xref: /PHP-8.3/ext/fileinfo/libmagic/ascmagic.c (revision c2b671cb)
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 1.110 2021/12/06 15:33:00 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 			rv = 0;
153 			goto done;
154 		}
155 		buffer_init(&bb, b->fd, &b->st, utf8_buf,
156 		    CAST(size_t, utf8_end - utf8_buf));
157 
158 		if ((rv = file_softmagic(ms, &bb, NULL, NULL,
159 		    TEXTTEST, text)) == 0)
160 			rv = -1;
161 		else
162 			need_separator = 1;
163 		buffer_fini(&bb);
164 		if ((ms->flags & (MAGIC_APPLE|MAGIC_EXTENSION))) {
165 			rv = rv == -1 ? 0 : 1;
166 			goto done;
167 		}
168 	}
169 
170 	if ((ms->flags & (MAGIC_APPLE|MAGIC_EXTENSION))) {
171 		rv = 0;
172 		goto done;
173 	}
174 
175 	/* Now try to discover other details about the file. */
176 	for (i = 0; i < ulen; i++) {
177 		if (ubuf[i] == '\n') {
178 			if (seen_cr)
179 				n_crlf++;
180 			else
181 				n_lf++;
182 			last_line_end = i;
183 		} else if (seen_cr)
184 			n_cr++;
185 
186 		seen_cr = (ubuf[i] == '\r');
187 		if (seen_cr)
188 			last_line_end = i;
189 
190 		if (ubuf[i] == 0x85) { /* X3.64/ECMA-43 "next line" character */
191 			n_nel++;
192 			last_line_end = i;
193 		}
194 
195 		/* If this line is _longer_ than MAXLINELEN, remember it. */
196 		if (i > last_line_end + MAXLINELEN) {
197 			size_t ll = i - last_line_end;
198 			if (ll > has_long_lines)
199 				has_long_lines = ll;
200 		}
201 
202 		if (ubuf[i] == '\033')
203 			has_escapes = 1;
204 		if (ubuf[i] == '\b')
205 			has_backspace = 1;
206 	}
207 
208 	/* Beware, if the data has been truncated, the final CR could have
209 	   been followed by a LF.  If we have ms->bytes_max bytes, it indicates
210 	   that the data might have been truncated, probably even before
211 	   this function was called. */
212 	if (seen_cr && nbytes < ms->bytes_max)
213 		n_cr++;
214 
215 	if (strcmp(type, "binary") == 0) {
216 		rv = 0;
217 		goto done;
218 	}
219 	len = file_printedlen(ms);
220 	if (mime) {
221 		if ((mime & MAGIC_MIME_TYPE) != 0) {
222 			if (len) {
223 				/*
224 				 * Softmagic printed something, we
225 				 * are either done, or we need a separator
226 				 */
227 				if ((ms->flags & MAGIC_CONTINUE) == 0) {
228 					rv = 1;
229 					goto done;
230 				}
231 				if (need_separator && file_separator(ms) == -1)
232 					goto done;
233 			} else {
234 				if (file_printf(ms, "text/plain") == -1)
235 					goto done;
236 			}
237 		}
238 	} else {
239 		if (len) {
240 			switch (file_replace(ms, " text$", ", ")) {
241 			case 0:
242 				switch (file_replace(ms, " text executable$",
243 				    ", ")) {
244 				case 0:
245 					if (file_printf(ms, ", ") == -1)
246 						goto done;
247 					break;
248 				case -1:
249 					goto done;
250 				default:
251 					executable = 1;
252 					break;
253 				}
254 				break;
255 			case -1:
256 				goto done;
257 			default:
258 				break;
259 			}
260 		}
261 
262 		if (file_printf(ms, "%s", code) == -1)
263 			goto done;
264 
265 		if (subtype) {
266 			if (file_printf(ms, " %s", subtype) == -1)
267 				goto done;
268 		}
269 
270 		if (file_printf(ms, " %s", type) == -1)
271 			goto done;
272 
273 		if (executable)
274 			if (file_printf(ms, " executable") == -1)
275 				goto done;
276 
277 		if (has_long_lines)
278 			if (file_printf(ms, ", with very long lines (%zu)",
279 			    has_long_lines) == -1)
280 				goto done;
281 
282 		/*
283 		 * Only report line terminators if we find one other than LF,
284 		 * or if we find none at all.
