1 /*        $NetBSD: uxtext.c,v 1.3 2020/03/18 19:05:16 christos Exp $  */
2 
3 /*++
4 /* NAME
5 /*        uxtext 3
6 /* SUMMARY
7 /*        quote/unquote text, xtext style.
8 /* SYNOPSIS
9 /*        #include <uxtext.h>
10 /*
11 /*        VSTRING   *uxtext_quote(quoted, unquoted, special)
12 /*        VSTRING   *quoted;
13 /*        const char *unquoted;
14 /*        const char *special;
15 /*
16 /*        VSTRING   *uxtext_quote_append(unquoted, quoted, special)
17 /*        VSTRING   *unquoted;
18 /*        const char *quoted;
19 /*        const char *special;
20 /*
21 /*        VSTRING   *uxtext_unquote(unquoted, quoted)
22 /*        VSTRING   *unquoted;
23 /*        const char *quoted;
24 /*
25 /*        VSTRING   *uxtext_unquote_append(unquoted, quoted)
26 /*        VSTRING   *unquoted;
27 /*        const char *quoted;
28 /* DESCRIPTION
29 /*        uxtext_quote() takes a null-terminated UTF8 string and
30 /*        replaces characters \, <33(10) and >126(10), as well as
31 /*        characters specified with "special" with \x{XX}, XX being
32 /*        a 2-6-digit uppercase hexadecimal equivalent.
33 /*
34 /*        uxtext_quote_append() is like uxtext_quote(), but appends
35 /*        the conversion result to the result buffer.
36 /*
37 /*        uxtext_unquote() performs the opposite transformation. This
38 /*        function understands lowercase, uppercase, and mixed case
39 /*        \x{XX...} sequences.  The result value is the unquoted
40 /*        argument in case of success, a null pointer otherwise.
41 /*
42 /*        uxtext_unquote_append() is like uxtext_unquote(), but appends
43 /*        the conversion result to the result buffer.
44 /* BUGS
45 /*        This module cannot process null characters in data.
46 /* LICENSE
47 /* .ad
48 /* .fi
49 /*        The Secure Mailer license must be distributed with this software.
50 /* AUTHOR(S)
51 /*        Arnt Gulbrandsen
52 /*
53 /*        Wietse Venema
54 /*        IBM T.J. Watson Research
55 /*        P.O. Box 704
56 /*        Yorktown Heights, NY 10598, USA
57 /*
58 /*        Wietse Venema
59 /*        Google, Inc.
60 /*        111 8th Avenue
61 /*        New York, NY 10011, USA
62 /*--*/
63 
64 /* System library. */
65 
66 #include <sys_defs.h>
67 #include <string.h>
68 #include <ctype.h>
69 
70 /* Utility library. */
71 
72 #include "msg.h"
73 #include "vstring.h"
74 #include "uxtext.h"
75 
76 /* Application-specific. */
77 
78 #define STR(x)      vstring_str(x)
79 #define LEN(x)      VSTRING_LEN(x)
80 
81 /* uxtext_quote_append - append unquoted data to quoted data */
82 
uxtext_quote_append(VSTRING * quoted,const char * unquoted,const char * special)83 VSTRING *uxtext_quote_append(VSTRING *quoted, const char *unquoted,
84                                            const char *special)
85 {
86     unsigned const char *cp;
87     int     ch;
88 
89     for (cp = (unsigned const char *) unquoted; (ch = *cp) != 0; cp++) {
90           /* Fix 20140709: the '\' character must always be quoted. */
91           if (ch != '\\' && ch > 32 && ch < 127
92               && (*special == 0 || strchr(special, ch) == 0)) {
93               VSTRING_ADDCH(quoted, ch);
94           } else {
95 
96               /*
97                * had RFC6533 been written like 6531 and 6532, this else clause
98                * would be one line long.
