Fix response rewriting
[metaproxy-moved-to-github.git] / src / test_filter_rewrite.cpp
index 4f6f0f4..7753f0e 100644 (file)
@@ -21,6 +21,7 @@ Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
 #include <stdexcept>
 
 #include "filter_http_client.hpp"
+#include "filter_http_rewrite.hpp"
 #include <metaproxy/util.hpp>
 #include "router_chain.hpp"
 #include <metaproxy/package.hpp>
@@ -28,6 +29,8 @@ Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
 #include <boost/regex.hpp>
 #include <boost/lexical_cast.hpp>
 
+#include <yaz/log.h>
+
 #define BOOST_AUTO_TEST_MAIN
 #define BOOST_TEST_DYN_LINK
 
@@ -36,253 +39,25 @@ Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
 using namespace boost::unit_test;
 namespace mp = metaproxy_1;
 
-class FilterHeaderRewrite: public mp::filter::Base {
-public:
-    void process(mp::Package & package) const {
-        Z_GDU *gdu = package.request().get();
-        //map of request/response vars
-        std::map<std::string, std::string> vars;
-        //we have an http req
-        if (gdu && gdu->which == Z_GDU_HTTP_Request)
-        {
-            Z_HTTP_Request *hreq = gdu->u.HTTP_Request;
-            mp::odr o;
-            //rewrite the request line
-            std::string path;
-            if (strstr(hreq->path, "http://") == hreq->path)
-            {
-                std::cout << "Path in the method line is absolute, " 
-                    "possibly a proxy request\n";
-                path += hreq->path;
-            }
-            else
-            {
-               path += z_HTTP_header_lookup(hreq->headers, "Host");
-               path += hreq->path; 
-            }
-            std::cout << "Proxy request URL is " << path << std::endl;
-            std::string npath = 
-                search_replace(vars, path, req_uri_rx, req_uri_pat);
-            std::cout << "Resp request URL is " << npath << std::endl;
-            if (!npath.empty())
-                hreq->path = odr_strdup(o, npath.c_str());
-            std::cout << ">> Request headers" << std::endl;
-            //iterate headers
-            for (Z_HTTP_Header *header = hreq->headers;
-                    header != 0; 
-                    header = header->next) 
-            {
-                std::cout << header->name << ": " << header->value << std::endl;
-                std::string out = search_replace(vars, 
-                        std::string(header->value), 
-                        req_uri_rx, req_uri_pat);
-                if (!out.empty())
-                    header->value = odr_strdup(o, out.c_str());
-            }
-            package.request() = gdu;
-        }
-        package.move();
-        gdu = package.response().get();
-        if (gdu && gdu->which == Z_GDU_HTTP_Response)
-        {
-            std::cout << "<< Respose headers" << std::endl;
-            Z_HTTP_Response *hr = gdu->u.HTTP_Response;
-            mp::odr o;
-            //iterate headers
-            for (Z_HTTP_Header *header = hr->headers;
-                    header != 0; 
-                    header = header->next) 
-            {
-                std::cout << header->name << ": " << header->value << std::endl;
-                std::string out = search_replace(vars, 
-                        std::string(header->value), 
-                        resp_uri_rx, resp_uri_pat);
-                if (!out.empty())
-                    header->value = odr_strdup(o, out.c_str());
-            }
-            package.response() = gdu;
-        }
-    };
-
-    void configure(const xmlNode* ptr, bool test_only, const char *path) {};
-
-    const std::string search_replace(
-            std::map<std::string, std::string> & vars,
-            const std::string txt,
-            const std::string & uri_re,
-            const std::string & uri_pat) const
-    {
-        //exec regex against value
-        boost::regex re(uri_re);
-        boost::smatch what;
-        std::string::const_iterator start, end;
-        start = txt.begin();
-        end = txt.end();
-        std::string out;
-        while (regex_search(start, end, what, re)) //find next full match
-        {
-            unsigned i;
-            for (i = 1; i < what.size(); ++i)
-            {
-                //check if the group is named
-                std::map<int, std::string>::const_iterator it
-                    = groups_by_num.find(i);
-                if (it != groups_by_num.end()) 
-                {   //it is
-                    std::string name = it->second;
-                    vars[name] = what[i];
-                }
-
-            }
-            //prepare replacement string
-            std::string rvalue = sub_vars(uri_pat, vars);
-            //rewrite value
-            std::string rhvalue = what.prefix().str() 
-                + rvalue + what.suffix().str();
-            std::cout << "! Rewritten '"+what.str(0)+"' to '"+rvalue+"'\n";
-            out += rhvalue;
-            start = what[0].second; //move search forward
-        }
-        return out;
-    };
-
-    static void parse_groups(const std::string & str,
-            std::map<int, std::string> & groups_bynum,
