]>
Commit | Line | Data |
---|---|---|
638794cd | 1 | #include "cache.h" |
c2e86add | 2 | #include "url.h" |
638794cd JK |
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 | { | |
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 | ||
66c84485 JK |
32 | static char *url_decode_internal(const char **query, int len, |
33 | const char *stop_at, struct strbuf *out, | |
34 | int decode_plus) | |
638794cd JK |
35 | { |
36 | const char *q = *query; | |
638794cd | 37 | |
66c84485 | 38 | while (len) { |
638794cd JK |
39 | unsigned char c = *q; |
40 | ||
41 | if (!c) | |
42 | break; | |
43 | if (stop_at && strchr(stop_at, c)) { | |
44 | q++; | |
66c84485 | 45 | len--; |
638794cd JK |
46 | break; |
47 | } | |
48 | ||
49 | if (c == '%') { | |
d2330973 | 50 | int val = hex2chr(q + 1); |
638794cd | 51 | if (0 <= val) { |
ce83eda1 | 52 | strbuf_addch(out, val); |
638794cd | 53 | q += 3; |
66c84485 | 54 | len -= 3; |
638794cd JK |
55 | continue; |
56 | } | |
57 | } | |
58 | ||
730220de | 59 | if (decode_plus && c == '+') |
ce83eda1 | 60 | strbuf_addch(out, ' '); |
638794cd | 61 | else |
ce83eda1 | 62 | strbuf_addch(out, c); |
638794cd | 63 | q++; |
66c84485 JK |
64 | len--; |
65 | } | |
638794cd | 66 | *query = q; |
ce83eda1 | 67 | return strbuf_detach(out, NULL); |
638794cd JK |
68 | } |
69 | ||
70 | char *url_decode(const char *url) | |
66c84485 JK |
71 | { |
72 | return url_decode_mem(url, strlen(url)); | |
73 | } | |
74 | ||
75 | char *url_decode_mem(const char *url, int len) | |
638794cd | 76 | { |
ce83eda1 | 77 | struct strbuf out = STRBUF_INIT; |
66c84485 | 78 | const char *colon = memchr(url, ':', len); |
ce83eda1 JH |
79 | |
80 | /* Skip protocol part if present */ | |
3c73a1d5 JH |
81 | if (colon && url < colon) { |
82 | strbuf_add(&out, url, colon - url); | |
66c84485 | 83 | len -= colon - url; |
3c73a1d5 | 84 | url = colon; |
ce83eda1 | 85 | } |
66c84485 | 86 | return url_decode_internal(&url, len, NULL, &out, 0); |
638794cd JK |
87 | } |
88 | ||
89 | char *url_decode_parameter_name(const char **query) | |
90 | { | |
ce83eda1 | 91 | struct strbuf out = STRBUF_INIT; |
66c84485 | 92 | return url_decode_internal(query, -1, "&=", &out, 1); |
638794cd JK |
93 | } |
94 | ||
95 | char *url_decode_parameter_value(const char **query) | |
96 | { | |
ce83eda1 | 97 | struct strbuf out = STRBUF_INIT; |
66c84485 | 98 | return url_decode_internal(query, -1, "&", &out, 1); |
638794cd | 99 | } |
1966d9f3 TRC |
100 | |
101 | void end_url_with_slash(struct strbuf *buf, const char *url) | |
102 | { | |
103 | strbuf_addstr(buf, url); | |
00b6c178 | 104 | strbuf_complete(buf, '/'); |
1966d9f3 | 105 | } |
3793a309 | 106 | |
3b335762 NTND |
107 | void str_end_url_with_slash(const char *url, char **dest) |
108 | { | |
3793a309 TRC |
109 | struct strbuf buf = STRBUF_INIT; |
110 | end_url_with_slash(&buf, url); | |
111 | free(*dest); | |
112 | *dest = strbuf_detach(&buf, NULL); | |
113 | } |