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 slightly 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         /* Is "scheme" part reasonable? */
  22         if (!url || !is_urlschemechar(1, *url++))
  23                 return 0;
  24         while (*url && *url != ':') {
  25                 if (!is_urlschemechar(0, *url++))
  26                         return 0;
  27         }
  28         /* We've seen "scheme"; we want colon-slash-slash */
  29         return (url[0] == ':' && url[1] == '/' && url[2] == '/');
  30 }
  31
  32 static char *url_decode_internal(const char **query, int len,
  33                                  const char *stop_at, struct strbuf *out,
  34                                  int decode_plus)
  35 {
  36         const char *q = *query;
  37
  38         while (len) {
  39                 unsigned char c = *q;
  40
  41                 if (!c)
  42                         break;
  43                 if (stop_at && strchr(stop_at, c)) {
  44                         q++;
  45                         len--;
  46                         break;
  47                 }
  48
  49                 if (c == '%' && (len < 0 || len >= 3)) {
  50                         int val = hex2chr(q + 1);
  51                         if (0 < val) {
  52                                 strbuf_addch(out, val);
  53                                 q += 3;
  54                                 len -= 3;
  55                                 continue;
  56                         }
  57                 }
  58
  59                 if (decode_plus && c == '+')
  60                         strbuf_addch(out, ' ');
  61                 else
  62                         strbuf_addch(out, c);
  63                 q++;
  64                 len--;
  65         }
  66         *query = q;
  67         return strbuf_detach(out, NULL);
  68 }
  69
  70 char *url_decode(const char *url)
  71 {
  72         return url_decode_mem(url, strlen(url));
  73 }
  74
  75 char *url_decode_mem(const char *url, int len)
  76 {
  77         struct strbuf out = STRBUF_INIT;
  78         const char *colon = memchr(url, ':', len);
  79
  80         /* Skip protocol part if present */
  81         if (colon && url < colon) {
  82                 strbuf_add(&out, url, colon - url);
  83                 len -= colon - url;
  84                 url = colon;
  85         }
  86         return url_decode_internal(&url, len, NULL, &out, 0);
  87 }
  88
  89 char *url_percent_decode(const char *encoded)
  90 {
  91         struct strbuf out = STRBUF_INIT;
  92         return url_decode_internal(&encoded, strlen(encoded), NULL, &out, 0);
  93 }
  94
  95 char *url_decode_parameter_name(const char **query)
  96 {
  97         struct strbuf out = STRBUF_INIT;
  98         return url_decode_internal(query, -1, "&=", &out, 1);
  99 }
 100
 101 char *url_decode_parameter_value(const char **query)
 102 {
 103         struct strbuf out = STRBUF_INIT;
 104         return url_decode_internal(query, -1, "&", &out, 1);
 105 }
 106
 107 void end_url_with_slash(struct strbuf *buf, const char *url)
 108 {
 109         strbuf_addstr(buf, url);
 110         strbuf_complete(buf, '/');
 111 }
 112
 113 void str_end_url_with_slash(const char *url, char **dest)
 114 {
 115         struct strbuf buf = STRBUF_INIT;
 116         end_url_with_slash(&buf, url);
 117         free(*dest);
 118         *dest = strbuf_detach(&buf, NULL);
 119 }