url.c

   1 #include "cache.h"
   2
   3 int is_urlschemechar(int first_flag, int ch)
   4 {
   5         /*
   6          * The set of valid URL schemes, as per STD66 (RFC3986) is
   7          * '[A-Za-z][A-Za-z0-9+.-]*'. But use sightly looser check
   8          * of '[A-Za-z0-9][A-Za-z0-9+.-]*' because earlier version
   9          * of check used '[A-Za-z0-9]+' so not to break any remote
  10          * helpers.
  11          */
  12         int alphanumeric, special;
  13         alphanumeric = ch > 0 && isalnum(ch);
  14         special = ch == '+' || ch == '-' || ch == '.';
  15         return alphanumeric || (!first_flag && special);
  16 }
  17
  18 int is_url(const char *url)
  19 {
  20         const char *url2, *first_slash;
  21
  22         if (!url)
  23                 return 0;
  24         url2 = url;
  25         first_slash = strchr(url, '/');
  26
  27         /* Input with no slash at all or slash first can't be URL. */
  28         if (!first_slash || first_slash == url)
  29                 return 0;
  30         /* Character before must be : and next must be /. */
  31         if (first_slash[-1] != ':' || first_slash[1] != '/')
  32                 return 0;
  33         /* There must be something before the :// */
  34         if (first_slash == url + 1)
  35                 return 0;
  36         /*
  37          * Check all characters up to first slash - 1. Only alphanum
  38          * is allowed.
  39          */
  40         url2 = url;
  41         while (url2 < first_slash - 1) {
  42                 if (!is_urlschemechar(url2 == url, (unsigned char)*url2))
  43                         return 0;
  44                 url2++;
  45         }
  46
  47         /* Valid enough. */
  48         return 1;
  49 }
  50
  51 static int url_decode_char(const char *q)
  52 {
  53         int i;
  54         unsigned char val = 0;
  55         for (i = 0; i < 2; i++) {
  56                 unsigned char c = *q++;
  57                 val <<= 4;
  58                 if (c >= '0' && c <= '9')
  59                         val += c - '0';
  60                 else if (c >= 'a' && c <= 'f')
  61                         val += c - 'a' + 10;
  62                 else if (c >= 'A' && c <= 'F')
  63                         val += c - 'A' + 10;
  64                 else
  65                         return -1;
  66         }
  67         return val;
  68 }
  69
  70 static char *url_decode_internal(const char **query, const char *stop_at)
  71 {
  72         const char *q = *query;
  73         struct strbuf out;
  74
  75         strbuf_init(&out, 16);
  76         do {
  77                 unsigned char c = *q;
  78
  79                 if (!c)
  80                         break;
  81                 if (stop_at && strchr(stop_at, c)) {
  82                         q++;
  83                         break;
  84                 }
  85
  86                 if (c == '%') {
  87                         int val = url_decode_char(q + 1);
  88                         if (0 <= val) {
  89                                 strbuf_addch(&out, val);
  90                                 q += 3;
  91                                 continue;
  92                         }
  93                 }
  94
  95                 if (c == '+')
  96                         strbuf_addch(&out, ' ');
  97                 else
  98                         strbuf_addch(&out, c);
  99                 q++;
 100         } while (1);
 101         *query = q;
 102         return strbuf_detach(&out, NULL);
 103 }
 104
 105 char *url_decode(const char *url)
 106 {
 107         return url_decode_internal(&url, NULL);
 108 }
 109
 110 char *url_decode_parameter_name(const char **query)
 111 {
 112         return url_decode_internal(query, "&=");
 113 }
 114
 115 char *url_decode_parameter_value(const char **query)
 116 {
 117         return url_decode_internal(query, "&");
 118 }