 0x9bd8, 0xabbb, 0xbb9a,
179 0x4a75, 0x5a54, 0x6a37, 0x7a16, 0x0af1, 0x1ad0, 0x2ab3, 0x3a92,
180 0xfd2e, 0xed0f, 0xdd6c, 0xcd4d, 0xbdaa, 0xad8b, 0x9de8, 0x8dc9,
181 0x7c26, 0x6c07, 0x5c64, 0x4c45, 0x3ca2, 0x2c83, 0x1ce0, 0x0cc1,
182 0xef1f, 0xff3e, 0xcf5d, 0xdf7c, 0xaf9b, 0xbfba, 0x8fd9, 0x9ff8,
183 0x6e17, 0x7e36, 0x4e55, 0x5e74, 0x2e93, 0x3eb2, 0x0ed1, 0x1ef0,
184};
185
186PyDoc_STRVAR(doc_a2b_uu, "(ascii) -> bin. Decode a line of uuencoded data");
187
188static PyObject *
189binascii_a2b_uu(PyObject *self, PyObject *args)
190{
191 unsigned char *ascii_data, *bin_data;
192 int leftbits = 0;
193 unsigned char this_ch;
194 unsigned int leftchar = 0;
195 PyObject *rv;
196 Py_ssize_t ascii_len, bin_len;
197
198 if ( !PyArg_ParseTuple(args, "t#:a2b_uu", &ascii_data, &ascii_len) )
199 return NULL;
200
201 assert(ascii_len >= 0);
202
203 /* First byte: binary data length (in bytes) */
204 bin_len = (*ascii_data++ - ' ') & 077;
205 ascii_len--;
206
207 /* Allocate the buffer */
208 if ( (rv=PyString_FromStringAndSize(NULL, bin_len)) == NULL )
209 return NULL;
210 bin_data = (unsigned char *)PyString_AsString(rv);
211
212 for( ; bin_len > 0 ; ascii_len--, ascii_data++ ) {
213 /* XXX is it really best to add NULs if there's no more data */
214 this_ch = (ascii_len > 0) ? *ascii_data : 0;
215 if ( this_ch == '\n' || this_ch == '\r' || ascii_len <= 0) {
216 /*
217 ** Whitespace. Assume some spaces got eaten at
218 ** end-of-line. (We check this later)
219 */
220 this_ch = 0;
221 } else {
222 /* Check the character for legality
223 ** The 64 in stead of the expected 63 is because
224 ** there are a few uuencodes out there that use
225 ** '`' as zero instead of space.
226 */
227 if ( this_ch < ' ' || this_ch > (' ' + 64)) {
228 PyErr_SetString(Error, "Illegal char");
229 Py_DECREF(rv);
230 return NULL;
231 }
232 this_ch = (this_ch - ' ') & 077;
233 }
234 /*
235 ** Shift it in on the low end, and see if there's
236 ** a byte ready for output.
237 */
238 leftchar = (leftchar << 6) | (this_ch);
239 leftbits += 6;
240 if ( leftbits >= 8 ) {
241 leftbits -= 8;
242 *bin_data++ = (leftchar >> leftbits) & 0xff;
243 leftchar &= ((1 << leftbits) - 1);
244 bin_len--;
245 }
246 }
247 /*
248 ** Finally, check that if there's anything left on the line
249 ** that it's whitespace only.
250 */
251 while( ascii_len-- > 0 ) {
252 this_ch = *ascii_data++;
253 /* Extra '`' may be written as padding in some cases */
254 if ( this_ch != ' ' && this_ch != ' '+64 &&
255 this_ch != '\n' && this_ch != '\r' ) {
256 PyErr_SetString(Error, "Trailing garbage");
257 Py_DECREF(rv);
258 return NULL;
259 }
260 }
261 return rv;
262}
263
264PyDoc_STRVAR(doc_b2a_uu, "(bin) -> ascii. Uuencode line of data");
265
266static PyObject *
267binascii_b2a_uu(PyObject *self, PyObject *args)
268{
269 unsigned char *ascii_data, *bin_data;
270 int leftbits = 0;
271 unsigned char this_ch;
272 unsigned int leftchar = 0;
273 PyObject *rv;
274 Py_ssize_t bin_len;
275
276 if ( !PyArg_ParseTuple(args, "s#:b2a_uu", &bin_data, &bin_len) )
277 return NULL;
278 if ( bin_len > 45 ) {
279 /* The 45 is a limit that appears in all uuencode's */
280 PyErr_SetString(Error, "At most 45 bytes at once");
281 return NULL;
282 }
283
284 /* We're lazy and allocate to much (fixed up later) */
285 if ( (rv=PyString_FromStringAndSize(NULL, 2 + (bin_len+2)/3*4)) == NULL )
286 return NULL;
287 ascii_data = (unsigned char *)PyString_AsString(rv);
288
289 /* Store the length */
290 *ascii_data++ = ' ' + (bin_len & 077);
291
292 for( ; bin_len > 0 || leftbits != 0 ; bin_len--, bin_data++ ) {
293 /* Shift the data (or padding) into our buffer */
294 if ( bin_len > 0 ) /* Data */
295 leftchar = (leftchar << 8) | *bin_data;
296 else /* Padding */
297 leftchar <<= 8;
298 leftbits += 8;
299
300 /* See if there are 6-bit groups ready */
301 while ( leftbits >= 6 ) {
302 this_ch = (leftchar >> (leftbits-6)) & 0x3f;
303 leftbits -= 6;
304 *ascii_data++ = this_ch + ' ';
305 }
306 }
307 *ascii_data++ = '\n'; /* Append a courtesy newline */
308
309 _PyString_Resize(&rv, (ascii_data -
310 (unsigned char *)PyString_AsString(rv)));
311 return rv;
312}
313
314
315static int
316binascii_find_valid(unsigned char *s, Py_ssize_t slen, int num)
317{
318 /* Finds & returns the (num+1)th
319 ** valid character for base64, or -1 if none.
