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