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
38FILE_RCSID("@(#)$File: ascmagic.c,v 1.107 2020/06/08 19:58:36 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
53private unsigned char *encode_utf8(unsigned char *, size_t, unichar *, size_t);
54private size_t trim_nuls(const unsigned char *, size_t);
55
56/*
57 * Undo the NUL-termination kindly provided by process()
58 * but leave at least one byte to look at
59 */
60private size_t
61trim_nuls(const unsigned char *buf, size_t nbytes)
62{
63	while (nbytes > 1 && buf[nbytes - 1] == '\0')
64		nbytes--;
65
66	return nbytes;
67}
68
69protected int
70file_ascmagic(struct magic_set *ms, const struct buffer *b, int text)
71{
72	unichar *ubuf = NULL;
73	size_t ulen = 0;
74	int rv = 1;
75	struct buffer bb;
76
77	const char *code = NULL;
78	const char *code_mime = NULL;
79	const char *type = NULL;
80
81	bb = *b;
82	bb.flen = trim_nuls(CAST(const unsigned char *, b->fbuf), b->flen);
83	/*
84	 * Avoid trimming at an odd byte if the original buffer was evenly
85	 * sized; this avoids losing the last character on UTF-16 LE text
86	 */
87	if ((bb.flen & 1) && !(b->flen & 1))
88		bb.flen++;
89
90	/* If file doesn't look like any sort of text, give up. */
91	if (file_encoding(ms, &bb, &ubuf, &ulen, &code, &code_mime,
92	    &type) == 0)
93		rv = 0;
94        else
95		rv = file_ascmagic_with_encoding(ms, &bb,
96		    ubuf, ulen, code, type, text);
97
98	free(ubuf);
99
100	return rv;
101}
102
103protected int
104file_ascmagic_with_encoding(struct magic_set *ms,
105    const struct buffer *b, unichar *ubuf, size_t ulen, const char *code,
106    const char *type, int text)
107{
108	struct buffer bb;
109	const unsigned char *buf = CAST(const unsigned char *, b->fbuf);
110	size_t nbytes = b->flen;
111	unsigned char *utf8_buf = NULL, *utf8_end;
112	size_t mlen, i, len;
113	int rv = -1;
114	int mime = ms->flags & MAGIC_MIME;
115	int need_separator = 0;
116
117	const char *subtype = NULL;
118
119	int has_escapes = 0;
120	int has_backspace = 0;
121	int seen_cr = 0;
122
123	int n_crlf = 0;
124	int n_lf = 0;
125	int n_cr = 0;
126	int n_nel = 0;
127	int executable = 0;
128
129	size_t last_line_end = CAST(size_t, -1);
130	int has_long_lines = 0;
131
132	nbytes = trim_nuls(buf, nbytes);
133
134	/* If we have fewer than 2 bytes, give up. */
135	if (nbytes <= 1) {
136		rv = 0;
137		goto done;
138	}
139
140	if (ulen > 0 && (ms->flags & MAGIC_NO_CHECK_SOFT) == 0) {
141		/* Convert ubuf to UTF-8 and try text soft magic */
142		/* malloc size is a conservative overestimate; could be
143		   improved, or at least realloced after conversion. */
144		mlen = ulen * 6;
145		if ((utf8_buf = CAST(unsigned char *, malloc(mlen))) == NULL) {
146			file_oomem(ms, mlen);
147			goto done;
148		}
149		if ((utf8_end = encode_utf8(utf8_buf, mlen, ubuf, ulen))
150		    == NULL)
151			goto done;
152		buffer_init(&bb, b->fd, &b->st, utf8_buf,
153		    CAST(size_t, utf8_end - utf8_buf));
154
155		if ((rv = file_softmagic(ms, &bb, NULL, NULL,
156		    TEXTTEST, text)) == 0)
157			rv = -1;
158		else
159			need_separator = 1;
160		buffer_fini(&bb);
161		if ((ms->flags & (MAGIC_APPLE|MAGIC_EXTENSION))) {
162			rv = rv == -1 ? 0 : 1;
163			goto done;
164		}
165	}
166
167	if ((ms->flags & (MAGIC_APPLE|MAGIC_EXTENSION))) {
168		rv = 0;
169		goto done;
170	}
171
172	/* Now try to discover other details about the file. */
173	for (i = 0; i < ulen; i++) {
174		if (ubuf[i] == '\n') {
175			if (seen_cr)
176				n_crlf++;
177			else
178				n_lf++;
179			last_line_end = i;
180		} else if (seen_cr)
181			n_cr++;
182
183		seen_cr = (ubuf[i] == '\r');
184		if (seen_cr)
185			last_line_end = i;
186