320 */
321
322 int ret = -1;
323 unsigned char c, b64val;
324
325 while ((slen > 0) && (ret == -1)) {
326 c = *s;
327 b64val = table_a2b_base64[c & 0x7f];
328 if ( ((c <= 0x7f) && (b64val != (unsigned char)-1)) ) {
329 if (num == 0)
330 ret = *s;
331 num--;
332 }
333
334 s++;
335 slen--;
336 }
337 return ret;
338}
339
340PyDoc_STRVAR(doc_a2b_base64, "(ascii) -> bin. Decode a line of base64 data");
341
342static PyObject *
343binascii_a2b_base64(PyObject *self, PyObject *args)
344{
345 unsigned char *ascii_data, *bin_data;
346 int leftbits = 0;
347 unsigned char this_ch;
348 unsigned int leftchar = 0;
349 PyObject *rv;
350 Py_ssize_t ascii_len, bin_len;
351 int quad_pos = 0;
352
353 if ( !PyArg_ParseTuple(args, "t#:a2b_base64", &ascii_data, &ascii_len) )
354 return NULL;
355
356 assert(ascii_len >= 0);
357
358 if (ascii_len > PY_SSIZE_T_MAX - 3)
359 return PyErr_NoMemory();
360
361 bin_len = ((ascii_len+3)/4)*3; /* Upper bound, corrected later */
362
363 /* Allocate the buffer */
364 if ( (rv=PyString_FromStringAndSize(NULL, bin_len)) == NULL )
365 return NULL;
366 bin_data = (unsigned char *)PyString_AsString(rv);
367 bin_len = 0;
368
369 for( ; ascii_len > 0; ascii_len--, ascii_data++) {
370 this_ch = *ascii_data;
371
372 if (this_ch > 0x7f ||
373 this_ch == '\r' || this_ch == '\n' || this_ch == ' ')
374 continue;
375
376 /* Check for pad sequences and ignore
377 ** the invalid ones.
378 */
379 if (this_ch == BASE64_PAD) {
380 if ( (quad_pos < 2) ||
381 ((quad_pos == 2) &&
382 (binascii_find_valid(ascii_data, ascii_len, 1)
383 != BASE64_PAD)) )
384 {
385 continue;
386 }
387 else {
388 /* A pad sequence means no more input.
389 ** We've already interpreted the data
390 ** from the quad at this point.
391 */
392 leftbits = 0;
393 break;
394 }
395 }
396
397 this_ch = table_a2b_base64[*ascii_data];
398 if ( this_ch == (unsigned char) -1 )
399 continue;
400
401 /*
402 ** Shift it in on the low end, and see if there's
403 ** a byte ready for output.
404 */
405 quad_pos = (quad_pos + 1) & 0x03;
406 leftchar = (leftchar << 6) | (this_ch);
407 leftbits += 6;
408
409 if ( leftbits >= 8 ) {
410 leftbits -= 8;
411 *bin_data++ = (leftchar >> leftbits) & 0xff;
412 bin_len++;
413 leftchar &= ((1 << leftbits) - 1);
414 }
415 }
416
417 if (leftbits != 0) {
418 PyErr_SetString(Error, "Incorrect padding");
419 Py_DECREF(rv);
420 return NULL;
421 }
422
423 /* And set string size correctly. If the result string is empty
424 ** (because the input was all invalid) return the shared empty
425 ** string instead; _PyString_Resize() won't do this for us.
426 */
427 if (bin_len > 0)
428 _PyString_Resize(&rv, bin_len);
429 else {
430 Py_DECREF(rv);
431 rv = PyString_FromString("");
432 }
433 return rv;
434}
435
436PyDoc_STRVAR(doc_b2a_base64, "(bin) -> ascii. Base64-code line of data");
437
438static PyObject *
439binascii_b2a_base64(PyObject *self, PyObject *args)
440{
441 unsigned char *ascii_data, *bin_data;
442 int leftbits = 0;
443 unsigned char this_ch;
444 unsigned int leftchar = 0;
445 PyObject *rv;
446 Py_ssize_t bin_len;
447
448 if ( !PyArg_ParseTuple(args, "s#:b2a_base64", &bin_data, &bin_len) )
449 return NULL;
450
451 assert(bin_len >= 0);
452
453 if ( bin_len > BASE64_MAXBIN ) {
454 PyErr_SetString(Error, "Too much data for base64 line");
455 return NULL;
456 }
457
458 /* We're lazy and allocate too much (fixed up later).
459 "+3" leaves room for up to two pad characters and a trailing
460 newline. Note that 'b' gets encoded as 'Yg==\n' (1 in, 5 out). */
461 if ( (rv=PyString_FromStringAndSize(NULL, bin_len*2 + 3)) == NULL )
462 return NULL;
463 ascii_data = (unsigned char *)PyString_AsString(rv);
464
465 for( ; bin_len > 0 ; bin_len--, bin_data++ ) {
466 /* Shift the data into our buffer */
467 leftchar = (leftchar << 8) | *bin_data;
468 leftbits += 8;
469
470 /* See if there are 6-bit groups ready */
471 while ( leftbits >= 6 ) {
472 this_ch = (leftchar >> (leftbits-6)) & 0x3f;
473 leftbits -= 6;
474 *ascii_data++ = table_b2a_base64[this_ch];
475 }
476 }
477 if ( leftbits == 2 ) {
478 *ascii_data++ = table_b2a_base64[(leftchar&3) << 4];
479 *ascii_data++ = BASE64_PAD;
480 *ascii_data++ = BASE64_PAD;
481 } else if ( leftbits == 4 ) {
482 *ascii_data++ = table_b2a_base64[(leftchar&0xf) << 2];
483 *ascii_data++ = BASE64_PAD;
484 }
485 *ascii_data++ = '\n'; /* Append a courtesy newline */
486
487 _PyString_Resize(&rv, (ascii_data -
488 (unsigned char *)PyString_AsString(rv)));
489 return rv;
490}
491
492PyDoc_STRVAR(doc_a2b_hqx, "ascii -> bin, done. Decode .hqx coding");
493
494static PyObject *
495binascii_a2b_hqx(PyObject *self, PyObject *args)
496{
497 unsigned char *ascii_data, *bin_data;
498 int leftbits = 0;
499 unsigned char this_ch;
500 unsigned int leftchar = 0;
501 PyObject *rv;
502 Py_ssize_t len;
503 int done = 0;
504
505 if ( !PyArg_ParseTuple(args, "t#:a2b_hqx", &ascii_data, &len) )
506 return NULL;
507
508 assert(len >= 0);
509
510 if (len > PY_SSIZE_T_MAX - 2)
511 return PyErr_NoMemory();
512
513 /* Allocate a string that is too big (fixed later)
514 Add two to the initial length to prevent interning which
515 would preclude subsequent resizing. */
516 if ( (rv=PyString_FromStringAndSize(NULL, len+2)) == NULL )
517 return NULL;
518 bin_data = (unsigned char *)PyString_AsString(rv);
519