99                */
100               int     unicode = 0;
101               int     pick = 0;
102 
103               if (ch < 0x80) {
104                     //0000 0000 - 0000 007 F 0x xxxxxx
105                         unicode = ch;
106               } else if ((ch & 0xe0) == 0xc0) {
107                     //0000 0080 - 0000 07 FF 110 xxxxx 10 xxxxxx
108                         unicode = (ch & 0x1f);
109                     pick = 1;
110               } else if ((ch & 0xf0) == 0xe0) {
111                     //0000 0800 - 0000 FFFF 1110 xxxx 10 xxxxxx 10 xxxxxx
112                         unicode = (ch & 0x0f);
113                     pick = 2;
114               } else if ((ch & 0xf8) == 0xf0) {
115                     //0001 0000 - 001 F FFFF 11110 xxx 10 xxxxxx 10 xxxxxx 10 xxxxxx
116                         unicode = (ch & 0x07);
117                     pick = 3;
118               } else if ((ch & 0xfc) == 0xf8) {
119                     //0020 0000 - 03 FF FFFF 111110 xx 10 xxxxxx 10 xxxxxx...10 xxxxxx
120                         unicode = (ch & 0x03);
121                     pick = 4;
122               } else if ((ch & 0xfe) == 0xfc) {
123                     //0400 0000 - 7 FFF FFFF 1111110 x 10 xxxxxx...10 xxxxxx
124                         unicode = (ch & 0x01);
125                     pick = 5;
126               } else {
127                     return (0);
128               }
129               while (pick > 0) {
130                     ch = *++cp;
131                     if ((ch & 0xc0) != 0x80)
132                         return (0);
133                     unicode = unicode << 6 | (ch & 0x3f);
134                     pick--;
135               }
136               vstring_sprintf_append(quoted, "\\x{%02X}", unicode);
137           }
138     }
139     VSTRING_TERMINATE(quoted);
140     return (quoted);
141 }
142 
143 /* uxtext_quote - unquoted data to quoted */
144 
uxtext_quote(VSTRING * quoted,const char * unquoted,const char * special)145 VSTRING *uxtext_quote(VSTRING *quoted, const char *unquoted, const char *special)
146 {
147     VSTRING_RESET(quoted);
148     uxtext_quote_append(quoted, unquoted, special);
149     return (quoted);
150 }
151 
152 /* uxtext_unquote_append - quoted data to unquoted */
153 
uxtext_unquote_append(VSTRING * unquoted,const char * quoted)154 VSTRING *uxtext_unquote_append(VSTRING *unquoted, const char *quoted)
155 {
156     const unsigned char *cp;
157     int     ch;
158 
159     for (cp = (const unsigned char *) quoted; (ch = *cp) != 0; cp++) {
160           if (ch == '\\' && cp[1] == 'x' && cp[2] == '{') {
161               int     unicode = 0;
162 
163               cp += 2;
164               while ((ch = *++cp) != '}') {
165                     if (ISDIGIT(ch))
166                         unicode = (unicode << 4) + (ch - '0');
167                     else if (ch >= 'a' && ch <= 'f')
168                         unicode = (unicode << 4) + (ch - 'a' + 10);
169                     else if (ch >= 'A' && ch <= 'F')
170                         unicode = (unicode << 4) + (ch - 'A' + 10);
171                     else
172                         return (0);                         /* also covers the null
173                                                              * terminator */
174                     if (unicode > 0x10ffff)
175                         return (0);
176               }
177 
178               /*
179                * the following block is from
180                * https://github.com/aox/aox/blob/master/encodings/utf.cpp, with
181                * permission by the authors.
