1 files changed, 19 insertions, 0 deletions

diff --git a/lib/util.cc b/lib/util.cc
index 54d6535..b7bc437 100644
--- a/lib/util.cc
+++ b/lib/util.cc
@@ -132,64 +132,83 @@ namespace opkele {
                             ) != 6
                 ) && (
                     sscanf(
                         w.c_str(),
                         "%04d-%02d-%02dT%02d:%02d:%02d.%03dZ",
                         &tm_t.tm_year,&tm_t.tm_mon,&tm_t.tm_mday,
                         &tm_t.tm_hour,&tm_t.tm_min,&tm_t.tm_sec,
                         &fraction
                         ) != 7
                     ) )
                 throw failed_conversion(OPKELE_CP_ "failed to sscanf()");
             tm_t.tm_mon--;
             tm_t.tm_year-=1900;
             time_t rv = mktime(&tm_t);
             if(rv==(time_t)-1)
                 throw failed_conversion(OPKELE_CP_ "failed to mktime()");
             return rv-timezone;
         }
 
         /*
          *
          */
 
         string url_encode(const string& str) {
             char * t = curl_escape(str.c_str(),str.length());
             if(!t)
                 throw failed_conversion(OPKELE_CP_ "failed to curl_escape()");
             string rv(t);
             curl_free(t);
             return rv;
         }
 
+        string attr_escape(const string& str) {
+            static const char *unsafechars = "<>&\n\"'";
+            string rv;
+            string::size_type p=0;
+            while(true) {
+                string::size_type us = str.find_first_of(unsafechars,p);
+                if(us==string::npos) {
+                    if(p!=str.length())
+                        rv.append(str,p,str.length()-p);
+                    return rv;
+                }
+                rv.append(str,p,us-p);
+                rv += "&#";
+                rv += long_to_string((long)str[us]);
+                rv += ';';
+                p = us+1;
+            }
+        }
+
         string long_to_string(long l) {
             char rv[32];
             int r=snprintf(rv,sizeof(rv),"%ld",l);
             if(r<0 || r>=(int)sizeof(rv))
                 throw failed_conversion(OPKELE_CP_ "failed to snprintf()");
             return rv;
         }
 
         long string_to_long(const string& s) {
             char *endptr = 0;
             long rv = strtol(s.c_str(),&endptr,10);
             if((!endptr) || endptr==s.c_str())
                 throw failed_conversion(OPKELE_CP_ "failed to strtol()");
             return rv;
         }
 
         /*
          * Normalize URL according to the rules, described in rfc 3986, section 6
          *
          * - uppercase hex triplets (e.g. %ab -> %AB)
          * - lowercase scheme and host
          * - decode %-encoded characters, specified as unreserved in rfc 3986, section 2.3,
          *   that is - [:alpha:][:digit:]._~-
          * - remove dot segments
          * - remove empty and default ports
          * - if there's no path component, add '/'
          */
          string rfc_3986_normalize_uri(const string& uri) {
              static const char *whitespace = " \t\r\n";
              string rv;
              string::size_type ns = uri.find_first_not_of(whitespace);
              if(ns==string::npos)