520 for( ; len > 0 ; len--, ascii_data++ ) {
521 /* Get the byte and look it up */
522 this_ch = table_a2b_hqx[*ascii_data];
523 if ( this_ch == SKIP )
524 continue;
525 if ( this_ch == FAIL ) {
526 PyErr_SetString(Error, "Illegal char");
527 Py_DECREF(rv);
528 return NULL;
529 }
530 if ( this_ch == DONE ) {
531 /* The terminating colon */
532 done = 1;
533 break;
534 }
535
536 /* Shift it into the buffer and see if any bytes are ready */
537 leftchar = (leftchar << 6) | (this_ch);
538 leftbits += 6;
539 if ( leftbits >= 8 ) {
540 leftbits -= 8;
541 *bin_data++ = (leftchar >> leftbits) & 0xff;
542 leftchar &= ((1 << leftbits) - 1);
543 }
544 }
545
546 if ( leftbits && !done ) {
547 PyErr_SetString(Incomplete,
548 "String has incomplete number of bytes");
549 Py_DECREF(rv);
550 return NULL;
551 }
552 _PyString_Resize(
553 &rv, (bin_data - (unsigned char *)PyString_AsString(rv)));
554 if (rv) {
555 PyObject *rrv = Py_BuildValue("Oi", rv, done);
556 Py_DECREF(rv);
557 return rrv;
558 }
559
560 return NULL;
561}
562
563PyDoc_STRVAR(doc_rlecode_hqx, "Binhex RLE-code binary data");
564
565static PyObject *
566binascii_rlecode_hqx(PyObject *self, PyObject *args)
567{
568 unsigned char *in_data, *out_data;
569 PyObject *rv;
570 unsigned char ch;
571 Py_ssize_t in, inend, len;
572
573 if ( !PyArg_ParseTuple(args, "s#:rlecode_hqx", &in_data, &len) )
574 return NULL;
575
576 assert(len >= 0);
577
578 if (len > PY_SSIZE_T_MAX / 2 - 2)
579 return PyErr_NoMemory();
580
581 /* Worst case: output is twice as big as input (fixed later) */
582 if ( (rv=PyString_FromStringAndSize(NULL, len*2+2)) == NULL )
583 return NULL;
584 out_data = (unsigned char *)PyString_AsString(rv);
585
586 for( in=0; in<len; in++) {
587 ch = in_data[in];
588 if ( ch == RUNCHAR ) {
589 /* RUNCHAR. Escape it. */
590 *out_data++ = RUNCHAR;
591 *out_data++ = 0;
592 } else {
593 /* Check how many following are the same */
594 for(inend=in+1;
595 inend<len && in_data[inend] == ch &&
596 inend < in+255;
597 inend++) ;
598 if ( inend - in > 3 ) {
599 /* More than 3 in a row. Output RLE. */
600 *out_data++ = ch;
601 *out_data++ = RUNCHAR;
602 *out_data++ = inend-in;
603 in = inend-1;
604 } else {
605 /* Less than 3. Output the byte itself */
606 *out_data++ = ch;
607 }
608 }
609 }
610 _PyString_Resize(&rv, (out_data -
611 (unsigned char *)PyString_AsString(rv)));
612 return rv;
613}
614
615PyDoc_STRVAR(doc_b2a_hqx, "Encode .hqx data");
616
617static PyObject *
618binascii_b2a_hqx(PyObject *self, PyObject *args)
619{
620 unsigned char *ascii_data, *bin_data;
621 int leftbits = 0;
622 unsigned char this_ch;
623 unsigned int leftchar = 0;
624 PyObject *rv;
625 Py_ssize_t len;
626
627 if ( !PyArg_ParseTuple(args, "s#:b2a_hqx", &bin_data, &len) )
628 return NULL;
629
630 assert(len >= 0);
631
632 if (len > PY_SSIZE_T_MAX / 2 - 2)
633 return PyErr_NoMemory();
634
635 /* Allocate a buffer that is at least large enough */
636 if ( (rv=PyString_FromStringAndSize(NULL, len*2+2)) == NULL )
637 return NULL;
638 ascii_data = (unsigned char *)PyString_AsString(rv);
639
640 for( ; len > 0 ; len--, bin_data++ ) {
641 /* Shift into our buffer, and output any 6bits ready */
642 leftchar = (leftchar << 8) | *bin_data;
643 leftbits += 8;
644 while ( leftbits >= 6 ) {
645 this_ch = (leftchar >> (leftbits-6)) & 0x3f;
646 leftbits -= 6;
647 *ascii_data++ = table_b2a_hqx[this_ch];
648 }
649 }
650 /* Output a possible runt byte */
651 if ( leftbits ) {
652 leftchar <<= (6-leftbits);
653 *ascii_data++ = table_b2a_hqx[leftchar & 0x3f];
654 }
655 _PyString_Resize(&rv, (ascii_data -
656 (unsigned char *)PyString_AsString(rv)));
657 return rv;
658}
659
660PyDoc_STRVAR(doc_rledecode_hqx, "Decode hexbin RLE-coded string");
661
662static PyObject *
663binascii_rledecode_hqx(PyObject *self, PyObject *args)
664{
665 unsigned char *in_data, *out_data;
666 unsigned char in_byte, in_repeat;
667 PyObject *rv;
668 Py_ssize_t in_len, out_len, out_len_left;
669
670 if ( !PyArg_ParseTuple(args, "s#:rledecode_hqx", &in_data, &in_len) )
671 return NULL;
672
673 assert(in_len >= 0);
674
675 /* Empty string is a special case */
676 if ( in_len == 0 )
677 return PyString_FromString("");
678 else if (in_len > PY_SSIZE_T_MAX / 2)
679 return PyErr_NoMemory();
680
681 /* Allocate a buffer of reasonable size. Resized when needed */
682 out_len = in_len*2;
683 if ( (rv=PyString_FromStringAndSize(NULL, out_len)) == NULL )
684 return NULL;
685 out_len_left = out_len;
686 out_data = (unsigned char *)PyString_AsString(rv);
687
688 /*
689 ** We need two macros here to get/put bytes and handle
690 ** end-of-buffer for input and output strings.
691 */
692#define INBYTE(b) \
693 do { \
694 if ( --in_len < 0 ) { \
695 PyErr_SetString(Incomplete, ""); \
696 Py_DECREF(rv); \
697 return NULL; \
698 } \
699 b = *in_data++; \
700 } while(0)
701
702#define OUTBYTE(b) \
703 do { \
704 if ( --out_len_left < 0 ) { \
705 if ( out_len > PY_SSIZE_T_MAX / 2) return PyErr_NoMemory(); \
706 _PyString_Resize(&rv, 2*out_len); \
707 if ( rv == NULL ) return NULL; \
708 out_data = (unsigned char *)PyString_AsString(rv) \
709 + out_len; \
710 out_len_left = out_len-1; \
711 out_len = out_len * 2; \
712 } \
713 *out_data++ = b; \
714 } while(0)
715
716 /*
717 ** Handle first byte separately (since we have to get angry
718 ** in case of an orphaned RLE code).
