/[pcre]/code/trunk/pcre_valid_utf8.c
ViewVC logotype

Contents of /code/trunk/pcre_valid_utf8.c

Parent Directory Parent Directory | Revision Log Revision Log


Revision 598 - (hide annotations) (download)
Sat May 7 15:37:31 2011 UTC (2 years, 11 months ago) by ph10
File MIME type: text/plain
File size: 10056 byte(s)
Pass back detailed info when UTF-8 check fails at runtime.

1 nigel 77 /*************************************************
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 ph10 409 Copyright (c) 1997-2009 University of Cambridge
10 nigel 77
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 ph10 200 #ifdef HAVE_CONFIG_H
46 ph10 236 #include "config.h"
47 ph10 200 #endif
48 ph10 199
49 nigel 77 #include "pcre_internal.h"
50    
51    
52     /*************************************************
53     * Validate a UTF-8 string *
54     *************************************************/
55    
56     /* This function is called (optionally) at the start of compile or match, to
57 ph10 598 check that a supposed UTF-8 string is actually valid. The early check means
58 nigel 77 that subsequent code can assume it is dealing with a valid string. The check
59 ph10 598 can be turned off for maximum performance, but the consequences of supplying an
60     invalid string are then undefined.
61 nigel 77
62 ph10 211 Originally, this function checked according to RFC 2279, allowing for values in
63     the range 0 to 0x7fffffff, up to 6 bytes long, but ensuring that they were in
64     the canonical format. Once somebody had pointed out RFC 3629 to me (it
65 ph10 293 obsoletes 2279), additional restrictions were applied. The values are now
66 ph10 211 limited to be between 0 and 0x0010ffff, no more than 4 bytes long, and the
67 ph10 598 subrange 0xd000 to 0xdfff is excluded. However, the format of 5-byte and 6-byte
68     characters is still checked.
69 ph10 211
70 ph10 598 From release 8.13 more information about the details of the error are passed
71     back in the error code:
72    
73     PCRE_UTF8_ERR0 No error
74     PCRE_UTF8_ERR1 Missing 1 byte at the end of the string
75     PCRE_UTF8_ERR2 Missing 2 bytes at the end of the string
76     PCRE_UTF8_ERR3 Missing 3 bytes at the end of the string
77     PCRE_UTF8_ERR4 Missing 4 bytes at the end of the string
78     PCRE_UTF8_ERR5 Missing 5 bytes at the end of the string
79     PCRE_UTF8_ERR6 2nd-byte's two top bits are not 0x80
80     PCRE_UTF8_ERR7 3rd-byte's two top bits are not 0x80
81     PCRE_UTF8_ERR8 4th-byte's two top bits are not 0x80
82     PCRE_UTF8_ERR9 5th-byte's two top bits are not 0x80
83     PCRE_UTF8_ERR10 6th-byte's two top bits are not 0x80
84     PCRE_UTF8_ERR11 5-byte character is not permitted by RFC 3629
85     PCRE_UTF8_ERR12 6-byte character is not permitted by RFC 3629
86     PCRE_UTF8_ERR13 4-byte character with value > 0x10ffff is not permitted
87     PCRE_UTF8_ERR14 3-byte character with value 0xd000-0xdfff is not permitted
88     PCRE_UTF8_ERR15 Overlong 2-byte sequence
89     PCRE_UTF8_ERR16 Overlong 3-byte sequence
90     PCRE_UTF8_ERR17 Overlong 4-byte sequence
91     PCRE_UTF8_ERR18 Overlong 5-byte sequence (won't ever occur)
92     PCRE_UTF8_ERR19 Overlong 6-byte sequence (won't ever occur)
93     PCRE_UTF8_ERR20 Isolated 0x80 byte (not within UTF-8 character)