182                */
183               if (unicode < 0x80) {
184                     VSTRING_ADDCH(unquoted, (char) unicode);
185               } else if (unicode < 0x800) {
186                     VSTRING_ADDCH(unquoted, 0xc0 | ((char) (unicode >> 6)));
187                     VSTRING_ADDCH(unquoted, 0x80 | ((char) (unicode & 0x3f)));
188               } else if (unicode < 0x10000) {
189                     VSTRING_ADDCH(unquoted, 0xe0 | ((char) (unicode >> 12)));
190                     VSTRING_ADDCH(unquoted, 0x80 | ((char) (unicode >> 6) & 0x3f));
191                     VSTRING_ADDCH(unquoted, 0x80 | ((char) (unicode & 0x3f)));
192               } else if (unicode < 0x200000) {
193                     VSTRING_ADDCH(unquoted, 0xf0 | ((char) (unicode >> 18)));
194                     VSTRING_ADDCH(unquoted, 0x80 | ((char) (unicode >> 12) & 0x3f));
195                     VSTRING_ADDCH(unquoted, 0x80 | ((char) (unicode >> 6) & 0x3f));
196                     VSTRING_ADDCH(unquoted, 0x80 | ((char) (unicode & 0x3f)));
197               } else if (unicode < 0x4000000) {
198                     VSTRING_ADDCH(unquoted, 0xf8 | ((char) (unicode >> 24)));
199                     VSTRING_ADDCH(unquoted, 0x80 | ((char) (unicode >> 18) & 0x3f));
200                     VSTRING_ADDCH(unquoted, 0x80 | ((char) (unicode >> 12) & 0x3f));
201                     VSTRING_ADDCH(unquoted, 0x80 | ((char) (unicode >> 6) & 0x3f));
202                     VSTRING_ADDCH(unquoted, 0x80 | ((char) (unicode & 0x3f)));
203               } else {
204                     VSTRING_ADDCH(unquoted, 0xfc | ((char) (unicode >> 30)));
205                     VSTRING_ADDCH(unquoted, 0x80 | ((char) (unicode >> 24) & 0x3f));
206                     VSTRING_ADDCH(unquoted, 0x80 | ((char) (unicode >> 18) & 0x3f));
207                     VSTRING_ADDCH(unquoted, 0x80 | ((char) (unicode >> 12) & 0x3f));
208                     VSTRING_ADDCH(unquoted, 0x80 | ((char) (unicode >> 6) & 0x3f));
209                     VSTRING_ADDCH(unquoted, 0x80 | ((char) (unicode & 0x3f)));
210               }
211           } else {
212               VSTRING_ADDCH(unquoted, ch);
213           }
214     }
215     VSTRING_TERMINATE(unquoted);
216     return (unquoted);
217 }
218 
219 /* uxtext_unquote - quoted data to unquoted */
220 
uxtext_unquote(VSTRING * unquoted,const char * quoted)221 VSTRING *uxtext_unquote(VSTRING *unquoted, const char *quoted)
222 {
223     VSTRING_RESET(unquoted);
224     return (uxtext_unquote_append(unquoted, quoted) ? unquoted : 0);
225 }
226 
227 #ifdef TEST
228 
229  /*
230   * Proof-of-concept test program: convert to quoted and back.
231   */
232 #include <vstream.h>
233 
234 #define BUFLEN 1024
235 
read_buf(VSTREAM * fp,VSTRING * buf)236 static ssize_t read_buf(VSTREAM *fp, VSTRING *buf)
237 {
238     ssize_t len;
239 
240     len = vstream_fread_buf(fp, buf, BUFLEN);
241     VSTRING_TERMINATE(buf);
242     return (len);
243 }
244 
main(int unused_argc,char ** unused_argv)245 int     main(int unused_argc, char **unused_argv)
246 {
247     VSTRING *unquoted = vstring_alloc(BUFLEN);
248     VSTRING *quoted = vstring_alloc(100);
249     ssize_t len;
250 
251     /*
252      * Negative tests.
253      */
254     if (uxtext_unquote(unquoted, "\\x{x1}") != 0)
255           msg_warn("undetected error pattern 1");
256     if (uxtext_unquote(unquoted, "\\x{2x}") != 0)
257           msg_warn("undetected error pattern 2");
258     if (uxtext_unquote(unquoted, "\\x{33") != 0)
259           msg_warn("undetected error pattern 3");
260 
261     /*
262      * Positive tests.
263      */
264     while ((len = read_buf(VSTREAM_IN, unquoted)) > 0) {
265           uxtext_quote(quoted, STR(unquoted), "+=");
266           if (uxtext_unquote(unquoted, STR(quoted)) == 0)
267               msg_fatal("bad input: %.100s", STR(quoted));
268           if (LEN(unquoted) != len)
269               msg_fatal("len %ld != unquoted len %ld",
270                           (long) len, (long) LEN(unquoted));
271           if (vstream_fwrite(VSTREAM_OUT, STR(unquoted), LEN(unquoted)) != LEN(unquoted))
272               msg_fatal("write error: %m");
273     }
274     vstream_fflush(VSTREAM_OUT);
275     vstring_free(unquoted);
276     vstring_free(quoted);
277     return (0);
278 }
279 
280 #endif
281