719 */
720 INBYTE(in_byte);
721
722 if (in_byte == RUNCHAR) {
723 INBYTE(in_repeat);
724 if (in_repeat != 0) {
725 /* Note Error, not Incomplete (which is at the end
726 ** of the string only). This is a programmer error.
727 */
728 PyErr_SetString(Error, "Orphaned RLE code at start");
729 Py_DECREF(rv);
730 return NULL;
731 }
732 OUTBYTE(RUNCHAR);
733 } else {
734 OUTBYTE(in_byte);
735 }
736
737 while( in_len > 0 ) {
738 INBYTE(in_byte);
739
740 if (in_byte == RUNCHAR) {
741 INBYTE(in_repeat);
742 if ( in_repeat == 0 ) {
743 /* Just an escaped RUNCHAR value */
744 OUTBYTE(RUNCHAR);
745 } else {
746 /* Pick up value and output a sequence of it */
747 in_byte = out_data[-1];
748 while ( --in_repeat > 0 )
749 OUTBYTE(in_byte);
750 }
751 } else {
752 /* Normal byte */
753 OUTBYTE(in_byte);
754 }
755 }
756 _PyString_Resize(&rv, (out_data -
757 (unsigned char *)PyString_AsString(rv)));
758 return rv;
759}
760
761PyDoc_STRVAR(doc_crc_hqx,
762"(data, oldcrc) -> newcrc. Compute hqx CRC incrementally");
763
764static PyObject *
765binascii_crc_hqx(PyObject *self, PyObject *args)
766{
767 unsigned char *bin_data;
768 unsigned int crc;
769 Py_ssize_t len;
770
771 if ( !PyArg_ParseTuple(args, "s#i:crc_hqx", &bin_data, &len, &crc) )
772 return NULL;
773
774 while(len-- > 0) {
775 crc=((crc<<8)&0xff00)^crctab_hqx[((crc>>8)&0xff)^*bin_data++];
776 }
777
778 return Py_BuildValue("i", crc);
779}
780
781PyDoc_STRVAR(doc_crc32,
782"(data, oldcrc = 0) -> newcrc. Compute CRC-32 incrementally");
783
784#ifdef USE_ZLIB_CRC32
785/* This was taken from zlibmodule.c PyZlib_crc32 (but is PY_SSIZE_T_CLEAN) */
786static PyObject *
787binascii_crc32(PyObject *self, PyObject *args)
788{
789 unsigned int crc32val = 0; /* crc32(0L, Z_NULL, 0) */
790 Byte *buf;
791 Py_ssize_t len;
792 int signed_val;
793
794 if (!PyArg_ParseTuple(args, "s#|I:crc32", &buf, &len, &crc32val))
795 return NULL;
796 /* In Python 2.x we return a signed integer regardless of native platform
797 * long size (the 32bit unsigned long is treated as 32-bit signed and sign
798 * extended into a 64-bit long inside the integer object). 3.0 does the
799 * right thing and returns unsigned. http://bugs.python.org/issue1202 */
800 signed_val = crc32(crc32val, buf, len);
801 return PyInt_FromLong(signed_val);
802}
803#else /* USE_ZLIB_CRC32 */
804/* Crc - 32 BIT ANSI X3.66 CRC checksum files
805 Also known as: ISO 3307
806**********************************************************************|
807* *|
808* Demonstration program to compute the 32-bit CRC used as the frame *|
809* check sequence in ADCCP (ANSI X3.66, also known as FIPS PUB 71 *|
810* and FED-STD-1003, the U.S. versions of CCITT's X.25 link-level *|
811* protocol). The 32-bit FCS was added via the Federal Register, *|
812* 1 June 1982, p.23798. I presume but don't know for certain that *|
813* this polynomial is or will be included in CCITT V.41, which *|
814* defines the 16-bit CRC (often called CRC-CCITT) polynomial. FIPS *|
815* PUB 78 says that the 32-bit FCS reduces otherwise undetected *|
816* errors by a factor of 10^-5 over 16-bit FCS. *|
817* *|
818**********************************************************************|
819
820 Copyright (C) 1986 Gary S. Brown. You may use this program, or
821 code or tables extracted from it, as desired without restriction.
822
823 First, the polynomial itself and its table of feedback terms. The
824 polynomial is
825 X^32+X^26+X^23+X^22+X^16+X^12+X^11+X^10+X^8+X^7+X^5+X^4+X^2+X^1+X^0
826 Note that we take it "backwards" and put the highest-order term in
827 the lowest-order bit. The X^32 term is "implied"; the LSB is the
828 X^31 term, etc. The X^0 term (usually shown as "+1") results in
829 the MSB being 1.
830
831 Note that the usual hardware shift register implementation, which
832 is what we're using (we're merely optimizing it by doing eight-bit
833 chunks at a time) shifts bits into the lowest-order term. In our
834 implementation, that means shifting towards the right. Why do we
835 do it this way? Because the calculated CRC must be transmitted in
836 order from highest-order term to lowest-order term. UARTs transmit
837 characters in order from LSB to MSB. By storing the CRC this way,
838 we hand it to the UART in the order low-byte to high-byte; the UART
839 sends each low-bit to hight-bit; and the result is transmission bit
840 by bit from highest- to lowest-order term without requiring any bit
841 shuffling on our part. Reception works similarly.
842
843 The feedback terms table consists of 256, 32-bit entries. Notes:
844
845 1. The table can be generated at runtime if desired; code to do so
846 is shown later. It might not be obvious, but the feedback
847 terms simply represent the results of eight shift/xor opera-
848 tions for all combinations of data and CRC register values.
849
850 2. The CRC accumulation logic is the same for all CRC polynomials,
851 be they sixteen or thirty-two bits wide. You simply choose the
852 appropriate table. Alternatively, because the table can be
853 generated at runtime, you can start by generating the table for
854 the polynomial in question and use exactly the same "updcrc",
855 if your application needn't simultaneously handle two CRC
856 polynomials. (Note, however, that XMODEM is strange.)
857
858 3. For 16-bit CRCs, the table entries need be only 16 bits wide;
859 of course, 32-bit entries work OK if the high 16 bits are zero.
