xref: /PHP-8.1/ext/fileinfo/libmagic/ascmagic.c (revision 3b9173dc)
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 -- try to detect text encoding.
30  *
31  * Extensively modified by Eric Fischer <enf@pobox.com> in July, 2000,
32  * to handle character codes other than ASCII on a unified basis.
33  */
34 
35 #include "file.h"
36 
37 #ifndef	lint
38 FILE_RCSID("@(#)$File: ascmagic.c,v 1.109 2021/02/05 23:01:40 christos Exp $")
39 #endif	/* lint */
40 
41 #include "magic.h"
42 #include <string.h>
43 #include <ctype.h>
44 #include <stdlib.h>
45 #ifdef HAVE_UNISTD_H
46 #include <unistd.h>
47 #endif
48 
49 #define MAXLINELEN 300	/* longest sane line length */
50 #define ISSPC(x) ((x) == ' ' || (x) == '\t' || (x) == '\r' || (x) == '\n' \
51 		  || (x) == 0x85 || (x) == '\f')
52 
53 private unsigned char *encode_utf8(unsigned char *, size_t, file_unichar_t *,
54     size_t);
55 private size_t trim_nuls(const unsigned char *, size_t);
56 
57 /*
58  * Undo the NUL-termination kindly provided by process()
59  * but leave at least one byte to look at
60  */
61 private size_t
trim_nuls(const unsigned char * buf,size_t nbytes)62 trim_nuls(const unsigned char *buf, size_t nbytes)
63 {
64 	while (nbytes > 1 && buf[nbytes - 1] == '\0')
65 		nbytes--;
66 
67 	return nbytes;
68 }
69 
70 protected int
file_ascmagic(struct magic_set * ms,const struct buffer * b,int text)71 file_ascmagic(struct magic_set *ms, const struct buffer *b, int text)
72 {
73 	file_unichar_t *ubuf = NULL;
74 	size_t ulen = 0;
75 	int rv = 1;
76 	struct buffer bb;
77 
78 	const char *code = NULL;
79 	const char *code_mime = NULL;
80 	const char *type = NULL;
81 
82 	bb = *b;
83 	bb.flen = trim_nuls(CAST(const unsigned char *, b->fbuf), b->flen);
84 	/*
85 	 * Avoid trimming at an odd byte if the original buffer was evenly
86 	 * sized; this avoids losing the last character on UTF-16 LE text
87 	 */
88 	if ((bb.flen & 1) && !(b->flen & 1))
89 		bb.flen++;
90 
91 	/* If file doesn't look like any sort of text, give up. */
92 	if (file_encoding(ms, &bb, &ubuf, &ulen, &code, &code_mime,
93 	    &type) == 0)
94 		rv = 0;
95         else
96 		rv = file_ascmagic_with_encoding(ms, &bb,
97 		    ubuf, ulen, code, type, text);
98 
99 	efree(ubuf);
100 
101 	return rv;
102 }
103 
104 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)105 file_ascmagic_with_encoding(struct magic_set *ms, const struct buffer *b,
106     file_unichar_t *ubuf, size_t ulen, const char *code, const char *type,
107     int text)
108 {
109 	struct buffer bb;
110 	const unsigned char *buf = CAST(const unsigned char *, b->fbuf);
111 	size_t nbytes = b->flen;
112 	unsigned char *utf8_buf = NULL, *utf8_end;
113 	size_t mlen, i, len;
114 	int rv = -1;
115 	int mime = ms->flags & MAGIC_MIME;
116 	int need_separator = 0;
117 
118 	const char *subtype = 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 executable = 0;
129 
130 	size_t last_line_end = CAST(size_t, -1);
131 	size_t has_long_lines = 0;
132 
133 	nbytes = trim_nuls(buf, nbytes);
134 
135 	/* If we have fewer than 2 bytes, give up. */
136 	if (nbytes <= 1) {
137 		rv = 0;
138 		goto done;
139 	}
140 
141 	if (ulen > 0 && (ms->flags & MAGIC_NO_CHECK_SOFT) == 0) {
142 		/* Convert ubuf to UTF-8 and try text soft magic */
143 		/* malloc size is a conservative overestimate; could be
144 		   improved, or at least realloced after conversion. */
145 		mlen = ulen * 6;
146 		if ((utf8_buf = CAST(unsigned char *, emalloc(mlen))) == NULL) {
147 			file_oomem(ms, mlen);
148 			goto done;
149 		}
150 		if ((utf8_end = encode_utf8(utf8_buf, mlen, ubuf, ulen))
151 		    == NULL)
152 			goto done;
153 		buffer_init(&bb, b->fd, &b->st, utf8_buf,
154 		    CAST(size_t, utf8_end - utf8_buf));
155 
156 		if ((rv = file_softmagic(ms, &bb, NULL, NULL,
157 		    TEXTTEST, text)) == 0)
158 			rv = -1;
159 		else
160 			need_separator = 1;
161 		buffer_fini(&bb);
162 		if ((ms->flags & (MAGIC_APPLE|MAGIC_EXTENSION))) {
