1 /* base64.c -- Encode binary data using printable characters.
2 Copyright (C) 1999-2001, 2004-2006, 2009-2011 Free Software Foundation, Inc.
4 This program is free software; you can redistribute it and/or modify
5 it under the terms of the GNU General Public License as published by
6 the Free Software Foundation; either version 2, or (at your option)
9 This program is distributed in the hope that it will be useful,
10 but WITHOUT ANY WARRANTY; without even the implied warranty of
11 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
12 GNU General Public License for more details.
14 You should have received a copy of the GNU General Public License
15 along with this program; if not, write to the Free Software Foundation,
16 Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. */
18 /* Written by Simon Josefsson. Partially adapted from GNU MailUtils
19 * (mailbox/filter_trans.c, as of 2004-11-28). Improved by review
20 * from Paul Eggert, Bruno Haible, and Stepan Kasal.
22 * See also RFC 3548 <http://www.ietf.org/rfc/rfc3548.txt>.
24 * Be careful with error checking. Here is how you would typically
25 * use these functions:
27 * bool ok = base64_decode_alloc (in, inlen, &out, &outlen);
29 * FAIL: input was not valid base64
31 * FAIL: memory allocation error
32 * OK: data in OUT/OUTLEN
34 * size_t outlen = base64_encode_alloc (in, inlen, &out);
35 * if (out == NULL && outlen == 0 && inlen != 0)
36 * FAIL: input too long
38 * FAIL: memory allocation error
39 * OK: data in OUT/OUTLEN.
56 /* C89 compliant way to cast 'char' to 'unsigned char'. */
57 static inline unsigned char
63 /* Base64 encode IN array of size INLEN into OUT array of size OUTLEN.
64 If OUTLEN is less than BASE64_LENGTH(INLEN), write as many bytes as
65 possible. If OUTLEN is larger than BASE64_LENGTH(INLEN), also zero
66 terminate the output buffer. */
68 base64_encode (const char *restrict in, size_t inlen,
69 char *restrict out, size_t outlen)
71 static const char b64str[64] =
72 "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/";
74 while (inlen && outlen)
76 *out++ = b64str[(to_uchar (in[0]) >> 2) & 0x3f];
79 *out++ = b64str[((to_uchar (in[0]) << 4)
80 + (--inlen ? to_uchar (in[1]) >> 4 : 0))
86 ? b64str[((to_uchar (in[1]) << 2)
87 + (--inlen ? to_uchar (in[2]) >> 6 : 0))
92 *out++ = inlen ? b64str[to_uchar (in[2]) & 0x3f] : '=';
105 /* Allocate a buffer and store zero terminated base64 encoded data
106 from array IN of size INLEN, returning BASE64_LENGTH(INLEN), i.e.,
107 the length of the encoded data, excluding the terminating zero. On
108 return, the OUT variable will hold a pointer to newly allocated
109 memory that must be deallocated by the caller. If output string
110 length would overflow, 0 is returned and OUT is set to NULL. If
111 memory allocation failed, OUT is set to NULL, and the return value
112 indicates length of the requested memory block, i.e.,
113 BASE64_LENGTH(inlen) + 1. */
115 base64_encode_alloc (const char *in, size_t inlen, char **out)
117 size_t outlen = 1 + BASE64_LENGTH (inlen);
119 /* Check for overflow in outlen computation.
121 * If there is no overflow, outlen >= inlen.
123 * If the operation (inlen + 2) overflows then it yields at most +1, so
126 * If the multiplication overflows, we lose at least half of the
127 * correct value, so the result is < ((inlen + 2) / 3) * 2, which is
128 * less than (inlen + 2) * 0.66667, which is less than inlen as soon as
137 *out = malloc (outlen);
141 base64_encode (in, inlen, *out, outlen);
146 /* With this approach this file works independent of the charset used
147 (think EBCDIC). However, it does assume that the characters in the
148 Base64 alphabet (A-Za-z0-9+/) are encoded in 0..255. POSIX
149 1003.1-2001 require that char and unsigned char are 8-bit
150 quantities, though, taking care of that problem. But this may be a
151 potential problem on non-POSIX C99 platforms.
