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 1.105 2019/06/08 20:49:14 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	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 executable = 0;
129
130	size_t last_line_end = CAST(size_t, -1);
131	int 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 *, malloc(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	if ((ms->flags & (MAGIC_APPLE|MAGIC_EXTENSION)))
168		return 0;
169
170	/* Now try to discover other details about the file. */
171	for (i = 0; i < ulen; i++) {
172		if (ubuf[i] == '\n') {
173			if (seen_cr)
174				n_crlf++;
175			else
176				n_lf++;
177			last_line_end = i;
178		} else if (seen_cr)
179			n_cr++;
180
181		seen_cr = (ubuf[i] == '\r');
182		if (seen_cr)
183			last_line_end = i;
184
185		if (ubuf[i] == 0x85) { /* X3.64/ECMA-43 "next line" character */
186			n_nel++;
187			last_line_end = i;
188		}
189
190		/* If this line is _longer_ than MAXLINELEN, remember it. */
191		if (i > last_line_end + MAXLINELEN)
192			has_long_lines = 1;
193
194		if (ubuf[i] == '\033')
195			has_escapes = 1;
196		if (ubuf[i] == '\b')
197			has_backspace = 1;
198	}
199
200	/* Beware, if the data has been truncated, the final CR could have
201	   been followed by a LF.  If we have ms->bytes_max bytes, it indicates
202	   that the data might have been truncated, probably even before
203	   this function was called. */
204	if (seen_cr && nbytes < ms->bytes_max)
205		n_cr++;
206
207	if (strcmp(type, "binary") == 0) {
208		rv = 0;
209		goto done;
210	}
211	len = file_printedlen(ms);
212	if (mime) {
213		if ((mime & MAGIC_MIME_TYPE) != 0) {
214			if (len) {
215				/*
216				 * Softmagic printed something, we
217				 * are either done, or we need a separator
218				 */
219				if ((ms->flags & MAGIC_CONTINUE) == 0) {
220					rv = 1;
221					goto done;
222				}
223				if (need_separator && file_separator(ms) == -1)
224					goto done;
225			}
226			if (subtype_mime) {
227				if (file_printf(ms, "%s", subtype_mime) == -1)
228					goto done;
229			} else {
230				if (file_printf(ms, "text/plain") == -1)
231					goto done;
232			}
233		}
234	} else {
235		if (len) {
236			switch (file_replace(ms, " text$", ", ")) {
237			case 0:
238				switch (file_replace(ms, " text executable$",
239				    ", ")) {
240				case 0:
241					if (file_printf(ms, ", ") == -1)
242						goto done;
243					break;
244				case -1:
245					goto done;
246				default:
247					executable = 1;
248					break;
249				}
250				break;
251			case -1:
252				goto done;
253			default:
254				break;
255			}
256		}
257
258		if (file_printf(ms, "%s", code) == -1)
259			goto done;
260
261		if (subtype) {
262			if (file_printf(ms, " %s", subtype) == -1)
263				goto done;
264		}
265
266		if (file_printf(ms, " %s", type) == -1)
267			goto done;
268
269		if (executable)
270			if (file_printf(ms, " executable") == -1)
271				goto done;
272
273		if (has_long_lines)
274			if (file_printf(ms, ", with very long lines") == -1)
275				goto done;
276
277		/*
278		 * Only report line terminators if we find one other than LF,
279		 * or if we find none at all.