860
861 4. The values must be right-shifted by eight bits by the "updcrc"
862 logic; the shift must be unsigned (bring in zeroes). On some
863 hardware you could probably optimize the shift in assembler by
864 using byte-swap instructions.
865********************************************************************/
866
867static unsigned int crc_32_tab[256] = {
8680x00000000U, 0x77073096U, 0xee0e612cU, 0x990951baU, 0x076dc419U,
8690x706af48fU, 0xe963a535U, 0x9e6495a3U, 0x0edb8832U, 0x79dcb8a4U,
8700xe0d5e91eU, 0x97d2d988U, 0x09b64c2bU, 0x7eb17cbdU, 0xe7b82d07U,
8710x90bf1d91U, 0x1db71064U, 0x6ab020f2U, 0xf3b97148U, 0x84be41deU,
8720x1adad47dU, 0x6ddde4ebU, 0xf4d4b551U, 0x83d385c7U, 0x136c9856U,
8730x646ba8c0U, 0xfd62f97aU, 0x8a65c9ecU, 0x14015c4fU, 0x63066cd9U,
8740xfa0f3d63U, 0x8d080df5U, 0x3b6e20c8U, 0x4c69105eU, 0xd56041e4U,
8750xa2677172U, 0x3c03e4d1U, 0x4b04d447U, 0xd20d85fdU, 0xa50ab56bU,
8760x35b5a8faU, 0x42b2986cU, 0xdbbbc9d6U, 0xacbcf940U, 0x32d86ce3U,
8770x45df5c75U, 0xdcd60dcfU, 0xabd13d59U, 0x26d930acU, 0x51de003aU,
8780xc8d75180U, 0xbfd06116U, 0x21b4f4b5U, 0x56b3c423U, 0xcfba9599U,
8790xb8bda50fU, 0x2802b89eU, 0x5f058808U, 0xc60cd9b2U, 0xb10be924U,
8800x2f6f7c87U, 0x58684c11U, 0xc1611dabU, 0xb6662d3dU, 0x76dc4190U,
8810x01db7106U, 0x98d220bcU, 0xefd5102aU, 0x71b18589U, 0x06b6b51fU,
8820x9fbfe4a5U, 0xe8b8d433U, 0x7807c9a2U, 0x0f00f934U, 0x9609a88eU,
8830xe10e9818U, 0x7f6a0dbbU, 0x086d3d2dU, 0x91646c97U, 0xe6635c01U,
8840x6b6b51f4U, 0x1c6c6162U, 0x856530d8U, 0xf262004eU, 0x6c0695edU,
8850x1b01a57bU, 0x8208f4c1U, 0xf50fc457U, 0x65b0d9c6U, 0x12b7e950U,
8860x8bbeb8eaU, 0xfcb9887cU, 0x62dd1ddfU, 0x15da2d49U, 0x8cd37cf3U,
8870xfbd44c65U, 0x4db26158U, 0x3ab551ceU, 0xa3bc0074U, 0xd4bb30e2U,
8880x4adfa541U, 0x3dd895d7U, 0xa4d1c46dU, 0xd3d6f4fbU, 0x4369e96aU,
8890x346ed9fcU, 0xad678846U, 0xda60b8d0U, 0x44042d73U, 0x33031de5U,
8900xaa0a4c5fU, 0xdd0d7cc9U, 0x5005713cU, 0x270241aaU, 0xbe0b1010U,
8910xc90c2086U, 0x5768b525U, 0x206f85b3U, 0xb966d409U, 0xce61e49fU,
8920x5edef90eU, 0x29d9c998U, 0xb0d09822U, 0xc7d7a8b4U, 0x59b33d17U,
8930x2eb40d81U, 0xb7bd5c3bU, 0xc0ba6cadU, 0xedb88320U, 0x9abfb3b6U,
8940x03b6e20cU, 0x74b1d29aU, 0xead54739U, 0x9dd277afU, 0x04db2615U,
8950x73dc1683U, 0xe3630b12U, 0x94643b84U, 0x0d6d6a3eU, 0x7a6a5aa8U,
8960xe40ecf0bU, 0x9309ff9dU, 0x0a00ae27U, 0x7d079eb1U, 0xf00f9344U,
8970x8708a3d2U, 0x1e01f268U, 0x6906c2feU, 0xf762575dU, 0x806567cbU,
8980x196c3671U, 0x6e6b06e7U, 0xfed41b76U, 0x89d32be0U, 0x10da7a5aU,
8990x67dd4accU, 0xf9b9df6fU, 0x8ebeeff9U, 0x17b7be43U, 0x60b08ed5U,
9000xd6d6a3e8U, 0xa1d1937eU, 0x38d8c2c4U, 0x4fdff252U, 0xd1bb67f1U,
9010xa6bc5767U, 0x3fb506ddU, 0x48b2364bU, 0xd80d2bdaU, 0xaf0a1b4cU,
9020x36034af6U, 0x41047a60U, 0xdf60efc3U, 0xa867df55U, 0x316e8eefU,
9030x4669be79U, 0xcb61b38cU, 0xbc66831aU, 0x256fd2a0U, 0x5268e236U,
9040xcc0c7795U, 0xbb0b4703U, 0x220216b9U, 0x5505262fU, 0xc5ba3bbeU,
9050xb2bd0b28U, 0x2bb45a92U, 0x5cb36a04U, 0xc2d7ffa7U, 0xb5d0cf31U,
9060x2cd99e8bU, 0x5bdeae1dU, 0x9b64c2b0U, 0xec63f226U, 0x756aa39cU,
9070x026d930aU, 0x9c0906a9U, 0xeb0e363fU, 0x72076785U, 0x05005713U,
9080x95bf4a82U, 0xe2b87a14U, 0x7bb12baeU, 0x0cb61b38U, 0x92d28e9bU,
9090xe5d5be0dU, 0x7cdcefb7U, 0x0bdbdf21U, 0x86d3d2d4U, 0xf1d4e242U,
9100x68ddb3f8U, 0x1fda836eU, 0x81be16cdU, 0xf6b9265bU, 0x6fb077e1U,
9110x18b74777U, 0x88085ae6U, 0xff0f6a70U, 0x66063bcaU, 0x11010b5cU,
9120x8f659effU, 0xf862ae69U, 0x616bffd3U, 0x166ccf45U, 0xa00ae278U,
9130xd70dd2eeU, 0x4e048354U, 0x3903b3c2U, 0xa7672661U, 0xd06016f7U,
9140x4969474dU, 0x3e6e77dbU, 0xaed16a4aU, 0xd9d65adcU, 0x40df0b66U,
9150x37d83bf0U, 0xa9bcae53U, 0xdebb9ec5U, 0x47b2cf7fU, 0x30b5ffe9U,
9160xbdbdf21cU, 0xcabac28aU, 0x53b39330U, 0x24b4a3a6U, 0xbad03605U,
9170xcdd70693U, 0x54de5729U, 0x23d967bfU, 0xb3667a2eU, 0xc4614ab8U,
9180x5d681b02U, 0x2a6f2b94U, 0xb40bbe37U, 0xc30c8ea1U, 0x5a05df1bU,
9190x2d02ef8dU
920};
921
922static PyObject *
923binascii_crc32(PyObject *self, PyObject *args)
924{ /* By Jim Ahlstrom; All rights transferred to CNRI */
925 unsigned char *bin_data;
926 unsigned int crc = 0U; /* initial value of CRC */
927 Py_ssize_t len;
928 int result;
929
930 if ( !PyArg_ParseTuple(args, "s#|I:crc32", &bin_data, &len, &crc) )