153 IBM C V6 for AIX mishandles "#define B64(x) ...'x'...", so use "_"
154 as the formal parameter rather than "x". */
222 static const signed char b64[0x100] = {
223 B64 (0), B64 (1), B64 (2), B64 (3),
224 B64 (4), B64 (5), B64 (6), B64 (7),
225 B64 (8), B64 (9), B64 (10), B64 (11),
226 B64 (12), B64 (13), B64 (14), B64 (15),
227 B64 (16), B64 (17), B64 (18), B64 (19),
228 B64 (20), B64 (21), B64 (22), B64 (23),
229 B64 (24), B64 (25), B64 (26), B64 (27),
230 B64 (28), B64 (29), B64 (30), B64 (31),
231 B64 (32), B64 (33), B64 (34), B64 (35),
232 B64 (36), B64 (37), B64 (38), B64 (39),
233 B64 (40), B64 (41), B64 (42), B64 (43),
234 B64 (44), B64 (45), B64 (46), B64 (47),
235 B64 (48), B64 (49), B64 (50), B64 (51),
236 B64 (52), B64 (53), B64 (54), B64 (55),
237 B64 (56), B64 (57), B64 (58), B64 (59),
238 B64 (60), B64 (61), B64 (62), B64 (63),
239 B64 (64), B64 (65), B64 (66), B64 (67),
240 B64 (68), B64 (69), B64 (70), B64 (71),
241 B64 (72), B64 (73), B64 (74), B64 (75),
242 B64 (76), B64 (77), B64 (78), B64 (79),
243 B64 (80), B64 (81), B64 (82), B64 (83),
244 B64 (84), B64 (85), B64 (86), B64 (87),
245 B64 (88), B64 (89), B64 (90), B64 (91),
246 B64 (92), B64 (93), B64 (94), B64 (95),
247 B64 (96), B64 (97), B64 (98), B64 (99),
248 B64 (100), B64 (101), B64 (102), B64 (103),
249 B64 (104), B64 (105), B64 (106), B64 (107),
250 B64 (108), B64 (109), B64 (110), B64 (111),
251 B64 (112), B64 (113), B64 (114), B64 (115),
252 B64 (116), B64 (117), B64 (118), B64 (119),
253 B64 (120), B64 (121), B64 (122), B64 (123),
254 B64 (124), B64 (125), B64 (126), B64 (127),
255 B64 (128), B64 (129), B64 (130), B64 (131),
256 B64 (132), B64 (133), B64 (134), B64 (135),
257 B64 (136), B64 (137), B64 (138), B64 (139),
258 B64 (140), B64 (141), B64 (142), B64 (143),
259 B64 (144), B64 (145), B64 (146), B64 (147),
260 B64 (148), B64 (149), B64 (150), B64 (151),
261 B64 (152), B64 (153), B64 (154), B64 (155),
262 B64 (156), B64 (157), B64 (158), B64 (159),
263 B64 (160), B64 (161), B64 (162), B64 (163),
264 B64 (164), B64 (165), B64 (166), B64 (167),
265 B64 (168), B64 (169), B64 (170), B64 (171),
266 B64 (172), B64 (173), B64 (174), B64 (175),
267 B64 (176), B64 (177), B64 (178), B64 (179),
268 B64 (180), B64 (181), B64 (182), B64 (183),
269 B64 (184), B64 (185), B64 (186), B64 (187),
270 B64 (188), B64 (189), B64 (190), B64 (191),
271 B64 (192), B64 (193), B64 (194), B64 (195),
272 B64 (196), B64 (197), B64 (198), B64 (199),
273 B64 (200), B64 (201), B64 (202), B64 (203),
274 B64 (204), B64 (205), B64 (206), B64 (207),
275 B64 (208), B64 (209), B64 (210), B64 (211),
276 B64 (212), B64 (213), B64 (214), B64 (215),
277 B64 (216), B64 (217), B64 (218), B64 (219),
278 B64 (220), B64 (221), B64 (222), B64 (223),
279 B64 (224), B64 (225), B64 (226), B64 (227),
280 B64 (228), B64 (229), B64 (230), B64 (231),
281 B64 (232), B64 (233), B64 (234), B64 (235),
282 B64 (236), B64 (237), B64 (238), B64 (239),
283 B64 (240), B64 (241), B64 (242), B64 (243),
284 B64 (244), B64 (245), B64 (246), B64 (247),
285 B64 (248), B64 (249), B64 (250), B64 (251),
286 B64 (252), B64 (253), B64 (254), B64 (255)
290 # define uchar_in_range(c) true
292 # define uchar_in_range(c) ((c) <= 255)
295 /* Return true if CH is a character from the Base64 alphabet, and
296 false otherwise. Note that '=' is padding and not considered to be
297 part of the alphabet. */
301 return uchar_in_range (to_uchar (ch)) && 0 <= b64[to_uchar (ch)];
304 /* Initialize decode-context buffer, CTX. */
306 base64_decode_ctx_init (struct base64_decode_context *ctx)
311 /* If CTX->i is 0 or 4, there are four or more bytes in [*IN..IN_END), and
312 none of those four is a newline, then return *IN. Otherwise, copy up to
313 4 - CTX->i non-newline bytes from that range into CTX->buf, starting at
314 index CTX->i and setting CTX->i to reflect the number of bytes copied,
315 and return CTX->buf. In either case, advance *IN to point to the byte
316 after the last one processed, and set *N_NON_NEWLINE to the number of
317 verified non-newline bytes accessible through the returned pointer. */
319 get_4 (struct base64_decode_context *ctx,
320 char const *restrict *in, char const *restrict in_end,
321 size_t *n_non_newline)
329 if (4 <= in_end - *in && memchr (t, '\n', 4) == NULL)
331 /* This is the common case: no newline. */
339 /* Copy non-newline bytes into BUF. */
346 ctx->buf[ctx->i++] = c;
353 *n_non_newline = ctx->i;
