xref: /dflybsd-src/contrib/file/src/ascmagic.c (revision e90a7c45c3303ed54c0fde732b2ba32dc80ffd9b)
1 /*
2  * Copyright (c) Ian F. Darwin 1986-1995.
3  * Software written by Ian F. Darwin and others;
4  * maintained 1995-present by Christos Zoulas and others.
5  *
6  * Redistribution and use in source and binary forms, with or without
7  * modification, are permitted provided that the following conditions
8  * are met:
9  * 1. Redistributions of source code must retain the above copyright
10  *    notice immediately at the beginning of the file, without modification,
11  *    this list of conditions, and the following disclaimer.
12  * 2. Redistributions in binary form must reproduce the above copyright
13  *    notice, this list of conditions and the following disclaimer in the
14  *    documentation and/or other materials provided with the distribution.
15  *
16  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
17  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
18  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
19  * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE FOR
20  * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
21  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
22  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
23  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
24  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
25  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
26  * SUCH DAMAGE.
27  */
28 /*
29  * ASCII magic -- file types that we know based on keywords
30  * that can appear anywhere in the file.
31  *
32  * Extensively modified by Eric Fischer <enf@pobox.com> in July, 2000,
33  * to handle character codes other than ASCII on a unified basis.
34  */
35 
36 #include "file.h"
37 
38 #ifndef	lint
39 FILE_RCSID("@(#)$File: ascmagic.c,v 1.77 2010/11/30 14:58:53 rrt Exp $")
40 #endif	/* lint */
41 
42 #include "magic.h"
43 #include <string.h>
44 #include <memory.h>
45 #include <ctype.h>
46 #include <stdlib.h>
47 #ifdef HAVE_UNISTD_H
48 #include <unistd.h>
49 #endif
50 #include "names.h"
51 
52 #define MAXLINELEN 300	/* longest sane line length */
53 #define ISSPC(x) ((x) == ' ' || (x) == '\t' || (x) == '\r' || (x) == '\n' \
54 		  || (x) == 0x85 || (x) == '\f')
55 
56 private int ascmatch(const unsigned char *, const unichar *, size_t);
57 private unsigned char *encode_utf8(unsigned char *, size_t, unichar *, size_t);
58 private size_t trim_nuls(const unsigned char *, size_t);
59 
60 /*
61  * Undo the NUL-termination kindly provided by process()
62  * but leave at least one byte to look at
63  */
64 private size_t
65 trim_nuls(const unsigned char *buf, size_t nbytes)
66 {
67 	while (nbytes > 1 && buf[nbytes - 1] == '\0')
68 		nbytes--;
69 
70 	return nbytes;
71 }
72 
73 protected int
74 file_ascmagic(struct magic_set *ms, const unsigned char *buf, size_t nbytes)
75 {
76 	unichar *ubuf = NULL;
77 	size_t ulen;
78 	int rv = 1;
79 
80 	const char *code = NULL;
81 	const char *code_mime = NULL;
82 	const char *type = NULL;
83 
84 	if (ms->flags & MAGIC_APPLE)
85 		return 0;
86 
87 	nbytes = trim_nuls(buf, nbytes);
88 
89 	/* If file doesn't look like any sort of text, give up. */
90 	if (file_encoding(ms, buf, nbytes, &ubuf, &ulen, &code, &code_mime,
91 	    &type) == 0) {
92 		rv = 0;
93 		goto done;
94 	}
95 
96 	rv = file_ascmagic_with_encoding(ms, buf, nbytes, ubuf, ulen, code,
97 	    type);
98 
99  done:
100 	if (ubuf)
101 		free(ubuf);
102 
103 	return rv;
104 }
105 
106 protected int
107 file_ascmagic_with_encoding(struct magic_set *ms, const unsigned char *buf,
108     size_t nbytes, unichar *ubuf, size_t ulen, const char *code,
109     const char *type)
110 {
111 	unsigned char *utf8_buf = NULL, *utf8_end;
112 	size_t mlen, i;
113 	const struct names *p;
114 	int rv = -1;
115 	int mime = ms->flags & MAGIC_MIME;
116 
117 	const char *subtype = NULL;
118 	const char *subtype_mime = NULL;
119 
120 	int has_escapes = 0;
121 	int has_backspace = 0;
122 	int seen_cr = 0;
123 