-            std::map<std::string, int> & groups_byname)
+struct TestConfig {
+    TestConfig()   
     {
-       int gnum = 0;
-       bool esc = false;
-       for (int i = 0; i < str.size(); ++i)
-       {
-           if (!esc && str[i] == '\\')
-           {
-               esc = true;
-               continue;
-           }
-           if (!esc && str[i] == '(') //group starts
-           {
-               gnum++;
-               if (i+1 < str.size() && str[i+1] == '?') //group with attrs 
-               {
-                   i++;
-                   if (i+1 < str.size() && str[i+1] == ':') //non-capturing
-                   {
-                       if (gnum > 0) gnum--;
-                       i++;
-                       continue;
-                   }
-                   if (i+1 < str.size() && str[i+1] == 'P') //optional, python
-                       i++;
-                   if (i+1 < str.size() && str[i+1] == '<') //named
-                   {
-                       i++;
-                       std::string gname;
-                       bool term = false;
-                       while (++i < str.size())
-                       {
-                           if (str[i] == '>') { term = true; break; }
-                           if (!isalnum(str[i])) 
-                               throw mp::filter::FilterException
-                                   ("Only alphanumeric chars allowed, found "
-                                    " in '" 
-                                    + str 
-                                    + "' at " 
-                                    + boost::lexical_cast<std::string>(i)); 
-                           gname += str[i];
-                       }
-                       if (!term)
-                           throw mp::filter::FilterException
-                               ("Unterminated group name '" + gname 
-                                + " in '" + str +"'");
-                      groups_bynum[gnum] = gname;
-                      groups_byname[gname] = gnum;
-                      std::cout << "Found named group '" << gname 
-                          << "' at $" << gnum << std::endl;
-                   }
-               }
-           }
-           esc = false;
-       }
+        std::cout << "global setup\n"; 
+        yaz_log_init_level(YLOG_ALL);
     }
-
-    static std::string sub_vars (const std::string & in, 
-            const std::map<std::string, std::string> & vars)
-    {
-        std::string out;
-        bool esc = false;
-        for (int i = 0; i < in.size(); ++i)
-        {
-            if (!esc && in[i] == '\\')
-            {
-                esc = true;
-                continue;
-            }
-            if (!esc && in[i] == '$') //var
-            {
-                if (i+1 < in.size() && in[i+1] == '{') //ref prefix
-                {
-                    ++i;
-                    std::string name;
-                    bool term = false;
-                    while (++i < in.size()) 
-                    {
-                        if (in[i] == '}') { term = true; break; }
-                        name += in[i];
-                    }
-                    if (!term) throw mp::filter::FilterException
-                        ("Unterminated var ref in '"+in+"' at "
-                         + boost::lexical_cast<std::string>(i));
-                    std::map<std::string, std::string>::const_iterator it
-                        = vars.find(name);
-                    if (it != vars.end())
-                        out += it->second;
-                }
-                else
-                {
-                    throw mp::filter::FilterException
-                        ("Malformed or trimmed var ref in '"
-                         +in+"' at "+boost::lexical_cast<std::string>(i)); 
-                }
-                continue;
-            }
-            //passthru
-            out += in[i];
-            esc = false;
-        }
-        return out;
+    ~TestConfig() 
+    { 
+        std::cout << "global teardown\n"; 
     }
-    
-    void configure(
-            const std::string & req_uri_rx, 
-            const std::string & req_uri_pat, 
-            const std::string & resp_uri_rx, 
-            const std::string & resp_uri_pat) 
-    {
-       this->req_uri_rx = req_uri_rx;
-       this->req_uri_pat = req_uri_pat;
-       //pick up names
-       parse_groups(req_uri_rx, groups_by_num, groups_by_name);
-       this->resp_uri_rx = resp_uri_rx;
-       this->resp_uri_pat = resp_uri_pat;
-    };
-
-private:
-    std::map<std::string, std::string> vars;
-    std::string req_uri_rx;
-    std::string resp_uri_rx;
-    std::string req_uri_pat;
-    std::string resp_uri_pat;
-    std::map<int, std::string> groups_by_num;
-    std::map<std::string, int> groups_by_name;
-
 };
 
+BOOST_GLOBAL_FIXTURE( TestConfig );
 
 BOOST_AUTO_TEST_CASE( test_filter_rewrite_1 )
 {
     try
     {
-       FilterHeaderRewrite fhr;
+        mp::filter::HttpRewrite fhr;
     }
     catch ( ... ) {
         BOOST_CHECK (false);
@@ -295,16 +70,97 @@ BOOST_AUTO_TEST_CASE( test_filter_rewrite_2 )
     {
         mp::RouterChain router;
 
