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 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 {
  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 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
  51 static char *url_decode_internal(const char **query, int len,
  52                                  const char *stop_at, struct strbuf *out,
  53                                  int decode_plus)
  54 {
  55         const char *q = *query;
  56
  57         while (len) {
  58                 unsigned char c = *q;
  59
  60                 if (!c)
  61                         break;
  62                 if (stop_at && strchr(stop_at, c)) {
  63                         q++;
  64                         len--;
  65                         break;
  66                 }
  67
  68                 if (c == '%') {
  69                         int val = url_decode_char(q + 1);
  70                         if (0 <= val) {
  71                                 strbuf_addch(out, val);
  72                                 q += 3;
  73                                 len -= 3;
  74                                 continue;
  75                         }
  76                 }
  77
  78                 if (decode_plus && c == '+')
  79                         strbuf_addch(out, ' ');
  80                 else
  81                         strbuf_addch(out, c);
  82                 q++;
  83                 len--;
  84         }
  85         *query = q;
  86         return strbuf_detach(out, NULL);
  87 }
  88
  89 char *url_decode(const char *url)
  90 {
  91         return url_decode_mem(url, strlen(url));
  92 }
  93
  94 char *url_decode_mem(const char *url, int len)
  95 {
  96         struct strbuf out = STRBUF_INIT;
  97         const char *colon = memchr(url, ':', len);
  98
  99         /* Skip protocol part if present */
 100         if (colon && url < colon) {
 101                 strbuf_add(&out, url, colon - url);
 102                 len -= colon - url;
 103                 url = colon;
 104         }
 105         return url_decode_internal(&url, len, NULL, &out, 0);
 106 }
 107
 108 char *url_decode_parameter_name(const char **query)
 109 {
 110         struct strbuf out = STRBUF_INIT;
 111         return url_decode_internal(query, -1, "&=", &out, 1);
 112 }
 113
 114 char *url_decode_parameter_value(const char **query)
 115 {
 116         struct strbuf out = STRBUF_INIT;
 117         return url_decode_internal(query, -1, "&", &out, 1);
 118 }
 119
 120 void end_url_with_slash(struct strbuf *buf, const char *url)
 121 {
 122         strbuf_addstr(buf, url);
 123         if (buf->len && buf->buf[buf->len - 1] != '/')
 124                 strbuf_addch(buf, '/');
 125 }
 126
 127 void str_end_url_with_slash(const char *url, char **dest) {
 128         struct strbuf buf = STRBUF_INIT;
 129         end_url_with_slash(&buf, url);
 130         free(*dest);
 131         *dest = strbuf_detach(&buf, NULL);
 132 }