1266733Speter/* Licensed to the Apache Software Foundation (ASF) under one or more
2266733Speter * contributor license agreements.  See the NOTICE file distributed with
3266733Speter * this work for additional information regarding copyright ownership.
4266733Speter * The ASF licenses this file to You under the Apache License, Version 2.0
5266733Speter * (the "License"); you may not use this file except in compliance with
6266733Speter * the License.  You may obtain a copy of the License at
7266733Speter *
8266733Speter *     http://www.apache.org/licenses/LICENSE-2.0
9266733Speter *
10266733Speter * Unless required by applicable law or agreed to in writing, software
11266733Speter * distributed under the License is distributed on an "AS IS" BASIS,
12266733Speter * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
13266733Speter * See the License for the specific language governing permissions and
14266733Speter * limitations under the License.
15266733Speter */
16266733Speter
17266733Speter#if defined(WIN32) || defined(OS2)
18266733Speter#define NEED_ENHANCED_ESCAPES
19266733Speter#endif
20266733Speter
21266733Speter#include <stdio.h>
22266733Speter#include <string.h>
23266733Speter#include <stdio.h>
24266733Speter#include <ctype.h>
25266733Speter
26266733Speter/* A bunch of functions in util.c scan strings looking for certain characters.
27266733Speter * To make that more efficient we encode a lookup table.
28266733Speter */
29266733Speter#define T_ESCAPE_SHELL_CMD    (0x01)
30266733Speter#define T_ESCAPE_PATH_SEGMENT (0x02)
31266733Speter#define T_OS_ESCAPE_PATH      (0x04)
32266733Speter#define T_ESCAPE_ECHO         (0x08)
33266733Speter#define T_ESCAPE_URLENCODED   (0x10)
34266733Speter#define T_ESCAPE_XML          (0x20)
35266733Speter
36266733Speterint main(int argc, char *argv[])
37266733Speter{
38266733Speter    unsigned c;
39266733Speter    unsigned char flags;
40266733Speter
41266733Speter    printf("/* this file is automatically generated by gen_test_char, "
42266733Speter           "do not edit. \"make include/private/apr_escape_test_char.h\" to regenerate. */\n"
43266733Speter           "#define T_ESCAPE_SHELL_CMD     (%u)\n"
44266733Speter           "#define T_ESCAPE_PATH_SEGMENT  (%u)\n"
45266733Speter           "#define T_OS_ESCAPE_PATH       (%u)\n"
46266733Speter           "#define T_ESCAPE_ECHO          (%u)\n"
47266733Speter           "#define T_ESCAPE_URLENCODED    (%u)\n"
48266733Speter           "#define T_ESCAPE_XML           (%u)\n"
49266733Speter           "\n"
50266733Speter           "static const unsigned char test_char_table[256] = {",
51266733Speter           T_ESCAPE_SHELL_CMD,
52266733Speter           T_ESCAPE_PATH_SEGMENT,
53266733Speter           T_OS_ESCAPE_PATH,
54266733Speter           T_ESCAPE_ECHO,
55266733Speter           T_ESCAPE_URLENCODED,
56266733Speter           T_ESCAPE_XML);
57266733Speter
58266733Speter    for (c = 0; c < 256; ++c) {
59266733Speter        flags = 0;
60266733Speter        if (c % 20 == 0)
61266733Speter            printf("\n    ");
62266733Speter
63266733Speter        /* escape_shell_cmd */
64266733Speter#ifdef NEED_ENHANCED_ESCAPES
65266733Speter        /* Win32/OS2 have many of the same vulnerable characters
66266733Speter         * as Unix sh, plus the carriage return and percent char.
67266733Speter         * The proper escaping of these characters varies from unix
68266733Speter         * since Win32/OS2 use carets or doubled-double quotes,
69266733Speter         * and neither lf nor cr can be escaped.  We escape unix
70266733Speter         * specific as well, to assure that cross-compiled unix
71266733Speter         * applications behave similiarly when invoked on win32/os2.
72266733Speter         *
73266733Speter         * Rem please keep in-sync with apr's list in win32/filesys.c
74266733Speter         */
75266733Speter        if (c && strchr("&;`'\"|*?~<>^()[]{}$\\\n\r%", c)) {
76266733Speter            flags |= T_ESCAPE_SHELL_CMD;
77266733Speter        }
78266733Speter#else
79266733Speter        if (c && strchr("&;`'\"|*?~<>^()[]{}$\\\n", c)) {
80266733Speter            flags |= T_ESCAPE_SHELL_CMD;
81266733Speter        }
82266733Speter#endif
83266733Speter
84266733Speter        if (!isalnum(c) && !strchr("$-_.+!*'(),:@&=~", c)) {
85266733Speter            flags |= T_ESCAPE_PATH_SEGMENT;
86266733Speter        }
87266733Speter
88266733Speter        if (!isalnum(c) && !strchr("$-_.+!*'(),:@&=/~", c)) {
89266733Speter            flags |= T_OS_ESCAPE_PATH;
90266733Speter        }
91266733Speter
92266733Speter        if (!isalnum(c) && !strchr(".-*_ ", c)) {
93266733Speter            flags |= T_ESCAPE_URLENCODED;
94266733Speter        }
95266733Speter
96266733Speter        /* For logging, escape all control characters,
97266733Speter         * double quotes (because they delimit the request in the log file)
98266733Speter         * backslashes (because we use backslash for escaping)
99266733Speter         * and 8-bit chars with the high bit set
100266733Speter         */
101266733Speter        if (c && (!isprint(c) || c == '"' || c == '\\' || iscntrl(c))) {
102266733Speter            flags |= T_ESCAPE_ECHO;
103266733Speter        }
104266733Speter
105266733Speter        if (strchr("<>&\"", c)) {
106266733Speter            flags |= T_ESCAPE_XML;
107266733Speter        }
108266733Speter
109266733Speter        printf("%u%c", flags, (c < 255) ? ',' : ' ');
110266733Speter    }
111266733Speter
112266733Speter    printf("\n};\n");
113266733Speter
114266733Speter    return 0;
115266733Speter}
116