931 return NULL;
932
933 crc = ~ crc;
934 while (len-- > 0)
935 crc = crc_32_tab[(crc ^ *bin_data++) & 0xffU] ^ (crc >> 8);
936 /* Note: (crc >> 8) MUST zero fill on left */
937
938 result = (int)(crc ^ 0xFFFFFFFFU);
939 return PyInt_FromLong(result);
940}
941#endif /* USE_ZLIB_CRC32 */
942
943
944static PyObject *
945binascii_hexlify(PyObject *self, PyObject *args)
946{
947 char* argbuf;
948 Py_ssize_t arglen;
949 PyObject *retval;
950 char* retbuf;
951 Py_ssize_t i, j;
952
953 if (!PyArg_ParseTuple(args, "s#:b2a_hex", &argbuf, &arglen))
954 return NULL;
955
956 assert(arglen >= 0);
957 if (arglen > PY_SSIZE_T_MAX / 2)
958 return PyErr_NoMemory();
959
960 retval = PyString_FromStringAndSize(NULL, arglen*2);
961 if (!retval)
962 return NULL;
963 retbuf = PyString_AsString(retval);
964 if (!retbuf)
965 goto finally;
966
967 /* make hex version of string, taken from shamodule.c */
968 for (i=j=0; i < arglen; i++) {
969 char c;
970 c = (argbuf[i] >> 4) & 0xf;
971 c = (c>9) ? c+'a'-10 : c + '0';
972 retbuf[j++] = c;
973 c = argbuf[i] & 0xf;
974 c = (c>9) ? c+'a'-10 : c + '0';
975 retbuf[j++] = c;
976 }
977 return retval;
978
979 finally:
980 Py_DECREF(retval);
981 return NULL;
982}
983
984PyDoc_STRVAR(doc_hexlify,
985"b2a_hex(data) -> s; Hexadecimal representation of binary data.\n\
986\n\
987This function is also available as \"hexlify()\".");
988
989
990static int
991to_int(int c)
992{
993 if (isdigit(c))
994 return c - '0';
995 else {
996 if (isupper(c))
997 c = tolower(c);
998 if (c >= 'a' && c <= 'f')
999 return c - 'a' + 10;
1000 }
1001 return -1;
1002}
1003
1004
1005static PyObject *
1006binascii_unhexlify(PyObject *self, PyObject *args)
1007{
1008 char* argbuf;
1009 Py_ssize_t arglen;
1010 PyObject *retval;
1011 char* retbuf;
1012 Py_ssize_t i, j;
1013
1014 if (!PyArg_ParseTuple(args, "s#:a2b_hex", &argbuf, &arglen))
1015 return NULL;
1016
1017 assert(arglen >= 0);
1018
1019 /* XXX What should we do about strings with an odd length? Should
1020 * we add an implicit leading zero, or a trailing zero? For now,
1021 * raise an exception.
1022 */
1023 if (arglen % 2) {
1024 PyErr_SetString(PyExc_TypeError, "Odd-length string");
1025 return NULL;
1026 }
1027
1028 retval = PyString_FromStringAndSize(NULL, (arglen/2));
1029 if (!retval)
1030 return NULL;
1031 retbuf = PyString_AsString(retval);
1032 if (!retbuf)
1033 goto finally;
1034
1035 for (i=j=0; i < arglen; i += 2) {
1036 int top = to_int(Py_CHARMASK(argbuf[i]));
1037 int bot = to_int(Py_CHARMASK(argbuf[i+1]));
1038 if (top == -1 || bot == -1) {
1039 PyErr_SetString(PyExc_TypeError,
1040 "Non-hexadecimal digit found");
1041 goto finally;
1042 }
1043 retbuf[j++] = (top << 4) + bot;
1044 }
1045 return retval;
1046
1047 finally:
1048 Py_DECREF(retval);
1049 return NULL;
1050}
1051
1052PyDoc_STRVAR(doc_unhexlify,
1053"a2b_hex(hexstr) -> s; Binary data of hexadecimal representation.\n\
1054\n\
1055hexstr must contain an even number of hex digits (upper or lower case).\n\
1056This function is also available as \"unhexlify()\"");
1057
1058static int table_hex[128] = {
1059 -1,-1,-1,-1, -1,-1,-1,-1, -1,-1,-1,-1, -1,-1,-1,-1,
1060 -1,-1,-1,-1, -1,-1,-1,-1, -1,-1,-1,-1, -1,-1,-1,-1,
1061 -1,-1,-1,-1, -1,-1,-1,-1, -1,-1,-1,-1, -1,-1,-1,-1,
1062 0, 1, 2, 3, 4, 5, 6, 7, 8, 9,-1,-1, -1,-1,-1,-1,
1063 -1,10,11,12, 13,14,15,-1, -1,-1,-1,-1, -1,-1,-1,-1,
1064 -1,-1,-1,-1, -1,-1,-1,-1, -1,-1,-1,-1, -1,-1,-1,-1,
1065 -1,10,11,12, 13,14,15,-1, -1,-1,-1,-1, -1,-1,-1,-1,
1066 -1,-1,-1,-1, -1,-1,-1,-1, -1,-1,-1,-1, -1,-1,-1,-1
1067};
1068
1069#define hexval(c) table_hex[(unsigned int)(c)]
1070
1071#define MAXLINESIZE 76
1072
1073PyDoc_STRVAR(doc_a2b_qp, "Decode a string of qp-encoded data");
1074
1075static PyObject*
1076binascii_a2b_qp(PyObject *self, PyObject *args, PyObject *kwargs)
1077{
1078 Py_ssize_t in, out;
1079 char ch;
1080 unsigned char *data, *odata;
1081 Py_ssize_t datalen = 0;
1082 PyObject *rv;
1083 static char *kwlist[] = {"data", "header", NULL};
1084 int header = 0;
1085
1086 if (!PyArg_ParseTupleAndKeywords(args, kwargs, "s#|i", kwlist, &data,
1087 &datalen, &header))
1088 return NULL;
1089
1090 /* We allocate the output same size as input, this is overkill.
