1 /*
2                             __  __            _
3                          ___\ \/ /_ __   __ _| |_
4                         / _ \\  /| '_ \ / _` | __|
5                        |  __//  \| |_) | (_| | |_
6                         \___/_/\_\ .__/ \__,_|\__|
7                                  |_| XML parser
8 
9    Copyright (c) 1997-2000 Thai Open Source Software Center Ltd
10    Copyright (c) 2000-2017 Expat development team
11    Licensed under the MIT license:
12 
13    Permission is  hereby granted,  free of charge,  to any  person obtaining
14    a  copy  of  this  software   and  associated  documentation  files  (the
15    "Software"),  to  deal in  the  Software  without restriction,  including
16    without  limitation the  rights  to use,  copy,  modify, merge,  publish,
17    distribute, sublicense, and/or sell copies of the Software, and to permit
18    persons  to whom  the Software  is  furnished to  do so,  subject to  the
19    following conditions:
20 
21    The above copyright  notice and this permission notice  shall be included
22    in all copies or substantial portions of the Software.
23 
24    THE  SOFTWARE  IS  PROVIDED  "AS  IS",  WITHOUT  WARRANTY  OF  ANY  KIND,
25    EXPRESS  OR IMPLIED,  INCLUDING  BUT  NOT LIMITED  TO  THE WARRANTIES  OF
26    MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN
27    NO EVENT SHALL THE AUTHORS OR  COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM,
28    DAMAGES OR  OTHER LIABILITY, WHETHER  IN AN  ACTION OF CONTRACT,  TORT OR
29    OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
30    USE OR OTHER DEALINGS IN THE SOFTWARE.
31 */
32 
33 #include <string.h>
34 #include "xmlmime.h"
35 
36 static const char *
getTok(const char ** pp)37 getTok(const char **pp) {
38   /* inComment means one level of nesting; inComment+1 means two levels etc */
39   enum { inAtom, inString, init, inComment };
40   int state = init;
41   const char *tokStart = 0;
42   for (;;) {
43     switch (**pp) {
44     case '\0':
45       if (state == inAtom)
46         return tokStart;
47       return 0;
48     case ' ':
49     case '\r':
50     case '\t':
51     case '\n':
52       if (state == inAtom)
53         return tokStart;
54       break;
55     case '(':
56       if (state == inAtom)
57         return tokStart;
58       if (state != inString)
59         state++;
60       break;
61     case ')':
62       if (state > init)
63         --state;
64       else if (state != inString)
65         return 0;
66       break;
67     case ';':
68     case '/':
69     case '=':
70       if (state == inAtom)
71         return tokStart;
72       if (state == init)
73         return (*pp)++;
74       break;
75     case '\\':
76       ++*pp;
77       if (**pp == '\0')
78         return 0;
79       break;
80     case '"':
81       switch (state) {
82       case inString:
83         ++*pp;
84         return tokStart;
85       case inAtom:
86         return tokStart;
87       case init:
88         tokStart = *pp;
89         state = inString;
90         break;
91       }
92       break;
93     default:
94       if (state == init) {
95         tokStart = *pp;
96         state = inAtom;
97       }
98       break;
99     }
100     ++*pp;
101   }
102   /* not reached */
103 }
104 
105 /* key must be lowercase ASCII */
106 
107 static int
matchkey(const char * start,const char * end,const char * key)108 matchkey(const char *start, const char *end, const char *key) {
109   if (! start)
110     return 0;
111   for (; start != end; start++, key++)
112     if (*start != *key && *start != 'A' + (*key - 'a'))
113       return 0;
114   return *key == '\0';
115 }
116 
117 void
getXMLCharset(const char * buf,char * charset)118 getXMLCharset(const char *buf, char *charset) {
119   const char *next, *p;
120 
121   charset[0] = '\0';
122   next = buf;
123   p = getTok(&next);
124   if (matchkey(p, next, "text"))
125     strcpy(charset, "us-ascii");
126   else if (! matchkey(p, next, "application"))
127     return;
128   p = getTok(&next);
129   if (! p || *p != '/')
130     return;
131   p = getTok(&next);
132   /* BEGIN disabled code */
133   if (0) {
134     if (! matchkey(p, next, "xml") && charset[0] == '\0')
135       return;
136   }
137   /* END disabled code */
138   p = getTok(&next);
139   while (p) {
140     if (*p == ';') {
141       p = getTok(&next);
142       if (matchkey(p, next, "charset")) {
143         p = getTok(&next);
144         if (p && *p == '=') {
145           p = getTok(&next);
146           if (p) {
147             char *s = charset;
148             if (*p == '"') {
149               while (++p != next - 1) {
150                 if (*p == '\\')
151                   ++p;
152                 if (s == charset + CHARSET_MAX - 1) {
153                   charset[0] = '\0';
154                   break;
155                 }
156                 *s++ = *p;
157               }
158               *s++ = '\0';
159             } else {
160               if (next - p > CHARSET_MAX - 1)
161                 break;
162               while (p != next)
163                 *s++ = *p++;
164               *s = 0;
165               break;
166             }
167           }
168         }
169         break;
170       }
171     } else
172       p = getTok(&next);
173   }
174 }
175 
176 #ifdef TEST
177 
178 #  include <stdio.h>
179 
180 int
main(int argc,char * argv[])181 main(int argc, char *argv[]) {
182   char buf[CHARSET_MAX];
183   if (argc <= 1)
184     return 1;
185   printf("%s\n", argv[1]);
186   getXMLCharset(argv[1], buf);
187   printf("charset=\"%s\"\n", buf);
188   return 0;
189 }
190 
191 #endif /* TEST */
192