ascmagic.c revision e46ca9aa7ac7fb0d486b2bdca760e288f76b160e
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 1.95 2016/05/03 16:10:37 christos Exp $")
39#endif	/* lint */
40
41#include "magic.h"
42#include <string.h>
43#include <memory.h>
44#include <ctype.h>
45#include <stdlib.h>
46#ifdef HAVE_UNISTD_H
47#include <unistd.h>
48#endif
49
50#define MAXLINELEN 300	/* longest sane line length */
51#define ISSPC(x) ((x) == ' ' || (x) == '\t' || (x) == '\r' || (x) == '\n' \
52		  || (x) == 0x85 || (x) == '\f')
53
54private unsigned char *encode_utf8(unsigned char *, size_t, unichar *, size_t);
55private 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 */
61private size_t
62trim_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
70protected int
71file_ascmagic(struct magic_set *ms, const unsigned char *buf, size_t nbytes,
72	int text)
73{
74	unichar *ubuf = NULL;
75	size_t ulen = 0;
76	int rv = 1;
77
78	const char *code = NULL;
79	const char *code_mime = NULL;
80	const char *type = NULL;
81
82	nbytes = trim_nuls(buf, nbytes);
83
84	/* If file doesn't look like any sort of text, give up. */
85	if (file_encoding(ms, buf, nbytes, &ubuf, &ulen, &code, &code_mime,
86	    &type) == 0)
87		rv = 0;
88        else
89		rv = file_ascmagic_with_encoding(ms, buf, nbytes, ubuf, ulen, code,
90						 type, text);
91
92	free(ubuf);
93
94	return rv;
95}
96
97protected int
98file_ascmagic_with_encoding(struct magic_set *ms, const unsigned char *buf,
99    size_t nbytes, unichar *ubuf, size_t ulen, const char *code,
100    const char *type, int text)
101{
102	unsigned char *utf8_buf = NULL, *utf8_end;
103	size_t mlen, i;
104	int rv = -1;
105	int mime = ms->flags & MAGIC_MIME;
106
107	const char *subtype = NULL;
108	const char *subtype_mime = NULL;
109
110	int has_escapes = 0;
111	int has_backspace = 0;
112	int seen_cr = 0;
113
114	int n_crlf = 0;
115	int n_lf = 0;
116	int n_cr = 0;
117	int n_nel = 0;
118	int executable = 0;
119
120	size_t last_line_end = (size_t)-1;
121	int has_long_lines = 0;
122
123	nbytes = trim_nuls(buf, nbytes);
124
125	/* If we have fewer than 2 bytes, give up. */
126	if (nbytes <= 1) {
127		rv = 0;
128		goto done;
129	}
130
131	if (ulen > 0 && (ms->flags & MAGIC_NO_CHECK_SOFT) == 0) {
132		/* Convert ubuf to UTF-8 and try text soft magic */
133		/* malloc size is a conservative overestimate; could be
134		   improved, or at least realloced after conversion. */
135		mlen = ulen * 6;
136		if ((utf8_buf = CAST(unsigned char *, malloc(mlen))) == NULL) {
137			file_oomem(ms, mlen);
138			goto done;
139		}
140		if ((utf8_end = encode_utf8(utf8_buf, mlen, ubuf, ulen))
141		    == NULL)
142			goto done;
143		if ((rv = file_softmagic(ms, utf8_buf,
144		    (size_t)(utf8_end - utf8_buf), NULL, NULL,
145		    TEXTTEST, text)) == 0)
146			rv = -1;
147		if ((ms->flags & (MAGIC_APPLE|MAGIC_EXTENSION)))
148			return rv == -1 ? 0 : 1;
149	}
150	if ((ms->flags & (MAGIC_APPLE|MAGIC_EXTENSION)))
151		return 0;
152
153	/* Now try to discover other details about the file. */
154	for (i = 0; i < ulen; i++) {
155		if (ubuf[i] == '\n') {
156			if (seen_cr)
157				n_crlf++;
158			else
159				n_lf++;
160			last_line_end = i;
161		} else if (seen_cr)
162			n_cr++;
163
164		seen_cr = (ubuf[i] == '\r');
165		if (seen_cr)
166			last_line_end = i;
167
168		if (ubuf[i] == 0x85) { /* X3.64/ECMA-43 "next line" character */
169			n_nel++;
170			last_line_end = i;
171		}
172
173		/* If this line is _longer_ than MAXLINELEN, remember it. */
174		if (i > last_line_end + MAXLINELEN)
175			has_long_lines = 1;
176
177		if (ubuf[i] == '\033')
178			has_escapes = 1;
179		if (ubuf[i] == '\b')
180			has_backspace = 1;
181	}
182
183	/* Beware, if the data has been truncated, the final CR could have
184	   been followed by a LF.  If we have ms->bytes_max bytes, it indicates
185	   that the data might have been truncated, probably even before
186	   this function was called. */
187	if (seen_cr && nbytes < ms->bytes_max)
188		n_cr++;
189
190	if (strcmp(type, "binary") == 0) {
191		rv = 0;
192		goto done;
193	}
194	if (mime) {
195		if (!file_printedlen(ms) && (mime & MAGIC_MIME_TYPE) != 0) {
196			if (subtype_mime) {
197				if (file_printf(ms, "%s", subtype_mime) == -1)
198					goto done;
199			} else {
200				if (file_printf(ms, "text/plain") == -1)
201					goto done;
202			}
203		}
204	} else {
205		if (file_printedlen(ms)) {
206			switch (file_replace(ms, " text$", ", ")) {
207			case 0:
208				switch (file_replace(ms, " text executable$",
209				    ", ")) {
210				case 0:
211					if (file_printf(ms, ", ") == -1)
212						goto done;
213					break;
214				case -1:
215					goto done;
216				default:
217					executable = 1;
218					break;
219				}
220				break;
221			case -1:
222				goto done;
223			default:
224				break;
225			}
226		}
227
228		if (file_printf(ms, "%s", code) == -1)
229			goto done;
230
231		if (subtype) {
232			if (file_printf(ms, " %s", subtype) == -1)
233				goto done;
234		}
235
236		if (file_printf(ms, " %s", type) == -1)
237			goto done;
238
239		if (executable)
240			if (file_printf(ms, " executable") == -1)
241				goto done;
242
243		if (has_long_lines)
244			if (file_printf(ms, ", with very long lines") == -1)
245				goto done;
246
247		/*
248		 * Only report line terminators if we find one other than LF,
249		 * or if we find none at all.