1091 * The previous implementation used calloc() so we'll zero out the
1092 * memory here too, since PyMem_Malloc() does not guarantee that.
1093 */
1094 odata = (unsigned char *) PyMem_Malloc(datalen);
1095 if (odata == NULL) {
1096 PyErr_NoMemory();
1097 return NULL;
1098 }
1099 memset(odata, 0, datalen);
1100
1101 in = out = 0;
1102 while (in < datalen) {
1103 if (data[in] == '=') {
1104 in++;
1105 if (in >= datalen) break;
1106 /* Soft line breaks */
1107 if ((data[in] == '\n') || (data[in] == '\r')) {
1108 if (data[in] != '\n') {
1109 while (in < datalen && data[in] != '\n') in++;
1110 }
1111 if (in < datalen) in++;
1112 }
1113 else if (data[in] == '=') {
1114 /* broken case from broken python qp */
1115 odata[out++] = '=';
1116 in++;
1117 }
1118 else if (((data[in] >= 'A' && data[in] <= 'F') ||
1119 (data[in] >= 'a' && data[in] <= 'f') ||
1120 (data[in] >= '0' && data[in] <= '9')) &&
1121 ((data[in+1] >= 'A' && data[in+1] <= 'F') ||
1122 (data[in+1] >= 'a' && data[in+1] <= 'f') ||
1123 (data[in+1] >= '0' && data[in+1] <= '9'))) {
1124 /* hexval */
1125 ch = hexval(data[in]) << 4;
1126 in++;
1127 ch |= hexval(data[in]);
1128 in++;
1129 odata[out++] = ch;
1130 }
1131 else {
1132 odata[out++] = '=';
1133 }
1134 }
1135 else if (header && data[in] == '_') {
1136 odata[out++] = ' ';
1137 in++;
1138 }
1139 else {
1140 odata[out] = data[in];
1141 in++;
1142 out++;
1143 }
1144 }
1145 if ((rv = PyString_FromStringAndSize((char *)odata, out)) == NULL) {
1146 PyMem_Free(odata);
1147 return NULL;
1148 }
1149 PyMem_Free(odata);
1150 return rv;
1151}
1152
1153static int
1154to_hex (unsigned char ch, unsigned char *s)
1155{
1156 unsigned int uvalue = ch;
1157
1158 s[1] = "0123456789ABCDEF"[uvalue % 16];
1159 uvalue = (uvalue / 16);
1160 s[0] = "0123456789ABCDEF"[uvalue % 16];
1161 return 0;
1162}
1163
1164PyDoc_STRVAR(doc_b2a_qp,
1165"b2a_qp(data, quotetabs=0, istext=1, header=0) -> s; \n\
1166 Encode a string using quoted-printable encoding. \n\
1167\n\
1168On encoding, when istext is set, newlines are not encoded, and white \n\
1169space at end of lines is. When istext is not set, \\r and \\n (CR/LF) are \n\
1170both encoded. When quotetabs is set, space and tabs are encoded.");
1171
1172/* XXX: This is ridiculously complicated to be backward compatible
1173 * (mostly) with the quopri module. It doesn't re-create the quopri
1174 * module bug where text ending in CRLF has the CR encoded */
1175static PyObject*
1176binascii_b2a_qp (PyObject *self, PyObject *args, PyObject *kwargs)
1177{
1178 Py_ssize_t in, out;
1179 unsigned char *data, *odata;
1180 Py_ssize_t datalen = 0, odatalen = 0;
1181 PyObject *rv;
1182 unsigned int linelen = 0;
1183 static char *kwlist[] = {"data", "quotetabs", "istext",
1184 "header", NULL};
1185 int istext = 1;
1186 int quotetabs = 0;
1187 int header = 0;
1188 unsigned char ch;
1189 int crlf = 0;
1190 unsigned char *p;
1191
1192 if (!PyArg_ParseTupleAndKeywords(args, kwargs, "s#|iii", kwlist, &data,
1193 &datalen, &quotetabs, &istext, &header))
1194 return NULL;
1195
1196 /* See if this string is using CRLF line ends */
1197 /* XXX: this function has the side effect of converting all of
1198 * the end of lines to be the same depending on this detection
1199 * here */
1200 p = (unsigned char *) memchr(data, '\n', datalen);
1201 if ((p != NULL) && (p > data) && (*(p-1) == '\r'))
1202 crlf = 1;
1203
1204 /* First, scan to see how many characters need to be encoded */
1205 in = 0;
1206 while (in < datalen) {
1207 if ((data[in] > 126) ||
1208 (data[in] == '=') ||
1209 (header && data[in] == '_') ||
1210 ((data[in] == '.') && (linelen == 0) &&
1211 (data[in+1] == '\n' || data[in+1] == '\r' || data[in+1] == 0)) ||
1212 (!istext && ((data[in] == '\r') || (data[in] == '\n'))) ||
1213 ((data[in] == '\t' || data[in] == ' ') && (in + 1 == datalen)) ||
1214 ((data[in] < 33) &&
1215 (data[in] != '\r') && (data[in] != '\n') &&
1216 (quotetabs ||
1217 (!quotetabs && ((data[in] != '\t') && (data[in] != ' '))))))
1218 {
1219 if ((linelen + 3) >= MAXLINESIZE) {
1220 linelen = 0;
1221 if (crlf)
1222 odatalen += 3;
1223 else
1224 odatalen += 2;
1225 }
1226 linelen += 3;
1227 odatalen += 3;
1228 in++;
1229 }
1230 else {
1231 if (istext &&
1232 ((data[in] == '\n') ||
1233 ((in+1 < datalen) && (data[in] == '\r') &&
1234 (data[in+1] == '\n'))))
1235 {
1236 linelen = 0;
1237 /* Protect against whitespace on end of line */
1238 if (in && ((data[in-1] == ' ') || (data[in-1] == '\t')))
1239 odatalen += 2;
1240 if (crlf)
1241 odatalen += 2;
1242 else
1243 odatalen += 1;
1244 if (data[in] == '\r')
1245 in += 2;
1246 else
1247 in++;
1248 }
1249 else {
1250 if ((in + 1 != datalen) &&
1251 (data[in+1] != '\n') &&
1252 (linelen + 1) >= MAXLINESIZE) {
1253 linelen = 0;
1254 if (crlf)
1255 odatalen += 3;
1256 else
1257 odatalen += 2;
1258 }
1259 linelen++;
1260 odatalen++;
1261 in++;
1262 }
1263 }
1264 }
1265
1266 /* We allocate the output same size as input, this is overkill.
