url.c

   1 #include "cache.h"
   2 #include "hex.h"
   3 #include "url.h"
   4
   5 int is_urlschemechar(int first_flag, int ch)
   6 {
   7         /*
   8          * The set of valid URL schemes, as per STD66 (RFC3986) is
   9          * '[A-Za-z][A-Za-z0-9+.-]*'. But use slightly looser check
  10          * of '[A-Za-z0-9][A-Za-z0-9+.-]*' because earlier version
  11          * of check used '[A-Za-z0-9]+' so not to break any remote
  12          * helpers.
  13          */
  14         int alphanumeric, special;
  15         alphanumeric = ch > 0 && isalnum(ch);
  16         special = ch == '+' || ch == '-' || ch == '.';
  17         return alphanumeric || (!first_flag && special);
  18 }
  19
  20 int is_url(const char *url)
  21 {
  22         /* Is "scheme" part reasonable? */
  23         if (!url || !is_urlschemechar(1, *url++))
  24                 return 0;
  25         while (*url && *url != ':') {
  26                 if (!is_urlschemechar(0, *url++))
  27                         return 0;
  28         }
  29         /* We've seen "scheme"; we want colon-slash-slash */
  30         return (url[0] == ':' && url[1] == '/' && url[2] == '/');
  31 }
  32
  33 static char *url_decode_internal(const char **query, int len,
  34                                  const char *stop_at, struct strbuf *out,
  35                                  int decode_plus)
  36 {
  37         const char *q = *query;
  38
  39         while (len) {
  40                 unsigned char c = *q;
  41
  42                 if (!c)
  43                         break;
  44                 if (stop_at && strchr(stop_at, c)) {
  45                         q++;
  46                         len--;
  47                         break;
  48                 }
  49
  50                 if (c == '%' && (len < 0 || len >= 3)) {
  51                         int val = hex2chr(q + 1);
  52                         if (0 < val) {
  53                                 strbuf_addch(out, val);
  54                                 q += 3;
  55                                 len -= 3;
  56                                 continue;
  57                         }
  58                 }
  59
  60                 if (decode_plus && c == '+')
  61                         strbuf_addch(out, ' ');
  62                 else
  63                         strbuf_addch(out, c);
  64                 q++;
  65                 len--;
  66         }
  67         *query = q;
  68         return strbuf_detach(out, NULL);
  69 }
  70
  71 char *url_decode(const char *url)
  72 {
  73         return url_decode_mem(url, strlen(url));
  74 }
  75
  76 char *url_decode_mem(const char *url, int len)
  77 {
  78         struct strbuf out = STRBUF_INIT;
  79         const char *colon = memchr(url, ':', len);
  80
  81         /* Skip protocol part if present */
  82         if (colon && url < colon) {
  83                 strbuf_add(&out, url, colon - url);
  84                 len -= colon - url;
  85                 url = colon;
  86         }
  87         return url_decode_internal(&url, len, NULL, &out, 0);
  88 }
  89
  90 char *url_percent_decode(const char *encoded)
  91 {
  92         struct strbuf out = STRBUF_INIT;
  93         return url_decode_internal(&encoded, strlen(encoded), NULL, &out, 0);
  94 }
  95
  96 char *url_decode_parameter_name(const char **query)
  97 {
  98         struct strbuf out = STRBUF_INIT;
  99         return url_decode_internal(query, -1, "&=", &out, 1);
 100 }
 101
 102 char *url_decode_parameter_value(const char **query)
 103 {
 104         struct strbuf out = STRBUF_INIT;
 105         return url_decode_internal(query, -1, "&", &out, 1);
 106 }
 107
 108 void end_url_with_slash(struct strbuf *buf, const char *url)
 109 {
 110         strbuf_addstr(buf, url);
 111         strbuf_complete(buf, '/');
 112 }
 113
 114 void str_end_url_with_slash(const char *url, char **dest)
 115 {
 116         struct strbuf buf = STRBUF_INIT;
 117         end_url_with_slash(&buf, url);
 118         free(*dest);
 119         *dest = strbuf_detach(&buf, NULL);
 120 }