280		 */
281		if ((n_crlf == 0 && n_cr == 0 && n_nel == 0 && n_lf == 0) ||
282		    (n_crlf != 0 || n_cr != 0 || n_nel != 0)) {
283			if (file_printf(ms, ", with") == -1)
284				goto done;
285
286			if (n_crlf == 0 && n_cr == 0 && n_nel == 0 && n_lf == 0) {
287				if (file_printf(ms, " no") == -1)
288					goto done;
289			} else {
290				if (n_crlf) {
291					if (file_printf(ms, " CRLF") == -1)
292						goto done;
293					if (n_cr || n_lf || n_nel)
294						if (file_printf(ms, ",") == -1)
295							goto done;
296				}
297				if (n_cr) {
298					if (file_printf(ms, " CR") == -1)
299						goto done;
300					if (n_lf || n_nel)
301						if (file_printf(ms, ",") == -1)
302							goto done;
303				}
304				if (n_lf) {
305					if (file_printf(ms, " LF") == -1)
306						goto done;
307					if (n_nel)
308						if (file_printf(ms, ",") == -1)
309							goto done;
310				}
311				if (n_nel)
312					if (file_printf(ms, " NEL") == -1)
313						goto done;
314			}
315
316			if (file_printf(ms, " line terminators") == -1)
317				goto done;
318		}
319
320		if (has_escapes)
321			if (file_printf(ms, ", with escape sequences") == -1)
322				goto done;
323		if (has_backspace)
324			if (file_printf(ms, ", with overstriking") == -1)
325				goto done;
326	}
327	rv = 1;
328done:
329	free(utf8_buf);
330
331	return rv;
332}
333
334/*
335 * Encode Unicode string as UTF-8, returning pointer to character
336 * after end of string, or NULL if an invalid character is found.
337 */
338private unsigned char *
339encode_utf8(unsigned char *buf, size_t len, unichar *ubuf, size_t ulen)
340{
341	size_t i;
342	unsigned char *end = buf + len;
343
344	for (i = 0; i < ulen; i++) {
345		if (ubuf[i] <= 0x7f) {
346			if (end - buf < 1)
347				return NULL;
348			*buf++ = CAST(unsigned char, ubuf[i]);
349		} else if (ubuf[i] <= 0x7ff) {
350			if (end - buf < 2)
351				return NULL;
352			*buf++ = CAST(unsigned char, (ubuf[i] >> 6) + 0xc0);
353			*buf++ = CAST(unsigned char, (ubuf[i] & 0x3f) + 0x80);
354		} else if (ubuf[i] <= 0xffff) {
355			if (end - buf < 3)
356				return NULL;
357			*buf++ = CAST(unsigned char, (ubuf[i] >> 12) + 0xe0);
358			*buf++ = CAST(unsigned char, ((ubuf[i] >> 6) & 0x3f) + 0x80);
359			*buf++ = CAST(unsigned char, (ubuf[i] & 0x3f) + 0x80);
360		} else if (ubuf[i] <= 0x1fffff) {
361			if (end - buf < 4)
362				return NULL;
363			*buf++ = CAST(unsigned char, (ubuf[i] >> 18) + 0xf0);
364			*buf++ = CAST(unsigned char, ((ubuf[i] >> 12) & 0x3f) + 0x80);
365			*buf++ = CAST(unsigned char, ((ubuf[i] >>  6) & 0x3f) + 0x80);
366			*buf++ = CAST(unsigned char, (ubuf[i] & 0x3f) + 0x80);
367		} else if (ubuf[i] <= 0x3ffffff) {
368			if (end - buf < 5)
369				return NULL;
370			*buf++ = CAST(unsigned char, (ubuf[i] >> 24) + 0xf8);
371			*buf++ = CAST(unsigned char, ((ubuf[i] >> 18) & 0x3f) + 0x80);
372			*buf++ = CAST(unsigned char, ((ubuf[i] >> 12) & 0x3f) + 0x80);
373			*buf++ = CAST(unsigned char, ((ubuf[i] >>  6) & 0x3f) + 0x80);
374			*buf++ = CAST(unsigned char, (ubuf[i] & 0x3f) + 0x80);
375		} else if (ubuf[i] <= 0x7fffffff) {
376			if (end - buf < 6)
377				return NULL;
378			*buf++ = CAST(unsigned char, (ubuf[i] >> 30) + 0xfc);
379			*buf++ = CAST(unsigned char, ((ubuf[i] >> 24) & 0x3f) + 0x80);
380			*buf++ = CAST(unsigned char, ((ubuf[i] >> 18) & 0x3f) + 0x80);
381			*buf++ = CAST(unsigned char, ((ubuf[i] >> 12) & 0x3f) + 0x80);
382			*buf++ = CAST(unsigned char, ((ubuf[i] >>  6) & 0x3f) + 0x80);
383			*buf++ = CAST(unsigned char, (ubuf[i] & 0x3f) + 0x80);
384		} else /* Invalid character */
385			return NULL;
386	}
387
388	return buf;
389}
390