1 /* $NetBSD: ascmagic.c,v 1.1.1.16 2023/08/18 18:36:49 christos Exp $ */
2
3 /*
4 * Copyright (c) Ian F. Darwin 1986-1995.
5 * Software written by Ian F. Darwin and others;
6 * maintained 1995-present by Christos Zoulas and others.
7 *
8 * Redistribution and use in source and binary forms, with or without
9 * modification, are permitted provided that the following conditions
10 * are met:
11 * 1. Redistributions of source code must retain the above copyright
12 * notice immediately at the beginning of the file, without modification,
13 * this list of conditions, and the following disclaimer.
14 * 2. Redistributions in binary form must reproduce the above copyright
15 * notice, this list of conditions and the following disclaimer in the
16 * documentation and/or other materials provided with the distribution.
17 *
18 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
19 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
20 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
21 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE FOR
22 * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
23 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
24 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
25 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
26 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
27 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
28 * SUCH DAMAGE.
29 */
30 /*
31 * ASCII magic -- try to detect text encoding.
32 *
33 * Extensively modified by Eric Fischer <enf@pobox.com> in July, 2000,
34 * to handle character codes other than ASCII on a unified basis.
35 */
36
37 #include "file.h"
38
39 #ifndef lint
40 #if 0
41 FILE_RCSID("@(#)$File: ascmagic.c,v 1.116 2023/05/21 16:08:50 christos Exp $")
42 #else
43 __RCSID("$NetBSD: ascmagic.c,v 1.1.1.16 2023/08/18 18:36:49 christos Exp $");
44 #endif
45 #endif /* lint */
46
47 #include "magic.h"
48 #include <string.h>
49 #include <ctype.h>
50 #include <stdlib.h>
51 #ifdef HAVE_UNISTD_H
52 #include <unistd.h>
53 #endif
54
55 #define MAXLINELEN 300 /* longest sane line length */
56 #define ISSPC(x) ((x) == ' ' || (x) == '\t' || (x) == '\r' || (x) == '\n' \
57 || (x) == 0x85 || (x) == '\f')
58
59 file_private unsigned char *encode_utf8(unsigned char *, size_t, file_unichar_t *,
60 size_t);
61 file_private size_t trim_nuls(const unsigned char *, size_t);
62
63 /*
64 * Undo the NUL-termination kindly provided by process()
65 * but leave at least one byte to look at
66 */
67 file_private size_t
trim_nuls(const unsigned char * buf,size_t nbytes)68 trim_nuls(const unsigned char *buf, size_t nbytes)
69 {
70 while (nbytes > 1 && buf[nbytes - 1] == '\0')
71 nbytes--;
72
73 return nbytes;
74 }
75
76 file_protected int
file_ascmagic(struct magic_set * ms,const struct buffer * b,int text)77 file_ascmagic(struct magic_set *ms, const struct buffer *b, int text)
78 {
79 file_unichar_t *ubuf = NULL;
80 size_t ulen = 0;
81 int rv = 1;
82 struct buffer bb;
83
84 const char *code = NULL;
85 const char *code_mime = NULL;
86 const char *type = NULL;
87
88 bb = *b;
89 bb.flen = trim_nuls(CAST(const unsigned char *, b->fbuf), b->flen);
90 /*
91 * Avoid trimming at an odd byte if the original buffer was evenly
92 * sized; this avoids losing the last character on UTF-16 LE text
93 */
94 if ((bb.flen & 1) && !(b->flen & 1))
95 bb.flen++;
96
97 /* If file doesn't look like any sort of text, give up. */
98 if (file_encoding(ms, &bb, &ubuf, &ulen, &code, &code_mime,
99 &type) == 0)
100 rv = 0;
101 else
102 rv = file_ascmagic_with_encoding(ms, &bb,
103 ubuf, ulen, code, type, text);
104
105 free(ubuf);
106
107 return rv;
108 }
109
110 file_protected int
file_ascmagic_with_encoding(struct magic_set * ms,const struct buffer * b,file_unichar_t * ubuf,size_t ulen,const char * code,const char * type,int text)111 file_ascmagic_with_encoding(struct magic_set *ms, const struct buffer *b,
