Git 2.9.3
[git/git.git] / url.c
CommitLineData
638794cd 1#include "cache.h"
c2e86add 2#include "url.h"
638794cd
JK
3
4int 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
19int is_url(const char *url)
20{
b33a1b9f
JH
21 /* Is "scheme" part reasonable? */
22 if (!url || !is_urlschemechar(1, *url++))
638794cd 23 return 0;
b33a1b9f
JH
24 while (*url && *url != ':') {
25 if (!is_urlschemechar(0, *url++))
638794cd 26 return 0;
638794cd 27 }
b33a1b9f
JH
28 /* We've seen "scheme"; we want colon-slash-slash */
29 return (url[0] == ':' && url[1] == '/' && url[2] == '/');
638794cd
JK
30}
31
32static int url_decode_char(const char *q)
33{
34 int i;
35 unsigned char val = 0;
36 for (i = 0; i < 2; i++) {
37 unsigned char c = *q++;
38 val <<= 4;
39 if (c >= '0' && c <= '9')
40 val += c - '0';
41 else if (c >= 'a' && c <= 'f')
42 val += c - 'a' + 10;
43 else if (c >= 'A' && c <= 'F')
44 val += c - 'A' + 10;
45 else
46 return -1;
47 }
48 return val;
49}
50
66c84485
JK
51static char *url_decode_internal(const char **query, int len,
52 const char *stop_at, struct strbuf *out,
53 int decode_plus)
638794cd
JK
54{
55 const char *q = *query;
638794cd 56
66c84485 57 while (len) {
638794cd
JK
58 unsigned char c = *q;
59
60 if (!c)
61 break;
62 if (stop_at && strchr(stop_at, c)) {
63 q++;
66c84485 64 len--;
638794cd
JK
65 break;
66 }
67
68 if (c == '%') {
69 int val = url_decode_char(q + 1);
70 if (0 <= val) {
ce83eda1 71 strbuf_addch(out, val);
638794cd 72 q += 3;
66c84485 73 len -= 3;
638794cd
JK
74 continue;
75 }
76 }
77
730220de 78 if (decode_plus && c == '+')
ce83eda1 79 strbuf_addch(out, ' ');
638794cd 80 else
ce83eda1 81 strbuf_addch(out, c);
638794cd 82 q++;
66c84485
JK
83 len--;
84 }
638794cd 85 *query = q;
ce83eda1 86 return strbuf_detach(out, NULL);
638794cd
JK
87}
88
89char *url_decode(const char *url)
66c84485
JK
90{
91 return url_decode_mem(url, strlen(url));
92}
93
94char *url_decode_mem(const char *url, int len)
638794cd 95{
ce83eda1 96 struct strbuf out = STRBUF_INIT;
66c84485 97 const char *colon = memchr(url, ':', len);
ce83eda1
JH
98
99 /* Skip protocol part if present */
3c73a1d5
JH
100 if (colon && url < colon) {
101 strbuf_add(&out, url, colon - url);
66c84485 102 len -= colon - url;
3c73a1d5 103 url = colon;
ce83eda1 104 }
66c84485 105 return url_decode_internal(&url, len, NULL, &out, 0);
638794cd
JK
106}
107
108char *url_decode_parameter_name(const char **query)
109{
ce83eda1 110 struct strbuf out = STRBUF_INIT;
66c84485 111 return url_decode_internal(query, -1, "&=", &out, 1);
638794cd
JK
112}
113
114char *url_decode_parameter_value(const char **query)
115{
ce83eda1 116 struct strbuf out = STRBUF_INIT;
66c84485 117 return url_decode_internal(query, -1, "&", &out, 1);
638794cd 118}
1966d9f3
TRC
119
120void end_url_with_slash(struct strbuf *buf, const char *url)
121{
122 strbuf_addstr(buf, url);
00b6c178 123 strbuf_complete(buf, '/');
1966d9f3 124}
3793a309
TRC
125
126void str_end_url_with_slash(const char *url, char **dest) {
127 struct strbuf buf = STRBUF_INIT;
128 end_url_with_slash(&buf, url);
129 free(*dest);
130 *dest = strbuf_detach(&buf, NULL);
131}