94     PCRE_UTF8_ERR21 Byte with the illegal value 0xfe or 0xff
95    
96 nigel 77 Arguments:
97     string points to the string
98     length length of string, or -1 if the string is zero-terminated
99 ph10 598 errp pointer to an error code variable
100 nigel 77
101     Returns: < 0 if the string is a valid UTF-8 string
102 ph10 598 >= 0 otherwise; the value is the offset of the bad character
103 nigel 77 */
104    
105 nigel 87 int
106 ph10 598 _pcre_valid_utf8(USPTR string, int length, int *errorcode)
107 nigel 77 {
108 ph10 110 #ifdef SUPPORT_UTF8
109 ph10 409 register USPTR p;
110 nigel 77
111     if (length < 0)
112     {
113     for (p = string; *p != 0; p++);
114     length = p - string;
115     }
116    
117 ph10 598 *errorcode = PCRE_UTF8_ERR0;
118    
119 nigel 77 for (p = string; length-- > 0; p++)
120     {
121 ph10 598 register int ab, c, d;
122    
123     c = *p;
124     if (c < 128) continue; /* ASCII character */
125    
126     if (c < 0xc0) /* Isolated 10xx xxxx byte */
127     {
128     *errorcode = PCRE_UTF8_ERR20;
129     return p - string;
130     }
131    
132     if (c >= 0xfe) /* Invalid 0xfe or 0xff bytes */
133     {
134     *errorcode = PCRE_UTF8_ERR21;
135     return p - string;
136     }
137    
138 ph10 211 ab = _pcre_utf8_table4[c & 0x3f]; /* Number of additional bytes */
139 ph10 598 if (length < ab)
140     {
141     *errorcode = ab - length; /* Codes ERR1 to ERR5 */
142     return p - string; /* Missing bytes */
143     }
144     length -= ab; /* Length remaining */
145 nigel 77
146     /* Check top bits in the second byte */
147 ph10 598
148     if (((d = *(++p)) & 0xc0) != 0x80)
149     {
150     *errorcode = PCRE_UTF8_ERR6;
151     return p - string - 1;
152     }
153 nigel 77
154 ph10 598 /* For each length, check that the remaining bytes start with the 0x80 bit
155     set and not the 0x40 bit. Then check for an overlong sequence, and for the
156     excluded range 0xd800 to 0xdfff. */
157 ph10 211
158 nigel 77 switch (ab)
159     {
160 ph10 598 /* 2-byte character. No further bytes to check for 0x80. Check first byte
161     for for xx00 000x (overlong sequence). */
162    
163     case 1: if ((c & 0x3e) == 0)
164     {
165     *errorcode = PCRE_UTF8_ERR15;
166     return p - string - 1;
167     }
168     break;
169 ph10 211
170 ph10 598 /* 3-byte character. Check third byte for 0x80. Then check first 2 bytes
171     for 1110 0000, xx0x xxxx (overlong sequence) or
172     1110 1101, 1010 xxxx (0xd800 - 0xdfff) */
173    
174 nigel 77 case 2:
175 ph10 598 if ((*(++p) & 0xc0) != 0x80) /* Third byte */
176     {
177     *errorcode = PCRE_UTF8_ERR7;
178     return p - string - 2;
179     }
180     if (c == 0xe0 && (d & 0x20) == 0)
181     {
182     *errorcode = PCRE_UTF8_ERR16;
183     return p - string - 2;
184     }
185     if (c == 0xed && d >= 0xa0)
186     {
187     *errorcode = PCRE_UTF8_ERR14;
188     return p - string - 2;
189     }
190 nigel 77 break;
191    
192 ph10 598 /* 4-byte character. Check 3rd and 4th bytes for 0x80. Then check first 2
193     bytes for for 1111 0000, xx00 xxxx (overlong sequence), then check for a
194     character greater than 0x0010ffff (f4 8f bf bf) */
195    
196 nigel 77 case 3:
197 ph10 598 if ((*(++p) & 0xc0) != 0x80) /* Third byte */