112 file_unichar_t *ubuf, size_t ulen, const char *code, const char *type,
113 int text)
114 {
115 struct buffer bb;
116 const unsigned char *buf = CAST(const unsigned char *, b->fbuf);
117 size_t nbytes = b->flen;
118 unsigned char *utf8_buf = NULL, *utf8_end;
119 size_t mlen, i, len;
120 int rv = -1;
121 int mime = ms->flags & MAGIC_MIME;
122 int need_separator = 0;
123
124 const char *subtype = NULL;
125
126 int has_escapes = 0;
127 int has_backspace = 0;
128 int seen_cr = 0;
129
130 size_t n_crlf = 0;
131 size_t n_lf = 0;
132 size_t n_cr = 0;
133 size_t n_nel = 0;
134 int executable = 0;
135
136 size_t last_line_end = CAST(size_t, -1);
137 size_t has_long_lines = 0;
138
139 nbytes = trim_nuls(buf, nbytes);
140
141 /* If we have fewer than 2 bytes, give up. */
142 if (nbytes <= 1) {
143 rv = 0;
144 goto done;
145 }
146
147 if (ulen > 0 && (ms->flags & MAGIC_NO_CHECK_SOFT) == 0) {
148 /* Convert ubuf to UTF-8 and try text soft magic */
149 /* malloc size is a conservative overestimate; could be
150 improved, or at least realloced after conversion. */
151 mlen = ulen * 6;
152 if ((utf8_buf = CAST(unsigned char *, malloc(mlen))) == NULL) {
153 file_oomem(ms, mlen);
154 goto done;
155 }
156 if ((utf8_end = encode_utf8(utf8_buf, mlen, ubuf, ulen))
157 == NULL) {
158 rv = 0;
159 goto done;
160 }
161 buffer_init(&bb, b->fd, &b->st, utf8_buf,
162 CAST(size_t, utf8_end - utf8_buf));
163
164 if ((rv = file_softmagic(ms, &bb, NULL, NULL,
165 TEXTTEST, text)) == 0)
166 rv = -1;
167 else
168 need_separator = 1;
169 buffer_fini(&bb);
170 if ((ms->flags & (MAGIC_APPLE|MAGIC_EXTENSION))) {
171 rv = rv == -1 ? 0 : 1;
172 goto done;
173 }
174 }
175
176 if ((ms->flags & (MAGIC_APPLE|MAGIC_EXTENSION))) {
177 rv = 0;
178 goto done;
179 }
180
181 /* Now try to discover other details about the file. */
182 for (i = 0; i < ulen; i++) {
183 if (ubuf[i] == '\n') {
184 if (seen_cr)
185 n_crlf++;
186 else
187 n_lf++;
188 last_line_end = i;
189 } else if (seen_cr)
190 n_cr++;
191
192 seen_cr = (ubuf[i] == '\r');
193 if (seen_cr)
194 last_line_end = i;
195
196 if (ubuf[i] == 0x85) { /* X3.64/ECMA-43 "next line" character */
197 n_nel++;
198 last_line_end = i;
199 }
200
201 /* If this line is _longer_ than MAXLINELEN, remember it. */
202 if (i > last_line_end + MAXLINELEN) {
203 size_t ll = i - last_line_end;
204 if (ll > has_long_lines)
205 has_long_lines = ll;
206 }
207
208 if (ubuf[i] == '\033')
209 has_escapes = 1;
210 if (ubuf[i] == '\b')
211 has_backspace = 1;
212 }
213
214 if (strcmp(type, "binary") == 0) {
215 rv = 0;
216 goto done;
217 }
218 len = file_printedlen(ms);
219 if (mime) {
220 if ((mime & MAGIC_MIME_TYPE) != 0) {
221 if (len) {
222 /*
223 * Softmagic printed something, we
224 * are either done, or we need a separator
225 */
226 if ((ms->flags & MAGIC_CONTINUE) == 0) {
227 rv = 1;
228 goto done;
229 }
230 if (need_separator && file_separator(ms) == -1)
231 goto done;
232 }
233 if (file_printf(ms, "text/plain") == -1)
234 goto done;
235 }
236 } else {
237 if (len) {
238 switch (file_replace(ms, " text$", ", ")) {
239 case 0:
240 switch (file_replace(ms, " text executable$",
241 ", ")) {
242 case 0:
243 if (file_printf(ms, ", ") == -1)
244 goto done;
245 break;
246 case -1:
247 goto done;
248 default:
249 executable = 1;
250 break;
251 }
252 break;
253 case -1:
254 goto done;
255 default:
256 break;
257 }
258 }
259
260 if (file_printf(ms, "%s", code) == -1)
261 goto done;
262
263 if (subtype) {
264 if (file_printf(ms, " %s", subtype) == -1)
265 goto done;
266 }
267
268 if (file_printf(ms, " %s", type) == -1)
269 goto done;
270
271 if (executable)
272 if (file_printf(ms, " executable") == -1)
273 goto done;
274
275 if (has_long_lines)
276 if (file_printf(ms, ", with very long lines (%"
277 SIZE_T_FORMAT "u)", has_long_lines) == -1)
278 goto done;
279
280 /*
281 * Only report line terminators if we find one other than LF,
282 * or if we find none at all.