358 #define return_false \
366 /* Decode up to four bytes of base64-encoded data, IN, of length INLEN
367 into the output buffer, *OUT, of size *OUTLEN bytes. Return true if
368 decoding is successful, false otherwise. If *OUTLEN is too small,
369 as many bytes as possible are written to *OUT. On return, advance
370 *OUT to point to the byte after the last one written, and decrement
371 *OUTLEN to reflect the number of bytes remaining in *OUT. */
373 decode_4 (char const *restrict in, size_t inlen,
374 char *restrict *outp, size_t *outleft)
380 if (!isbase64 (in[0]) || !isbase64 (in[1]))
385 *out++ = ((b64[to_uchar (in[0])] << 2)
386 | (b64[to_uchar (in[1])] >> 4));
403 if (!isbase64 (in[2]))
408 *out++ = (((b64[to_uchar (in[1])] << 4) & 0xf0)
409 | (b64[to_uchar (in[2])] >> 2));
423 if (!isbase64 (in[3]))
428 *out++ = (((b64[to_uchar (in[2])] << 6) & 0xc0)
429 | b64[to_uchar (in[3])]);
439 /* Decode base64-encoded input array IN of length INLEN to output array
440 OUT that can hold *OUTLEN bytes. The input data may be interspersed
441 with newlines. Return true if decoding was successful, i.e. if the
442 input was valid base64 data, false otherwise. If *OUTLEN is too
443 small, as many bytes as possible will be written to OUT. On return,
444 *OUTLEN holds the length of decoded bytes in OUT. Note that as soon
445 as any non-alphabet, non-newline character is encountered, decoding
446 is stopped and false is returned. If INLEN is zero, then process
447 only whatever data is stored in CTX.
449 Initially, CTX must have been initialized via base64_decode_ctx_init.
450 Subsequent calls to this function must reuse whatever state is recorded
451 in that buffer. It is necessary for when a quadruple of base64 input
452 bytes spans two input buffers.
454 If CTX is NULL then newlines are treated as garbage and the input
455 buffer is processed as a unit. */
458 base64_decode_ctx (struct base64_decode_context *ctx,
459 const char *restrict in, size_t inlen,
460 char *restrict out, size_t *outlen)
462 size_t outleft = *outlen;
463 bool ignore_newlines = ctx != NULL;
464 bool flush_ctx = false;
465 unsigned int ctx_i = 0;
470 flush_ctx = inlen == 0;
476 size_t outleft_save = outleft;
477 if (ctx_i == 0 && !flush_ctx)
481 /* Save a copy of outleft, in case we need to re-parse this
482 block of four bytes. */
483 outleft_save = outleft;
484 if (!decode_4 (in, inlen, &out, &outleft))
492 if (inlen == 0 && !flush_ctx)
495 /* Handle the common case of 72-byte wrapped lines.
496 This also handles any other multiple-of-4-byte wrapping. */
497 if (inlen && *in == '\n' && ignore_newlines)
504 /* Restore OUT and OUTLEFT. */
505 out -= outleft_save - outleft;
506 outleft = outleft_save;
509 char const *in_end = in + inlen;
513 non_nl = get_4 (ctx, &in, in_end, &inlen);
515 non_nl = in; /* Might have nl in this case. */
517 /* If the input is empty or consists solely of newlines (0 non-newlines),
518 then we're done. Likewise if there are fewer than 4 bytes when not
519 flushing context and not treating newlines as garbage. */
520 if (inlen == 0 || (inlen < 4 && !flush_ctx && ignore_newlines))
525 if (!decode_4 (non_nl, inlen, &out, &outleft))
537 /* Allocate an output buffer in *OUT, and decode the base64 encoded
538 data stored in IN of size INLEN to the *OUT buffer. On return, the
539 size of the decoded data is stored in *OUTLEN. OUTLEN may be NULL,
540 if the caller is not interested in the decoded length. *OUT may be
541 NULL to indicate an out of memory error, in which case *OUTLEN
542 contains the size of the memory block needed. The function returns
543 true on successful decoding and memory allocation errors. (Use the
544 *OUT and *OUTLEN parameters to differentiate between successful
545 decoding and memory error.) The function returns false if the
546 input was invalid, in which case *OUT is NULL and *OUTLEN is
549 base64_decode_alloc_ctx (struct base64_decode_context *ctx,
550 const char *in, size_t inlen, char **out,
553 /* This may allocate a few bytes too many, depending on input,
554 but it's not worth the extra CPU time to compute the exact size.
555 The exact size is 3 * inlen / 4, minus 1 if the input ends
556 with "=" and minus another 1 if the input ends with "==".
557 Dividing before multiplying avoids the possibility of overflow. */
558 size_t needlen = 3 * (inlen / 4) + 2;
560 *out = malloc (needlen);
564 if (!base64_decode_ctx (ctx, in, inlen, *out, &needlen))