187		if (ubuf[i] == 0x85) { /* X3.64/ECMA-43 "next line" character */
188			n_nel++;
189			last_line_end = i;
190		}
191
192		/* If this line is _longer_ than MAXLINELEN, remember it. */
193		if (i > last_line_end + MAXLINELEN)
194			has_long_lines = 1;
195
196		if (ubuf[i] == '\033')
197			has_escapes = 1;
198		if (ubuf[i] == '\b')
199			has_backspace = 1;
200	}
201
202	/* Beware, if the data has been truncated, the final CR could have
203	   been followed by a LF.  If we have ms->bytes_max bytes, it indicates
204	   that the data might have been truncated, probably even before
205	   this function was called. */
206	if (seen_cr && nbytes < ms->bytes_max)
207		n_cr++;
208
209	if (strcmp(type, "binary") == 0) {
210		rv = 0;
211		goto done;
212	}
213	len = file_printedlen(ms);
214	if (mime) {
215		if ((mime & MAGIC_MIME_TYPE) != 0) {
216			if (len) {
217				/*
218				 * Softmagic printed something, we
219				 * are either done, or we need a separator
220				 */
221				if ((ms->flags & MAGIC_CONTINUE) == 0) {
222					rv = 1;
223					goto done;
224				}
225				if (need_separator && file_separator(ms) == -1)
226					goto done;
227			} else {
228				if (file_printf(ms, "text/plain") == -1)
229					goto done;
230			}
231		}
232	} else {
233		if (len) {
234			switch (file_replace(ms, " text$", ", ")) {
235			case 0:
236				switch (file_replace(ms, " text executable$",
237				    ", ")) {
238				case 0:
239					if (file_printf(ms, ", ") == -1)
240						goto done;
241					break;
242				case -1:
243					goto done;
244				default:
245					executable = 1;
246					break;
247				}
248				break;
249			case -1:
250				goto done;
251			default:
252				break;
253			}
254		}
255
256		if (file_printf(ms, "%s", code) == -1)
257			goto done;
258
259		if (subtype) {
260			if (file_printf(ms, " %s", subtype) == -1)
261				goto done;
262		}
263
264		if (file_printf(ms, " %s", type) == -1)
265			goto done;
266
267		if (executable)
268			if (file_printf(ms, " executable") == -1)
269				goto done;
270
271		if (has_long_lines)
272			if (file_printf(ms, ", with very long lines") == -1)
273				goto done;
274
275		/*
276		 * Only report line terminators if we find one other than LF,
277		 * or if we find none at all.
278		 */
279		if ((n_crlf == 0 && n_cr == 0 && n_nel == 0 && n_lf == 0) ||
280		    (n_crlf != 0 || n_cr != 0 || n_nel != 0)) {
281			if (file_printf(ms, ", with") == -1)
282				goto done;
283
284			if (n_crlf == 0 && n_cr == 0 && n_nel == 0 && n_lf == 0) {
285				if (file_printf(ms, " no") == -1)
286					goto done;
287			} else {
288				if (n_crlf) {
289					if (file_printf(ms, " CRLF") == -1)
290						goto done;
291					if (n_cr || n_lf || n_nel)
292						if (file_printf(ms, ",") == -1)
293							goto done;
294				}
295				if (n_cr) {
296					if (file_printf(ms, " CR") == -1)
297						goto done;
298					if (n_lf || n_nel)
299						if (file_printf(ms, ",") == -1)
300							goto done;
301				}
302				if (n_lf) {
303					if (file_printf(ms, " LF") == -1)
304						goto done;
305					if (n_nel)
306						if (file_printf(ms, ",") == -1)
307							goto done;
308				}
309				if (n_nel)
310					if (file_printf(ms, " NEL") == -1)
311						goto done;
312			}
313
314			if (file_printf(ms, " line terminators") == -1)
315				goto done;
316		}
317
318		if (has_escapes)
319			if (file_printf(ms, ", with escape sequences") == -1)
320				goto done;
321		if (has_backspace)
322			if (file_printf(ms, ", with overstriking") == -1)
323				goto done;
324	}
325	rv = 1;
326done:
327	free(utf8_buf);
328
329	return rv;
330}
331
332/*
333 * Encode Unicode string as UTF-8, returning pointer to character
334 * after end of string, or NULL if an invalid character is found.