250		 */
251		if ((n_crlf == 0 && n_cr == 0 && n_nel == 0 && n_lf == 0) ||
252		    (n_crlf != 0 || n_cr != 0 || n_nel != 0)) {
253			if (file_printf(ms, ", with") == -1)
254				goto done;
255
256			if (n_crlf == 0 && n_cr == 0 && n_nel == 0 && n_lf == 0) {
257				if (file_printf(ms, " no") == -1)
258					goto done;
259			} else {
260				if (n_crlf) {
261					if (file_printf(ms, " CRLF") == -1)
262						goto done;
263					if (n_cr || n_lf || n_nel)
264						if (file_printf(ms, ",") == -1)
265							goto done;
266				}
267				if (n_cr) {
268					if (file_printf(ms, " CR") == -1)
269						goto done;
270					if (n_lf || n_nel)
271						if (file_printf(ms, ",") == -1)
272							goto done;
273				}
274				if (n_lf) {
275					if (file_printf(ms, " LF") == -1)
276						goto done;
277					if (n_nel)
278						if (file_printf(ms, ",") == -1)
279							goto done;
280				}
281				if (n_nel)
282					if (file_printf(ms, " NEL") == -1)
283						goto done;
284			}
285
286			if (file_printf(ms, " line terminators") == -1)
287				goto done;
288		}
289
290		if (has_escapes)
291			if (file_printf(ms, ", with escape sequences") == -1)
292				goto done;
293		if (has_backspace)
294			if (file_printf(ms, ", with overstriking") == -1)
295				goto done;
296	}
297	rv = 1;
298done:
299	free(utf8_buf);
300
301	return rv;
302}
303
304/*
305 * Encode Unicode string as UTF-8, returning pointer to character
306 * after end of string, or NULL if an invalid character is found.
307 */
308private unsigned char *
309encode_utf8(unsigned char *buf, size_t len, unichar *ubuf, size_t ulen)
310{
311	size_t i;
312	unsigned char *end = buf + len;
313
314	for (i = 0; i < ulen; i++) {
315		if (ubuf[i] <= 0x7f) {
316			if (end - buf < 1)
317				return NULL;
318			*buf++ = (unsigned char)ubuf[i];
319		} else if (ubuf[i] <= 0x7ff) {
320			if (end - buf < 2)
321				return NULL;
322			*buf++ = (unsigned char)((ubuf[i] >> 6) + 0xc0);
323			*buf++ = (unsigned char)((ubuf[i] & 0x3f) + 0x80);
324		} else if (ubuf[i] <= 0xffff) {
325			if (end - buf < 3)
326				return NULL;
327			*buf++ = (unsigned char)((ubuf[i] >> 12) + 0xe0);
328			*buf++ = (unsigned char)(((ubuf[i] >> 6) & 0x3f) + 0x80);
329			*buf++ = (unsigned char)((ubuf[i] & 0x3f) + 0x80);
330		} else if (ubuf[i] <= 0x1fffff) {
331			if (end - buf < 4)
332				return NULL;
333			*buf++ = (unsigned char)((ubuf[i] >> 18) + 0xf0);
334			*buf++ = (unsigned char)(((ubuf[i] >> 12) & 0x3f) + 0x80);
335			*buf++ = (unsigned char)(((ubuf[i] >>  6) & 0x3f) + 0x80);
336			*buf++ = (unsigned char)((ubuf[i] & 0x3f) + 0x80);
337		} else if (ubuf[i] <= 0x3ffffff) {
338			if (end - buf < 5)
339				return NULL;
340			*buf++ = (unsigned char)((ubuf[i] >> 24) + 0xf8);
341			*buf++ = (unsigned char)(((ubuf[i] >> 18) & 0x3f) + 0x80);
342			*buf++ = (unsigned char)(((ubuf[i] >> 12) & 0x3f) + 0x80);
343			*buf++ = (unsigned char)(((ubuf[i] >>  6) & 0x3f) + 0x80);
344			*buf++ = (unsigned char)((ubuf[i] & 0x3f) + 0x80);
345		} else if (ubuf[i] <= 0x7fffffff) {
346			if (end - buf < 6)
347				return NULL;
348			*buf++ = (unsigned char)((ubuf[i] >> 30) + 0xfc);
349			*buf++ = (unsigned char)(((ubuf[i] >> 24) & 0x3f) + 0x80);
350			*buf++ = (unsigned char)(((ubuf[i] >> 18) & 0x3f) + 0x80);
351			*buf++ = (unsigned char)(((ubuf[i] >> 12) & 0x3f) + 0x80);
352			*buf++ = (unsigned char)(((ubuf[i] >>  6) & 0x3f) + 0x80);
353			*buf++ = (unsigned char)((ubuf[i] & 0x3f) + 0x80);
354		} else /* Invalid character */
355			return NULL;
356	}
357
358	return buf;
359}
360