Process Hacker
pcre_valid_utf8.c
Go to the documentation of this file.
1 /*************************************************
2 * Perl-Compatible Regular Expressions *
3 *************************************************/
4 
5 /* PCRE is a library of functions to support regular expressions whose syntax
6 and semantics are as close as possible to those of the Perl 5 language.
7 
8  Written by Philip Hazel
9  Copyright (c) 1997-2009 University of Cambridge
10 
11 -----------------------------------------------------------------------------
12 Redistribution and use in source and binary forms, with or without
13 modification, are permitted provided that the following conditions are met:
14 
15  * Redistributions of source code must retain the above copyright notice,
16  this list of conditions and the following disclaimer.
17 
18  * Redistributions in binary form must reproduce the above copyright
19  notice, this list of conditions and the following disclaimer in the
20  documentation and/or other materials provided with the distribution.
21 
22  * Neither the name of the University of Cambridge nor the names of its
23  contributors may be used to endorse or promote products derived from
24  this software without specific prior written permission.
25 
26 THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
27 AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
28 IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
29 ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
30 LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
31 CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
32 SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
33 INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
34 CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
35 ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
36 POSSIBILITY OF SUCH DAMAGE.
37 -----------------------------------------------------------------------------
38 */
39 
40 
41 /* This module contains an internal function for validating UTF-8 character
42 strings. */
43 
44 
45 #define HAVE_CONFIG_H
46 #ifdef HAVE_CONFIG_H
47 #include "config.h"
48 #endif
49 
50 #include "pcre_internal.h"
51 
52 
53 /*************************************************
54 * Validate a UTF-8 string *
55 *************************************************/
56 
57 /* This function is called (optionally) at the start of compile or match, to
58 validate that a supposed UTF-8 string is actually valid. The early check means
59 that subsequent code can assume it is dealing with a valid string. The check
60 can be turned off for maximum performance, but the consequences of supplying
61 an invalid string are then undefined.
62 
63 Originally, this function checked according to RFC 2279, allowing for values in
64 the range 0 to 0x7fffffff, up to 6 bytes long, but ensuring that they were in
65 the canonical format. Once somebody had pointed out RFC 3629 to me (it
66 obsoletes 2279), additional restrictions were applied. The values are now
67 limited to be between 0 and 0x0010ffff, no more than 4 bytes long, and the
68 subrange 0xd000 to 0xdfff is excluded.
69 
70 Arguments:
71  string points to the string
72  length length of string, or -1 if the string is zero-terminated
73 
74 Returns: < 0 if the string is a valid UTF-8 string
75  >= 0 otherwise; the value is the offset of the bad byte
76 
77 Bad bytes can be:
78 
79  . An isolated byte whose most significant bits are 0x80, because this
80  can only correctly appear within a UTF-8 character;
81 
82  . A byte whose most significant bits are 0xc0, but whose other bits indicate
83  that there are more than 3 additional bytes (i.e. an RFC 2279 starting
84  byte, which is no longer valid under RFC 3629);
85 
86  .
87 
88 The returned offset may also be equal to the length of the string; this means
89 that one or more bytes is missing from the final UTF-8 character.
90 */
91 
92 int
93 _pcre_valid_utf8(USPTR string, int length)
94 {
95 #ifdef SUPPORT_UTF8
96 register USPTR p;
97 
98 if (length < 0)
99  {
100  for (p = string; *p != 0; p++);
101  length = p - string;
102  }
103 
104 for (p = string; length-- > 0; p++)
105  {
106  register int ab;
107  register int c = *p;
108  if (c < 128) continue;
109  if (c < 0xc0) return p - string;
110  ab = _pcre_utf8_table4[c & 0x3f]; /* Number of additional bytes */
111  if (ab > 3) return p - string; /* Too many for RFC 3629 */
112  if (length < ab) return p + 1 + length - string; /* Missing bytes */
113  length -= ab;
114 
115  /* Check top bits in the second byte */
116  if ((*(++p) & 0xc0) != 0x80) return p - string;
117 
118  /* Check for overlong sequences for each different length, and for the
119  excluded range 0xd000 to 0xdfff. */
120 
121  switch (ab)
122  {
123  /* Check for xx00 000x (overlong sequence) */
124 
125  case 1:
126  if ((c & 0x3e) == 0) return p - string;
127  continue; /* We know there aren't any more bytes to check */
128 
129  /* Check for 1110 0000, xx0x xxxx (overlong sequence) or
130  1110 1101, 1010 xxxx (0xd000 - 0xdfff) */
131 
132  case 2:
133  if ((c == 0xe0 && (*p & 0x20) == 0) ||
134  (c == 0xed && *p >= 0xa0))
135  return p - string;
136  break;
137 
138  /* Check for 1111 0000, xx00 xxxx (overlong sequence) or
139  greater than 0x0010ffff (f4 8f bf bf) */
140 
141  case 3:
142  if ((c == 0xf0 && (*p & 0x30) == 0) ||
143  (c > 0xf4 ) ||
144  (c == 0xf4 && *p > 0x8f))
145  return p - string;
146  break;
147 
148 #if 0
149  /* These cases can no longer occur, as we restrict to a maximum of four
150  bytes nowadays. Leave the code here in case we ever want to add an option
151  for longer sequences. */
152 
153  /* Check for 1111 1000, xx00 0xxx */
154  case 4:
155  if (c == 0xf8 && (*p & 0x38) == 0) return p - string;
156  break;
157 
158  /* Check for leading 0xfe or 0xff, and then for 1111 1100, xx00 00xx */
159  case 5:
160  if (c == 0xfe || c == 0xff ||
161  (c == 0xfc && (*p & 0x3c) == 0)) return p - string;
162  break;
163 #endif
164 
165  }
166 
167  /* Check for valid bytes after the 2nd, if any; all must start 10 */
168  while (--ab > 0)
169  {
170  if ((*(++p) & 0xc0) != 0x80) return p - string;
171  }
172  }
173 #else
174 (void)(string); /* Keep picky compilers happy */
175 (void)(length);
176 #endif
177 
178 return -1;
179 }
180 
181 /* End of pcre_valid_utf8.c */