124 	int n_crlf = 0;
125 	int n_lf = 0;
126 	int n_cr = 0;
127 	int n_nel = 0;
128 	int score, curtype;
129 
130 	size_t last_line_end = (size_t)-1;
131 	int has_long_lines = 0;
132 
133 	if (ms->flags & MAGIC_APPLE)
134 		return 0;
135 
136 	nbytes = trim_nuls(buf, nbytes);
137 
138 	/* If we have fewer than 2 bytes, give up. */
139 	if (nbytes <= 1) {
140 		rv = 0;
141 		goto done;
142 	}
143 
144 	if ((ms->flags & MAGIC_NO_CHECK_SOFT) == 0) {
145 		/* Convert ubuf to UTF-8 and try text soft magic */
146 		/* malloc size is a conservative overestimate; could be
147 		   improved, or at least realloced after conversion. */
148 		mlen = ulen * 6;
149 		if ((utf8_buf = CAST(unsigned char *, malloc(mlen))) == NULL) {
150 			file_oomem(ms, mlen);
151 			goto done;
152 		}
153 		if ((utf8_end = encode_utf8(utf8_buf, mlen, ubuf, ulen)) == NULL)
154 			goto done;
155 		if ((rv = file_softmagic(ms, utf8_buf, (size_t)(utf8_end - utf8_buf),
156 					 TEXTTEST)) != 0)
157 			goto done;
158 		else
159 			rv = -1;
160 	}
161 
162 	/* look for tokens from names.h - this is expensive! */
163 	if ((ms->flags & MAGIC_NO_CHECK_TOKENS) != 0)
164 		goto subtype_identified;
165 
166 	i = 0;
167 	score = 0;
168 	curtype = -1;
169 	while (i < ulen) {
170 		size_t end;
171 
172 		/* skip past any leading space */
173 		while (i < ulen && ISSPC(ubuf[i]))
174 			i++;
175 		if (i >= ulen)
176 			break;
177 
178 		/* find the next whitespace */
179 		for (end = i + 1; end < nbytes; end++)
180 			if (ISSPC(ubuf[end]))
181 				break;
182 
183 		/* compare the word thus isolated against the token list */
184 		for (p = names; p < names + NNAMES; p++) {
185 			if (ascmatch((const unsigned char *)p->name, ubuf + i,
186 			    end - i)) {
187 				if (curtype == -1)
188 					curtype = p->type;
189 				else if (curtype != p->type) {
190 					score = p->score;
191 					curtype = p->type;
192 				} else
193 					score += p->score;
194 				if (score > 1) {
195 					subtype = types[p->type].human;
196 					subtype_mime = types[p->type].mime;
197 					goto subtype_identified;
198 				}
199 			}
200 		}
201 
202 		i = end;
203 	}
204 
205 subtype_identified:
206 
207 	/* Now try to discover other details about the file. */
208 	for (i = 0; i < ulen; i++) {
209 		if (ubuf[i] == '\n') {
210 			if (seen_cr)
211 				n_crlf++;
212 			else
213 				n_lf++;
214 			last_line_end = i;
215 		} else if (seen_cr)
216 			n_cr++;
217 
218 		seen_cr = (ubuf[i] == '\r');
219 		if (seen_cr)
220 			last_line_end = i;
221 
222 		if (ubuf[i] == 0x85) { /* X3.64/ECMA-43 "next line" character */
223 			n_nel++;
224 			last_line_end = i;
225 		}
226 
227 		/* If this line is _longer_ than MAXLINELEN, remember it. */
228 		if (i > last_line_end + MAXLINELEN)
229 			has_long_lines = 1;
230 
231 		if (ubuf[i] == '\033')
232 			has_escapes = 1;
233 		if (ubuf[i] == '\b')
234 			has_backspace = 1;
235 	}
236 
237 	/* Beware, if the data has been truncated, the final CR could have
238 	   been followed by a LF.  If we have HOWMANY bytes, it indicates
239 	   that the data might have been truncated, probably even before
240 	   this function was called. */
241 	if (seen_cr && nbytes < HOWMANY)
242 		n_cr++;
243 
244 	if (strcmp(type, "binary") == 0) {
245 		rv = 0;
246 		goto done;
247 	}
248 	if (mime) {
249 		if ((mime & MAGIC_MIME_TYPE) != 0) {
250 			if (subtype_mime) {
251 				if (file_printf(ms, "%s", subtype_mime) == -1)
252 					goto done;
253 			} else {
254 				if (file_printf(ms, "text/plain") == -1)
255 					goto done;
256 			}
257 		}
258 	} else {
259 		if (file_printf(ms, "%s", code) == -1)
260 			goto done;
261 
262 		if (subtype) {
263 			if (file_printf(ms, " %s", subtype) == -1)
264 				goto done;
265 		}
266 
267 		if (file_printf(ms, " %s", type) == -1)
268 			goto done;
269 
270 		if (has_long_lines)
271 			if (file_printf(ms, ", with very long lines") == -1)
272 				goto done;
273 
274 		/*
275 		 * Only report line terminators if we find one other than LF,
276 		 * or if we find none at all.