163 			rv = rv == -1 ? 0 : 1;
164 			goto done;
165 		}
166 	}
167 
168 	if ((ms->flags & (MAGIC_APPLE|MAGIC_EXTENSION))) {
169 		rv = 0;
170 		goto done;
171 	}
172 
173 	/* Now try to discover other details about the file. */
174 	for (i = 0; i < ulen; i++) {
175 		if (ubuf[i] == '\n') {
176 			if (seen_cr)
177 				n_crlf++;
178 			else
179 				n_lf++;
180 			last_line_end = i;
181 		} else if (seen_cr)
182 			n_cr++;
183 
184 		seen_cr = (ubuf[i] == '\r');
185 		if (seen_cr)
186 			last_line_end = i;
187 
188 		if (ubuf[i] == 0x85) { /* X3.64/ECMA-43 "next line" character */
189 			n_nel++;
190 			last_line_end = i;
191 		}
192 
193 		/* If this line is _longer_ than MAXLINELEN, remember it. */
194 		if (i > last_line_end + MAXLINELEN) {
195 			size_t ll = i - last_line_end;
196 			if (ll > has_long_lines)
197 				has_long_lines = ll;
198 		}
199 
200 		if (ubuf[i] == '\033')
201 			has_escapes = 1;
202 		if (ubuf[i] == '\b')
203 			has_backspace = 1;
204 	}
205 
206 	/* Beware, if the data has been truncated, the final CR could have
207 	   been followed by a LF.  If we have ms->bytes_max bytes, it indicates
208 	   that the data might have been truncated, probably even before
209 	   this function was called. */
210 	if (seen_cr && nbytes < ms->bytes_max)
211 		n_cr++;
212 
213 	if (strcmp(type, "binary") == 0) {
214 		rv = 0;
215 		goto done;
216 	}
217 	len = file_printedlen(ms);
218 	if (mime) {
219 		if ((mime & MAGIC_MIME_TYPE) != 0) {
220 			if (len) {
221 				/*
222 				 * Softmagic printed something, we
223 				 * are either done, or we need a separator
224 				 */
225 				if ((ms->flags & MAGIC_CONTINUE) == 0) {
226 					rv = 1;
227 					goto done;
228 				}
229 				if (need_separator && file_separator(ms) == -1)
230 					goto done;
231 			} else {
232 				if (file_printf(ms, "text/plain") == -1)
233 					goto done;
234 			}
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 (%zu)",
277 			    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 	if (utf8_buf)
334 		efree(utf8_buf);
335 
336 	return rv;
337 }
338 
339 /*
340  * Encode Unicode string as UTF-8, returning pointer to character
341  * after end of string, or NULL if an invalid character is found.
342  */
343 private unsigned char *
encode_utf8(unsigned char * buf,size_t len,file_unichar_t * ubuf,size_t ulen)344 encode_utf8(unsigned char *buf, size_t len, file_unichar_t *ubuf, size_t ulen)
345 {
346 	size_t i;
347 	unsigned char *end = buf + len;
348 
349 	for (i = 0; i < ulen; i++) {
350 		if (ubuf[i] <= 0x7f) {
351 			if (end - buf < 1)
352 				return NULL;
353 			*buf++ = CAST(unsigned char, ubuf[i]);
354 			continue;
355 		}
356 		if (ubuf[i] <= 0x7ff) {
357 			if (end - buf < 2)
358 				return NULL;
359 			*buf++ = CAST(unsigned char, (ubuf[i] >> 6) + 0xc0);
360 			goto out1;
361 		}
362 		if (ubuf[i] <= 0xffff) {
363 			if (end - buf < 3)
364 				return NULL;
365 			*buf++ = CAST(unsigned char, (ubuf[i] >> 12) + 0xe0);
366 			goto out2;
367 		}
368 		if (ubuf[i] <= 0x1fffff) {
369 			if (end - buf < 4)
370 				return NULL;
371 			*buf++ = CAST(unsigned char, (ubuf[i] >> 18) + 0xf0);
372 			goto out3;
373 		}
374 		if (ubuf[i] <= 0x3ffffff) {
375 			if (end - buf < 5)
376 				return NULL;
377 			*buf++ = CAST(unsigned char, (ubuf[i] >> 24) + 0xf8);
378 			goto out4;
379 		}
380 		if (ubuf[i] <= 0x7fffffff) {
381 			if (end - buf < 6)
382 				return NULL;
383 			*buf++ = CAST(unsigned char, (ubuf[i] >> 30) + 0xfc);
384 			goto out5;
385 		}
386 		/* Invalid character */
387 		return NULL;
388 	out5:	*buf++ = CAST(unsigned char, ((ubuf[i] >> 24) & 0x3f) + 0x80);
389 	out4:	*buf++ = CAST(unsigned char, ((ubuf[i] >> 18) & 0x3f) + 0x80);
390 	out3:	*buf++ = CAST(unsigned char, ((ubuf[i] >> 12) & 0x3f) + 0x80);
391 	out2:	*buf++ = CAST(unsigned char, ((ubuf[i] >>  6) & 0x3f) + 0x80);
392 	out1:	*buf++ = CAST(unsigned char, ((ubuf[i] >>  0) & 0x3f) + 0x80);
393 	}
394 
395 	return buf;
396 }
397