335 */
336private unsigned char *
337encode_utf8(unsigned char *buf, size_t len, unichar *ubuf, size_t ulen)
338{
339	size_t i;
340	unsigned char *end = buf + len;
341
342	for (i = 0; i < ulen; i++) {
343		if (ubuf[i] <= 0x7f) {
344			if (end - buf < 1)
345				return NULL;
346			*buf++ = CAST(unsigned char, ubuf[i]);
347		} else if (ubuf[i] <= 0x7ff) {
348			if (end - buf < 2)
349				return NULL;
350			*buf++ = CAST(unsigned char, (ubuf[i] >> 6) + 0xc0);
351			*buf++ = CAST(unsigned char, (ubuf[i] & 0x3f) + 0x80);
352		} else if (ubuf[i] <= 0xffff) {
353			if (end - buf < 3)
354				return NULL;
355			*buf++ = CAST(unsigned char, (ubuf[i] >> 12) + 0xe0);
356			*buf++ = CAST(unsigned char, ((ubuf[i] >> 6) & 0x3f) + 0x80);
357			*buf++ = CAST(unsigned char, (ubuf[i] & 0x3f) + 0x80);
358		} else if (ubuf[i] <= 0x1fffff) {
359			if (end - buf < 4)
360				return NULL;
361			*buf++ = CAST(unsigned char, (ubuf[i] >> 18) + 0xf0);
362			*buf++ = CAST(unsigned char, ((ubuf[i] >> 12) & 0x3f) + 0x80);
363			*buf++ = CAST(unsigned char, ((ubuf[i] >>  6) & 0x3f) + 0x80);
364			*buf++ = CAST(unsigned char, (ubuf[i] & 0x3f) + 0x80);
365		} else if (ubuf[i] <= 0x3ffffff) {
366			if (end - buf < 5)
367				return NULL;
368			*buf++ = CAST(unsigned char, (ubuf[i] >> 24) + 0xf8);
369			*buf++ = CAST(unsigned char, ((ubuf[i] >> 18) & 0x3f) + 0x80);
370			*buf++ = CAST(unsigned char, ((ubuf[i] >> 12) & 0x3f) + 0x80);
371			*buf++ = CAST(unsigned char, ((ubuf[i] >>  6) & 0x3f) + 0x80);
372			*buf++ = CAST(unsigned char, (ubuf[i] & 0x3f) + 0x80);
373		} else if (ubuf[i] <= 0x7fffffff) {
374			if (end - buf < 6)
375				return NULL;
376			*buf++ = CAST(unsigned char, (ubuf[i] >> 30) + 0xfc);
377			*buf++ = CAST(unsigned char, ((ubuf[i] >> 24) & 0x3f) + 0x80);
378			*buf++ = CAST(unsigned char, ((ubuf[i] >> 18) & 0x3f) + 0x80);
379			*buf++ = CAST(unsigned char, ((ubuf[i] >> 12) & 0x3f) + 0x80);
380			*buf++ = CAST(unsigned char, ((ubuf[i] >>  6) & 0x3f) + 0x80);
381			*buf++ = CAST(unsigned char, (ubuf[i] & 0x3f) + 0x80);
382		} else /* Invalid character */
383			return NULL;
384	}
385
386	return buf;
387}
388