added an identifier normalization utility function

* moved iname leader characters and whitespace characters strings to
  opkele::data namespace
* added opkele::util::normalize_identifier() function

Signed-off-by: Michael Krelin <hacker@klever.net>

3 files changed, 49 insertions, 12 deletions

diff --git a/lib/data.cc b/lib/data.cc
index c040430..f71788f 100644
--- a/lib/data.cc
+++ b/lib/data.cc
@@ -9,2 +9,5 @@ namespace opkele {
 
+	const char *_iname_leaders = "=@+$!(";
+	const char *_whitespace_chars = " \t\r\n";
+
     }
diff --git a/lib/discovery.cc b/lib/discovery.cc
index b7f2db6..5913ad4 100644
--- a/lib/discovery.cc
+++ b/lib/discovery.cc
@@ -8,2 +8,3 @@
 #include <opkele/tidy.h>
+#include <opkele/data.h>
 #include <opkele/debug.h>
@@ -23,4 +24,2 @@ namespace opkele {
 
-    static const char *whitespace = " \t\r\n";
-    static const char *i_leaders = "=@+$!(";
     static const size_t max_html = 16384;
@@ -130,6 +129,6 @@ namespace opkele {
 		idiscovery_t idis;
-		string::size_type fsc = identity.find_first_not_of(whitespace);
+		string::size_type fsc = identity.find_first_not_of(data::_whitespace_chars);
 		if(fsc==string::npos)
 		    throw bad_input(OPKELE_CP_ "whitespace-only identity");
-		string::size_type lsc = identity.find_last_not_of(whitespace);
+		string::size_type lsc = identity.find_last_not_of(data::_whitespace_chars);
 		assert(lsc!=string::npos);
@@ -141,3 +140,3 @@ namespace opkele {
 		idis.clear();
-		if(strchr(i_leaders,id[0])) {
+		if(strchr(data::_iname_leaders,id[0])) {
 		    /* TODO: further normalize xri identity? Like folding case
@@ -497,3 +496,3 @@ namespace opkele {
 			    href.assign(ns);
-			    string::size_type lns=href.find_last_not_of(whitespace);
+			    string::size_type lns=href.find_last_not_of(data::_whitespace_chars);
 			    href.erase(lns+1);
@@ -501,5 +500,5 @@ namespace opkele {
 		    }
-		    for(string::size_type ns=rels.find_first_not_of(whitespace);
-			    ns!=string::npos; ns=rels.find_first_not_of(whitespace,ns)) {
-			string::size_type s = rels.find_first_of(whitespace,ns);
+		    for(string::size_type ns=rels.find_first_not_of(data::_whitespace_chars);
+			    ns!=string::npos; ns=rels.find_first_not_of(data::_whitespace_chars,ns)) {
+			string::size_type s = rels.find_first_of(data::_whitespace_chars,ns);
 			string rel;
diff --git a/lib/util.cc b/lib/util.cc
index bb8a2e8..29e6738 100644
--- a/lib/util.cc
+++ b/lib/util.cc
@@ -15,2 +15,3 @@
 #include <opkele/exception.h>
+#include <opkele/data.h>
 #include <opkele/debug.h>
@@ -213,5 +214,4 @@ namespace opkele {
 	 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);
+	     string::size_type ns = uri.find_first_not_of(data::_whitespace_chars);
 	     if(ns==string::npos)
@@ -225,3 +225,3 @@ namespace opkele {
 	     bool s;
-	     string::size_type ul = uri.find_last_not_of(whitespace)+1;
+	     string::size_type ul = uri.find_last_not_of(data::_whitespace_chars)+1;
 	     if(ul <= (colon+3))
@@ -432,2 +432,37 @@ namespace opkele {
 
+	string normalize_identifier(const string& usi,bool strip_fragment) {
+	    if(usi.empty())
+		return usi;
+	    string rv;
+	    string::size_type fsc = usi.find_first_not_of(data::_whitespace_chars);
+	    if(fsc==string::npos)
+		return rv;
+	    string::size_type lsc = usi.find_last_not_of(data::_whitespace_chars);
+	    assert(lsc!=string::npos);
+	    if(!strncasecmp(usi.c_str()+fsc,"xri://",sizeof("xri://")-1))
+		fsc += sizeof("xri://")-1;
+	    if( (fsc+1) >= lsc )
+		return rv;
+	    rv.assign(usi,fsc,lsc-fsc+1);
+	    if(strchr(data::_iname_leaders,rv[0])) {
+		/* TODO: further normalize xri identity, fold case or
+		 * whatever... */
+	    }else{
+		if(rv.find("://")==string::npos)
+		    rv.insert(0,"http://");
+		if(strip_fragment) {
+		    string::size_type fp = rv.find('#');
+		    if(fp!=string::npos) {
+			string::size_type qp = rv.find('?');
+			if(qp==string::npos || qp<fp)
+			    rv.erase(fp);
+			else if(qp>fp)
+			    rv.erase(fp,qp-fp);
+		    }
+		}
+		rv = rfc_3986_normalize_uri(rv);
+	    }
+	    return rv;
+	}
+
     }