277 		 */
278 		if ((n_crlf == 0 && n_cr == 0 && n_nel == 0 && n_lf == 0) ||
279 		    (n_crlf != 0 || n_cr != 0 || n_nel != 0)) {
280 			if (file_printf(ms, ", with") == -1)
281 				goto done;
282 
283 			if (n_crlf == 0 && n_cr == 0 && n_nel == 0 && n_lf == 0) {
284 				if (file_printf(ms, " no") == -1)
285 					goto done;
286 			} else {
287 				if (n_crlf) {
288 					if (file_printf(ms, " CRLF") == -1)
289 						goto done;
290 					if (n_cr || n_lf || n_nel)
291 						if (file_printf(ms, ",") == -1)
292 							goto done;
293 				}
294 				if (n_cr) {
295 					if (file_printf(ms, " CR") == -1)
296 						goto done;
297 					if (n_lf || n_nel)
298 						if (file_printf(ms, ",") == -1)
299 							goto done;
300 				}
301 				if (n_lf) {
302 					if (file_printf(ms, " LF") == -1)
303 						goto done;
304 					if (n_nel)
305 						if (file_printf(ms, ",") == -1)
306 							goto done;
307 				}
308 				if (n_nel)
309 					if (file_printf(ms, " NEL") == -1)
310 						goto done;
311 			}
312 
313 			if (file_printf(ms, " line terminators") == -1)
314 				goto done;
315 		}
316 
317 		if (has_escapes)
318 			if (file_printf(ms, ", with escape sequences") == -1)
319 				goto done;
320 		if (has_backspace)
321 			if (file_printf(ms, ", with overstriking") == -1)
322 				goto done;
323 	}
324 	rv = 1;
325 done:
326 	if (utf8_buf)
327 		free(utf8_buf);
328 
329 	return rv;
330 }
331 
332 private int
333 ascmatch(const unsigned char *s, const unichar *us, size_t ulen)
334 {
335 	size_t i;
336 
337 	for (i = 0; i < ulen; i++) {
338 		if (s[i] != us[i])
339 			return 0;
340 	}
341 
342 	if (s[i])
343 		return 0;
344 	else
345 		return 1;
346 }
347 
348 /*
349  * Encode Unicode string as UTF-8, returning pointer to character
350  * after end of string, or NULL if an invalid character is found.
351  */
352 private unsigned char *
353 encode_utf8(unsigned char *buf, size_t len, unichar *ubuf, size_t ulen)
354 {
355 	size_t i;
356 	unsigned char *end = buf + len;
357 
358 	for (i = 0; i < ulen; i++) {
359 		if (ubuf[i] <= 0x7f) {
360 			if (end - buf < 1)
361 				return NULL;
362 			*buf++ = (unsigned char)ubuf[i];
363 		} else if (ubuf[i] <= 0x7ff) {
364 			if (end - buf < 2)
365 				return NULL;
366 			*buf++ = (unsigned char)((ubuf[i] >> 6) + 0xc0);
367 			*buf++ = (unsigned char)((ubuf[i] & 0x3f) + 0x80);
368 		} else if (ubuf[i] <= 0xffff) {
369 			if (end - buf < 3)
370 				return NULL;
371 			*buf++ = (unsigned char)((ubuf[i] >> 12) + 0xe0);
372 			*buf++ = (unsigned char)(((ubuf[i] >> 6) & 0x3f) + 0x80);
373 			*buf++ = (unsigned char)((ubuf[i] & 0x3f) + 0x80);
374 		} else if (ubuf[i] <= 0x1fffff) {
375 			if (end - buf < 4)
376 				return NULL;
377 			*buf++ = (unsigned char)((ubuf[i] >> 18) + 0xf0);
378 			*buf++ = (unsigned char)(((ubuf[i] >> 12) & 0x3f) + 0x80);
379 			*buf++ = (unsigned char)(((ubuf[i] >>  6) & 0x3f) + 0x80);
380 			*buf++ = (unsigned char)((ubuf[i] & 0x3f) + 0x80);
381 		} else if (ubuf[i] <= 0x3ffffff) {
382 			if (end - buf < 5)
383 				return NULL;
384 			*buf++ = (unsigned char)((ubuf[i] >> 24) + 0xf8);
385 			*buf++ = (unsigned char)(((ubuf[i] >> 18) & 0x3f) + 0x80);
386 			*buf++ = (unsigned char)(((ubuf[i] >> 12) & 0x3f) + 0x80);
387 			*buf++ = (unsigned char)(((ubuf[i] >>  6) & 0x3f) + 0x80);
388 			*buf++ = (unsigned char)((ubuf[i] & 0x3f) + 0x80);
389 		} else if (ubuf[i] <= 0x7fffffff) {
390 			if (end - buf < 6)
391 				return NULL;
392 			*buf++ = (unsigned char)((ubuf[i] >> 30) + 0xfc);
393 			*buf++ = (unsigned char)(((ubuf[i] >> 24) & 0x3f) + 0x80);
394 			*buf++ = (unsigned char)(((ubuf[i] >> 18) & 0x3f) + 0x80);
395 			*buf++ = (unsigned char)(((ubuf[i] >> 12) & 0x3f) + 0x80);
396 			*buf++ = (unsigned char)(((ubuf[i] >>  6) & 0x3f) + 0x80);
397 			*buf++ = (unsigned char)((ubuf[i] & 0x3f) + 0x80);
398 		} else /* Invalid character */
399 			return NULL;
400 	}
401 
402 	return buf;
403 }
404