url: decode buffers that are not NUL-terminated
[git/git.git] / url.c
1 #include "cache.h"
2 #include "url.h"
3
4 int is_urlschemechar(int first_flag, int ch)
5 {
6 /*
7 * The set of valid URL schemes, as per STD66 (RFC3986) is
8 * '[A-Za-z][A-Za-z0-9+.-]*'. But use sightly looser check
9 * of '[A-Za-z0-9][A-Za-z0-9+.-]*' because earlier version
10 * of check used '[A-Za-z0-9]+' so not to break any remote
11 * helpers.
12 */
13 int alphanumeric, special;
14 alphanumeric = ch > 0 && isalnum(ch);
15 special = ch == '+' || ch == '-' || ch == '.';
16 return alphanumeric || (!first_flag && special);
17 }
18
19 int is_url(const char *url)
20 {
21 const char *url2, *first_slash;
22
23 if (!url)
24 return 0;
25 url2 = url;
26 first_slash = strchr(url, '/');
27
28 /* Input with no slash at all or slash first can't be URL. */
29 if (!first_slash || first_slash == url)
30 return 0;
31 /* Character before must be : and next must be /. */
32 if (first_slash[-1] != ':' || first_slash[1] != '/')
33 return 0;
34 /* There must be something before the :// */
35 if (first_slash == url + 1)
36 return 0;
37 /*
38 * Check all characters up to first slash - 1. Only alphanum
39 * is allowed.
40 */
41 url2 = url;
42 while (url2 < first_slash - 1) {
43 if (!is_urlschemechar(url2 == url, (unsigned char)*url2))
44 return 0;
45 url2++;
46 }
47
48 /* Valid enough. */
49 return 1;
50 }
51
52 static int url_decode_char(const char *q)
53 {
54 int i;
55 unsigned char val = 0;
56 for (i = 0; i < 2; i++) {
57 unsigned char c = *q++;
58 val <<= 4;
59 if (c >= '0' && c <= '9')
60 val += c - '0';
61 else if (c >= 'a' && c <= 'f')
62 val += c - 'a' + 10;
63 else if (c >= 'A' && c <= 'F')
64 val += c - 'A' + 10;
65 else
66 return -1;
67 }
68 return val;
69 }
70
71 static char *url_decode_internal(const char **query, int len,
72 const char *stop_at, struct strbuf *out,
73 int decode_plus)
74 {
75 const char *q = *query;
76
77 while (len) {
78 unsigned char c = *q;
79
80 if (!c)
81 break;
82 if (stop_at && strchr(stop_at, c)) {
83 q++;
84 len--;
85 break;
86 }
87
88 if (c == '%') {
89 int val = url_decode_char(q + 1);
90 if (0 <= val) {
91 strbuf_addch(out, val);
92 q += 3;
93 len -= 3;
94 continue;
95 }
96 }
97
98 if (decode_plus && c == '+')
99 strbuf_addch(out, ' ');
100 else
101 strbuf_addch(out, c);
102 q++;
103 len--;
104 }
105 *query = q;
106 return strbuf_detach(out, NULL);
107 }
108
109 char *url_decode(const char *url)
110 {
111 return url_decode_mem(url, strlen(url));
112 }
113
114 char *url_decode_mem(const char *url, int len)
115 {
116 struct strbuf out = STRBUF_INIT;
117 const char *colon = memchr(url, ':', len);
118
119 /* Skip protocol part if present */
120 if (colon && url < colon) {
121 strbuf_add(&out, url, colon - url);
122 len -= colon - url;
123 url = colon;
124 }
125 return url_decode_internal(&url, len, NULL, &out, 0);
126 }
127
128 char *url_decode_parameter_name(const char **query)
129 {
130 struct strbuf out = STRBUF_INIT;
131 return url_decode_internal(query, -1, "&=", &out, 1);
132 }
133
134 char *url_decode_parameter_value(const char **query)
135 {
136 struct strbuf out = STRBUF_INIT;
137 return url_decode_internal(query, -1, "&", &out, 1);
138 }
139
140 void end_url_with_slash(struct strbuf *buf, const char *url)
141 {
142 strbuf_addstr(buf, url);
143 if (buf->len && buf->buf[buf->len - 1] != '/')
144 strbuf_addstr(buf, "/");
145 }
146
147 void str_end_url_with_slash(const char *url, char **dest) {
148 struct strbuf buf = STRBUF_INIT;
149 end_url_with_slash(&buf, url);
150 free(*dest);
151 *dest = strbuf_detach(&buf, NULL);
152 }