1267 * The previous implementation used calloc() so we'll zero out the
1268 * memory here too, since PyMem_Malloc() does not guarantee that.
1269 */
1270 odata = (unsigned char *) PyMem_Malloc(odatalen);
1271 if (odata == NULL) {
1272 PyErr_NoMemory();
1273 return NULL;
1274 }
1275 memset(odata, 0, odatalen);
1276
1277 in = out = linelen = 0;
1278 while (in < datalen) {
1279 if ((data[in] > 126) ||
1280 (data[in] == '=') ||
1281 (header && data[in] == '_') ||
1282 ((data[in] == '.') && (linelen == 0) &&
1283 (data[in+1] == '\n' || data[in+1] == '\r' || data[in+1] == 0)) ||
1284 (!istext && ((data[in] == '\r') || (data[in] == '\n'))) ||
1285 ((data[in] == '\t' || data[in] == ' ') && (in + 1 == datalen)) ||
1286 ((data[in] < 33) &&
1287 (data[in] != '\r') && (data[in] != '\n') &&
1288 (quotetabs ||
1289 (!quotetabs && ((data[in] != '\t') && (data[in] != ' '))))))
1290 {
1291 if ((linelen + 3 )>= MAXLINESIZE) {
1292 odata[out++] = '=';
1293 if (crlf) odata[out++] = '\r';
1294 odata[out++] = '\n';
1295 linelen = 0;
1296 }
1297 odata[out++] = '=';
1298 to_hex(data[in], &odata[out]);
1299 out += 2;
1300 in++;
1301 linelen += 3;
1302 }
1303 else {
1304 if (istext &&
1305 ((data[in] == '\n') ||
1306 ((in+1 < datalen) && (data[in] == '\r') &&
1307 (data[in+1] == '\n'))))
1308 {
1309 linelen = 0;
1310 /* Protect against whitespace on end of line */
1311 if (out && ((odata[out-1] == ' ') || (odata[out-1] == '\t'))) {
1312 ch = odata[out-1];
1313 odata[out-1] = '=';
1314 to_hex(ch, &odata[out]);
1315 out += 2;
1316 }
1317
1318 if (crlf) odata[out++] = '\r';
1319 odata[out++] = '\n';
1320 if (data[in] == '\r')
1321 in += 2;
1322 else
1323 in++;
1324 }
1325 else {
1326 if ((in + 1 != datalen) &&
1327 (data[in+1] != '\n') &&
1328 (linelen + 1) >= MAXLINESIZE) {
1329 odata[out++] = '=';
1330 if (crlf) odata[out++] = '\r';
1331 odata[out++] = '\n';
1332 linelen = 0;
1333 }
1334 linelen++;
1335 if (header && data[in] == ' ') {
1336 odata[out++] = '_';
1337 in++;
1338 }
1339 else {
1340 odata[out++] = data[in++];
1341 }
1342 }
1343 }
1344 }
1345 if ((rv = PyString_FromStringAndSize((char *)odata, out)) == NULL) {
1346 PyMem_Free(odata);
1347 return NULL;
1348 }
1349 PyMem_Free(odata);
1350 return rv;
1351}
1352
1353/* List of functions defined in the module */
1354
1355static struct PyMethodDef binascii_module_methods[] = {
1356 {"a2b_uu", binascii_a2b_uu, METH_VARARGS, doc_a2b_uu},
1357 {"b2a_uu", binascii_b2a_uu, METH_VARARGS, doc_b2a_uu},
1358 {"a2b_base64", binascii_a2b_base64, METH_VARARGS, doc_a2b_base64},
1359 {"b2a_base64", binascii_b2a_base64, METH_VARARGS, doc_b2a_base64},
1360 {"a2b_hqx", binascii_a2b_hqx, METH_VARARGS, doc_a2b_hqx},
1361 {"b2a_hqx", binascii_b2a_hqx, METH_VARARGS, doc_b2a_hqx},
1362 {"b2a_hex", binascii_hexlify, METH_VARARGS, doc_hexlify},
1363 {"a2b_hex", binascii_unhexlify, METH_VARARGS, doc_unhexlify},
1364 {"hexlify", binascii_hexlify, METH_VARARGS, doc_hexlify},
1365 {"unhexlify", binascii_unhexlify, METH_VARARGS, doc_unhexlify},
1366 {"rlecode_hqx", binascii_rlecode_hqx, METH_VARARGS, doc_rlecode_hqx},
1367 {"rledecode_hqx", binascii_rledecode_hqx, METH_VARARGS,
1368 doc_rledecode_hqx},
1369 {"crc_hqx", binascii_crc_hqx, METH_VARARGS, doc_crc_hqx},
1370 {"crc32", binascii_crc32, METH_VARARGS, doc_crc32},
1371 {"a2b_qp", (PyCFunction)binascii_a2b_qp, METH_VARARGS | METH_KEYWORDS,
1372 doc_a2b_qp},
1373 {"b2a_qp", (PyCFunction)binascii_b2a_qp, METH_VARARGS | METH_KEYWORDS,
1374 doc_b2a_qp},
1375 {NULL, NULL} /* sentinel */
1376};
1377
1378
1379/* Initialization function for the module (*must* be called initbinascii) */
1380PyDoc_STRVAR(doc_binascii, "Conversion between binary data and ASCII");
1381
1382PyMODINIT_FUNC
1383initbinascii(void)
1384{
1385 PyObject *m, *d, *x;
1386
1387 /* Create the module and add the functions */
1388 m = Py_InitModule("binascii", binascii_module_methods);
1389 if (m == NULL)
1390 return;
1391
1392 d = PyModule_GetDict(m);
1393 x = PyString_FromString(doc_binascii);
1394 PyDict_SetItemString(d, "__doc__", x);
1395 Py_XDECREF(x);
1396
1397 Error = PyErr_NewException("binascii.Error", NULL, NULL);
1398 PyDict_SetItemString(d, "Error", Error);
1399 Incomplete = PyErr_NewException("binascii.Incomplete", NULL, NULL);
1400 PyDict_SetItemString(d, "Incomplete", Incomplete);
1401}
Note: See TracBrowser for help on using the repository browser.