283 */
284 if ((n_crlf == 0 && n_cr == 0 && n_nel == 0 && n_lf == 0) ||
285 (n_crlf != 0 || n_cr != 0 || n_nel != 0)) {
286 if (file_printf(ms, ", with") == -1)
287 goto done;
288
289 if (n_crlf == 0 && n_cr == 0 &&
290 n_nel == 0 && n_lf == 0) {
291 if (file_printf(ms, " no") == -1)
292 goto done;
293 } else {
294 if (n_crlf) {
295 if (file_printf(ms, " CRLF") == -1)
296 goto done;
297 if (n_cr || n_lf || n_nel)
298 if (file_printf(ms, ",") == -1)
299 goto done;
300 }
301 if (n_cr) {
302 if (file_printf(ms, " CR") == -1)
303 goto done;
304 if (n_lf || n_nel)
305 if (file_printf(ms, ",") == -1)
306 goto done;
307 }
308 if (n_lf) {
309 if (file_printf(ms, " LF") == -1)
310 goto done;
311 if (n_nel)
312 if (file_printf(ms, ",") == -1)
313 goto done;
314 }
315 if (n_nel)
316 if (file_printf(ms, " NEL") == -1)
317 goto done;
318 }
319
320 if (file_printf(ms, " line terminators") == -1)
321 goto done;
322 }
323
324 if (has_escapes)
325 if (file_printf(ms, ", with escape sequences") == -1)
326 goto done;
327 if (has_backspace)
328 if (file_printf(ms, ", with overstriking") == -1)
329 goto done;
330 }
331 rv = 1;
332 done:
333 free(utf8_buf);
334
335 return rv;
336 }
337
338 /*
339 * Encode Unicode string as UTF-8, returning pointer to character
340 * after end of string, or NULL if an invalid character is found.
341 */
342 file_private unsigned char *
encode_utf8(unsigned char * buf,size_t len,file_unichar_t * ubuf,size_t ulen)343 encode_utf8(unsigned char *buf, size_t len, file_unichar_t *ubuf, size_t ulen)
344 {
345 size_t i;
346 unsigned char *end = buf + len;
347
348 for (i = 0; i < ulen; i++) {
349 if (ubuf[i] <= 0x7f) {
350 if (end - buf < 1)
351 return NULL;
352 *buf++ = CAST(unsigned char, ubuf[i]);
353 continue;
354 }
355 if (ubuf[i] <= 0x7ff) {
356 if (end - buf < 2)
357 return NULL;
358 *buf++ = CAST(unsigned char, (ubuf[i] >> 6) + 0xc0);
359 goto out1;
360 }
361 if (ubuf[i] <= 0xffff) {
362 if (end - buf < 3)
363 return NULL;
364 *buf++ = CAST(unsigned char, (ubuf[i] >> 12) + 0xe0);
365 goto out2;
366 }
367 if (ubuf[i] <= 0x1fffff) {
368 if (end - buf < 4)
369 return NULL;
370 *buf++ = CAST(unsigned char, (ubuf[i] >> 18) + 0xf0);
371 goto out3;
372 }
373 if (ubuf[i] <= 0x3ffffff) {
374 if (end - buf < 5)
375 return NULL;
376 *buf++ = CAST(unsigned char, (ubuf[i] >> 24) + 0xf8);
377 goto out4;
378 }
379 if (ubuf[i] <= 0x7fffffff) {
380 if (end - buf < 6)
381 return NULL;
382 *buf++ = CAST(unsigned char, (ubuf[i] >> 30) + 0xfc);
383 goto out5;
384 }
385 /* Invalid character */
386 return NULL;
387 out5: *buf++ = CAST(unsigned char, ((ubuf[i] >> 24) & 0x3f) + 0x80);
388 out4: *buf++ = CAST(unsigned char, ((ubuf[i] >> 18) & 0x3f) + 0x80);
389 out3: *buf++ = CAST(unsigned char, ((ubuf[i] >> 12) & 0x3f) + 0x80);
390 out2: *buf++ = CAST(unsigned char, ((ubuf[i] >> 6) & 0x3f) + 0x80);
391 out1: *buf++ = CAST(unsigned char, ((ubuf[i] >> 0) & 0x3f) + 0x80);
392 }
393
394 return buf;
395 }
396