285 		 */
286 		if ((n_crlf == 0 && n_cr == 0 && n_nel == 0 && n_lf == 0) ||
287 		    (n_crlf != 0 || n_cr != 0 || n_nel != 0)) {
288 			if (file_printf(ms, ", with") == -1)
289 				goto done;
290 
291 			if (n_crlf == 0 && n_cr == 0 &&
292 			    n_nel == 0 && n_lf == 0) {
293 				if (file_printf(ms, " no") == -1)
294 					goto done;
295 			} else {
296 				if (n_crlf) {
297 					if (file_printf(ms, " CRLF") == -1)
298 						goto done;
299 					if (n_cr || n_lf || n_nel)
300 						if (file_printf(ms, ",") == -1)
301 							goto done;
302 				}
303 				if (n_cr) {
304 					if (file_printf(ms, " CR") == -1)
305 						goto done;
306 					if (n_lf || n_nel)
307 						if (file_printf(ms, ",") == -1)
308 							goto done;
309 				}
310 				if (n_lf) {
311 					if (file_printf(ms, " LF") == -1)
312 						goto done;
313 					if (n_nel)
314 						if (file_printf(ms, ",") == -1)
315 							goto done;
316 				}
317 				if (n_nel)
318 					if (file_printf(ms, " NEL") == -1)
319 						goto done;
320 			}
321 
322 			if (file_printf(ms, " line terminators") == -1)
323 				goto done;
324 		}
325 
326 		if (has_escapes)
327 			if (file_printf(ms, ", with escape sequences") == -1)
328 				goto done;
329 		if (has_backspace)
330 			if (file_printf(ms, ", with overstriking") == -1)
331 				goto done;
332 	}
333 	rv = 1;
334 done:
335 	if (utf8_buf)
336 		efree(utf8_buf);
337 
338 	return rv;
339 }
340 
341 /*
342  * Encode Unicode string as UTF-8, returning pointer to character
343  * after end of string, or NULL if an invalid character is found.
344  */
345 private unsigned char *
encode_utf8(unsigned char * buf,size_t len,file_unichar_t * ubuf,size_t ulen)346 encode_utf8(unsigned char *buf, size_t len, file_unichar_t *ubuf, size_t ulen)
347 {
348 	size_t i;
349 	unsigned char *end = buf + len;
350 
351 	for (i = 0; i < ulen; i++) {
352 		if (ubuf[i] <= 0x7f) {
353 			if (end - buf < 1)
354 				return NULL;
355 			*buf++ = CAST(unsigned char, ubuf[i]);
356 			continue;
357 		}
358 		if (ubuf[i] <= 0x7ff) {
359 			if (end - buf < 2)
360 				return NULL;
361 			*buf++ = CAST(unsigned char, (ubuf[i] >> 6) + 0xc0);
362 			goto out1;
363 		}
364 		if (ubuf[i] <= 0xffff) {
365 			if (end - buf < 3)
366 				return NULL;
367 			*buf++ = CAST(unsigned char, (ubuf[i] >> 12) + 0xe0);
368 			goto out2;
369 		}
370 		if (ubuf[i] <= 0x1fffff) {
371 			if (end - buf < 4)
372 				return NULL;
373 			*buf++ = CAST(unsigned char, (ubuf[i] >> 18) + 0xf0);
374 			goto out3;
375 		}
376 		if (ubuf[i] <= 0x3ffffff) {
377 			if (end - buf < 5)
378 				return NULL;
379 			*buf++ = CAST(unsigned char, (ubuf[i] >> 24) + 0xf8);
380 			goto out4;
381 		}
382 		if (ubuf[i] <= 0x7fffffff) {
383 			if (end - buf < 6)
384 				return NULL;
385 			*buf++ = CAST(unsigned char, (ubuf[i] >> 30) + 0xfc);
386 			goto out5;
387 		}
388 		/* Invalid character */
389 		return NULL;
390 	out5:	*buf++ = CAST(unsigned char, ((ubuf[i] >> 24) & 0x3f) + 0x80);
391 	out4:	*buf++ = CAST(unsigned char, ((ubuf[i] >> 18) & 0x3f) + 0x80);
392 	out3:	*buf++ = CAST(unsigned char, ((ubuf[i] >> 12) & 0x3f) + 0x80);
393 	out2:	*buf++ = CAST(unsigned char, ((ubuf[i] >>  6) & 0x3f) + 0x80);
394 	out1:	*buf++ = CAST(unsigned char, ((ubuf[i] >>  0) & 0x3f) + 0x80);
395 	}
396 
397 	return buf;
398 }
399