]>
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 | ||
32 | static 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 | ||
730220de TR |
51 | static char *url_decode_internal(const char **query, const char *stop_at, |
52 | struct strbuf *out, int decode_plus) | |
638794cd JK |
53 | { |
54 | const char *q = *query; | |
638794cd | 55 | |
638794cd JK |
56 | do { |
57 | unsigned char c = *q; | |
58 | ||
59 | if (!c) | |
60 | break; | |
61 | if (stop_at && strchr(stop_at, c)) { | |
62 | q++; | |
63 | break; | |
64 | } | |
65 | ||
66 | if (c == '%') { | |
67 | int val = url_decode_char(q + 1); | |
68 | if (0 <= val) { | |
ce83eda1 | 69 | strbuf_addch(out, val); |
638794cd JK |
70 | q += 3; |
71 | continue; | |
72 | } | |
73 | } | |
74 | ||
730220de | 75 | if (decode_plus && c == '+') |
ce83eda1 | 76 | strbuf_addch(out, ' '); |
638794cd | 77 | else |
ce83eda1 | 78 | strbuf_addch(out, c); |
638794cd JK |
79 | q++; |
80 | } while (1); | |
81 | *query = q; | |
ce83eda1 | 82 | return strbuf_detach(out, NULL); |
638794cd JK |
83 | } |
84 | ||
85 | char *url_decode(const char *url) | |
86 | { | |
ce83eda1 | 87 | struct strbuf out = STRBUF_INIT; |
3c73a1d5 | 88 | const char *colon = strchr(url, ':'); |
ce83eda1 JH |
89 | |
90 | /* Skip protocol part if present */ | |
3c73a1d5 JH |
91 | if (colon && url < colon) { |
92 | strbuf_add(&out, url, colon - url); | |
93 | url = colon; | |
ce83eda1 | 94 | } |
730220de | 95 | return url_decode_internal(&url, NULL, &out, 0); |
638794cd JK |
96 | } |
97 | ||
98 | char *url_decode_parameter_name(const char **query) | |
99 | { | |
ce83eda1 | 100 | struct strbuf out = STRBUF_INIT; |
730220de | 101 | return url_decode_internal(query, "&=", &out, 1); |
638794cd JK |
102 | } |
103 | ||
104 | char *url_decode_parameter_value(const char **query) | |
105 | { | |
ce83eda1 | 106 | struct strbuf out = STRBUF_INIT; |
730220de | 107 | return url_decode_internal(query, "&", &out, 1); |
638794cd | 108 | } |
1966d9f3 TRC |
109 | |
110 | void end_url_with_slash(struct strbuf *buf, const char *url) | |
111 | { | |
112 | strbuf_addstr(buf, url); | |
113 | if (buf->len && buf->buf[buf->len - 1] != '/') | |
114 | strbuf_addstr(buf, "/"); | |
115 | } | |
3793a309 TRC |
116 | |
117 | void str_end_url_with_slash(const char *url, char **dest) { | |
118 | struct strbuf buf = STRBUF_INIT; | |
119 | end_url_with_slash(&buf, url); | |
120 | free(*dest); | |
121 | *dest = strbuf_detach(&buf, NULL); | |
122 | } |