url.con commit Git 1.7.2.5 (b91779f)
   1#include "cache.h"
   2
   3int 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
  18int 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
  51static 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
  70static char *url_decode_internal(const char **query, const char *stop_at, struct strbuf *out)
  71{
  72        const char *q = *query;
  73
  74        do {
  75                unsigned char c = *q;
  76
  77                if (!c)
  78                        break;
  79                if (stop_at && strchr(stop_at, c)) {
  80                        q++;
  81                        break;
  82                }
  83
  84                if (c == '%') {
  85                        int val = url_decode_char(q + 1);
  86                        if (0 <= val) {
  87                                strbuf_addch(out, val);
  88                                q += 3;
  89                                continue;
  90                        }
  91                }
  92
  93                if (c == '+')
  94                        strbuf_addch(out, ' ');
  95                else
  96                        strbuf_addch(out, c);
  97                q++;
  98        } while (1);
  99        *query = q;
 100        return strbuf_detach(out, NULL);
 101}
 102
 103char *url_decode(const char *url)
 104{
 105        struct strbuf out = STRBUF_INIT;
 106        const char *colon = strchr(url, ':');
 107
 108        /* Skip protocol part if present */
 109        if (colon && url < colon) {
 110                strbuf_add(&out, url, colon - url);
 111                url = colon;
 112        }
 113        return url_decode_internal(&url, NULL, &out);
 114}
 115
 116char *url_decode_parameter_name(const char **query)
 117{
 118        struct strbuf out = STRBUF_INIT;
 119        return url_decode_internal(query, "&=", &out);
 120}
 121
 122char *url_decode_parameter_value(const char **query)
 123{
 124        struct strbuf out = STRBUF_INIT;
 125        return url_decode_internal(query, "&", &out);
 126}