198     {
199     *errorcode = PCRE_UTF8_ERR7;
200     return p - string - 2;
201     }
202     if ((*(++p) & 0xc0) != 0x80) /* Fourth byte */
203     {
204     *errorcode = PCRE_UTF8_ERR8;
205     return p - string - 3;
206     }
207     if (c == 0xf0 && (d & 0x30) == 0)
208     {
209     *errorcode = PCRE_UTF8_ERR17;
210     return p - string - 3;
211     }
212     if (c > 0xf4 || (c == 0xf4 && d > 0x8f))
213     {
214     *errorcode = PCRE_UTF8_ERR13;
215     return p - string - 3;
216     }
217 nigel 77 break;
218    
219 ph10 598 /* 5-byte and 6-byte characters are not allowed by RFC 3629, and will be
220     rejected by the length test below. However, we do the appropriate tests
221     here so that overlong sequences get diagnosed, and also in case there is
222     ever an option for handling these larger code points. */
223 ph10 211
224 ph10 598 /* 5-byte character. Check 3rd, 4th, and 5th bytes for 0x80. Then check for
225     1111 1000, xx00 0xxx */
226    
227     case 4:
228     if ((*(++p) & 0xc0) != 0x80) /* Third byte */
229     {
230     *errorcode = PCRE_UTF8_ERR7;
231     return p - string - 2;
232     }
233     if ((*(++p) & 0xc0) != 0x80) /* Fourth byte */
234     {
235     *errorcode = PCRE_UTF8_ERR8;
236     return p - string - 3;
237     }
238     if ((*(++p) & 0xc0) != 0x80) /* Fifth byte */
239     {
240     *errorcode = PCRE_UTF8_ERR9;
241     return p - string - 4;
242     }
243     if (c == 0xf8 && (d & 0x38) == 0)
244     {
245     *errorcode = PCRE_UTF8_ERR18;
246     return p - string - 4;
247     }
248 nigel 77 break;
249    
250 ph10 598 /* 6-byte character. Check 3rd-6th bytes for 0x80. Then check for
251     1111 1100, xx00 00xx. */
252    
253 nigel 77 case 5:
254 ph10 598 if ((*(++p) & 0xc0) != 0x80) /* Third byte */
255     {
256     *errorcode = PCRE_UTF8_ERR7;
257     return p - string - 2;
258     }
259     if ((*(++p) & 0xc0) != 0x80) /* Fourth byte */
260     {
261     *errorcode = PCRE_UTF8_ERR8;
262     return p - string - 3;
263     }
264     if ((*(++p) & 0xc0) != 0x80) /* Fifth byte */
265     {
266     *errorcode = PCRE_UTF8_ERR9;
267     return p - string - 4;
268     }
269     if ((*(++p) & 0xc0) != 0x80) /* Sixth byte */
270     {
271     *errorcode = PCRE_UTF8_ERR10;
272     return p - string - 5;
273     }
274     if (c == 0xfc && (d & 0x3c) == 0)
275     {
276     *errorcode = PCRE_UTF8_ERR19;
277     return p - string - 5;
278     }
279 nigel 77 break;
280     }
281 ph10 598
282     /* Character is valid under RFC 2279, but 4-byte and 5-byte characters are
283     excluded by RFC 3629. The pointer p is currently at the last byte of the
284     character. */
285 nigel 77
286 ph10 598 if (ab > 3)
287 nigel 77 {
288 ph10 598 *errorcode = (ab == 4)? PCRE_UTF8_ERR11 : PCRE_UTF8_ERR12;
289     return p - string - ab;
290     }
291 nigel 77 }
292 ph10 598
293     #else /* SUPPORT_UTF8 */
294 ph10 369 (void)(string); /* Keep picky compilers happy */
295 ph10 371 (void)(length);
296 ph10 110 #endif
297 nigel 77
298 ph10 598 return -1; /* This indicates success */
299 nigel 77 }
300    
301     /* End of pcre_valid_utf8.c */

Properties

Name Value
svn:eol-style native
svn:keywords "Author Date Id Revision Url"

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12