url.c

   1 #include "git-compat-util.h"
   2 #include "hex.h"
   3 #include "strbuf.h"
   4 #include "url.h"
   5
   6 int is_urlschemechar(int first_flag, int ch)
   7 {
   8         /*
   9          * The set of valid URL schemes, as per STD66 (RFC3986) is
  10          * '[A-Za-z][A-Za-z0-9+.-]*'. But use slightly looser check
  11          * of '[A-Za-z0-9][A-Za-z0-9+.-]*' because earlier version
  12          * of check used '[A-Za-z0-9]+' so not to break any remote
  13          * helpers.
  14          */
  15         int alphanumeric, special;
  16         alphanumeric = ch > 0 && isalnum(ch);
  17         special = ch == '+' || ch == '-' || ch == '.';
  18         return alphanumeric || (!first_flag && special);
  19 }
  20
  21 int is_url(const char *url)
  22 {
  23         /* Is "scheme" part reasonable? */
  24         if (!url || !is_urlschemechar(1, *url++))
  25                 return 0;
  26         while (*url && *url != ':') {
  27                 if (!is_urlschemechar(0, *url++))
  28                         return 0;
  29         }
  30         /* We've seen "scheme"; we want colon-slash-slash */
  31         return (url[0] == ':' && url[1] == '/' && url[2] == '/');
  32 }
  33
  34 static char *url_decode_internal(const char **query, int len,
  35                                  const char *stop_at, struct strbuf *out,
  36                                  int decode_plus)
  37 {
  38         const char *q = *query;
  39
  40         while (len) {
  41                 unsigned char c = *q;
  42
  43                 if (!c)
  44                         break;
  45                 if (stop_at && strchr(stop_at, c)) {
  46                         q++;
  47                         len--;
  48                         break;
  49                 }
  50
  51                 if (c == '%' && (len < 0 || len >= 3)) {
  52                         int val = hex2chr(q + 1);
  53                         if (0 < val) {
  54                                 strbuf_addch(out, val);
  55                                 q += 3;
  56                                 len -= 3;
  57                                 continue;
  58                         }
  59                 }
  60
  61                 if (decode_plus && c == '+')
  62                         strbuf_addch(out, ' ');
  63                 else
  64                         strbuf_addch(out, c);
  65                 q++;
  66                 len--;
  67         }
  68         *query = q;
  69         return strbuf_detach(out, NULL);
  70 }
  71
  72 char *url_decode(const char *url)
  73 {
  74         return url_decode_mem(url, strlen(url));
  75 }
  76
  77 char *url_decode_mem(const char *url, int len)
  78 {
  79         struct strbuf out = STRBUF_INIT;
  80         const char *colon = memchr(url, ':', len);
  81
  82         /* Skip protocol part if present */
  83         if (colon && url < colon) {
  84                 strbuf_add(&out, url, colon - url);
  85                 len -= colon - url;
  86                 url = colon;
  87         }
  88         return url_decode_internal(&url, len, NULL, &out, 0);
  89 }
  90
  91 char *url_percent_decode(const char *encoded)
  92 {
  93         struct strbuf out = STRBUF_INIT;
  94         return url_decode_internal(&encoded, strlen(encoded), NULL, &out, 0);
  95 }
  96
  97 char *url_decode_parameter_name(const char **query)
  98 {
  99         struct strbuf out = STRBUF_INIT;
 100         return url_decode_internal(query, -1, "&=", &out, 1);
 101 }
 102
 103 char *url_decode_parameter_value(const char **query)
 104 {
 105         struct strbuf out = STRBUF_INIT;
 106         return url_decode_internal(query, -1, "&", &out, 1);
 107 }
 108
 109 void end_url_with_slash(struct strbuf *buf, const char *url)
 110 {
 111         strbuf_addstr(buf, url);
 112         strbuf_complete(buf, '/');
 113 }
 114
 115 void str_end_url_with_slash(const char *url, char **dest)
 116 {
 117         struct strbuf buf = STRBUF_INIT;
 118         end_url_with_slash(&buf, url);
 119         free(*dest);
 120         *dest = strbuf_detach(&buf, NULL);
 121 }