-        FilterHeaderRewrite fhr;
-        fhr.configure(
-                "(?:http\\:\\/\\/s?)?(?<host>[A-Za-z.]+):(?<port>\\d+)",
-                "http://${host}:${port}/somepath",
-                //rewrite connection close
-                "close",
-                "open for ${host}");
+        mp::filter::HttpRewrite fhr;
+        
+        mp::filter::HttpRewrite::spair_vec vec_req;
+        vec_req.push_back(std::make_pair(
+        "(?<proto>http\\:\\/\\/s?)(?<pxhost>[^\\/?#]+)\\/(?<pxpath>[^\\/]+)"
+        "\\/(?<host>[^\\/]+)(?<path>.*)",
+        "${proto}${host}${path}"
+        ));
+        vec_req.push_back(std::make_pair(
+        "(?:Host\\: )(.*)",
+        "Host: localhost"
+        ));
+
+        mp::filter::HttpRewrite::spair_vec vec_res;
+        vec_res.push_back(std::make_pair(
+        "(?<proto>http\\:\\/\\/s?)(?<host>[^\\/?#]+)\\/(?<path>[^ >]+)",
+        "THAT WAS MATCHED"
+        ));
+        
+        fhr.configure(vec_req, vec_res);
+
+        mp::filter::HTTPClient hc;
+        
+        router.append(fhr);
+        router.append(hc);
+
+        // create an http request
+        mp::Package pack;
+
+        mp::odr odr;
+        Z_GDU *gdu_req = z_get_HTTP_Request_uri(odr, 
+        "http://proxyhost/proxypath/localhost:80/~jakub/targetsite.php", 0, 1);
+
+        pack.request() = gdu_req;
+
+        //feed to the router
+        pack.router(router).move();
+
+        //analyze the response
+        Z_GDU *gdu_res = pack.response().get();
+        BOOST_CHECK(gdu_res);
+        BOOST_CHECK_EQUAL(gdu_res->which, Z_GDU_HTTP_Response);
+        
+        Z_HTTP_Response *hres = gdu_res->u.HTTP_Response;
+        BOOST_CHECK(hres);
+
+    }
+    catch (std::exception & e) {
+        std::cout << e.what();
+        std::cout << std::endl;
+        BOOST_CHECK (false);
+    }
+}
+
+BOOST_AUTO_TEST_CASE( test_filter_rewrite_3 )
+{
+    try
+    {
+        std::string xmlconf =
+            "<?xml version='1.0'?>\n"
+            "<filter xmlns='http://indexdata.com/metaproxy'\n"
+            "        id='rewrite1' type='http_rewrite'>\n"
+            " <request>\n"
+            "   <rewrite from='"
+    "(?&lt;proto>http\\:\\/\\/s?)(?&lt;pxhost>[^\\/?#]+)\\/(?&lt;pxpath>[^\\/]+)"
+    "\\/(?&lt;host>[^\\/]+)(?&lt;path>.*)'\n"
+            "            to='${proto}${host}${path}' />\n"
+            "   <rewrite from='(?:Host\\: )(.*)'\n"
+            "            to='Host: localhost' />\n" 
+            " </request>\n"
+            " <response>\n"
+            "   <rewrite from='"
+    "(?&lt;proto>http\\:\\/\\/s?)(?&lt;host>[^\\/?#]+)\\/(?&lt;path>[^ >]+)'\n"
+            "            to='THAT WAS MATCHED' />\n" 
+            " </response>\n"
+            "</filter>\n"
+            ;
+
+        std::cout << xmlconf;
+
+        // reading and parsing XML conf
+        xmlDocPtr doc = xmlParseMemory(xmlconf.c_str(), xmlconf.size());
+        BOOST_CHECK(doc);
+        xmlNode *root_element = xmlDocGetRootElement(doc);
+        mp::filter::HttpRewrite fhr; 
+        fhr.configure(root_element, true, "");
+        xmlFreeDoc(doc);
 
         mp::filter::HTTPClient hc;
         
+        mp::RouterChain router;
         router.append(fhr);
         router.append(hc);
 
@@ -313,7 +169,7 @@ BOOST_AUTO_TEST_CASE( test_filter_rewrite_2 )
 
         mp::odr odr;
         Z_GDU *gdu_req = z_get_HTTP_Request_uri(odr, 
-            "http://localhost:80/~jakub/targetsite.php", 0, 1);
+        "http://proxyhost/proxypath/localhost:80/~jakub/targetsite.php", 0, 1);
 
         pack.request() = gdu_req;
 
@@ -331,6 +187,7 @@ BOOST_AUTO_TEST_CASE( test_filter_rewrite_2 )
     }
     catch (std::exception & e) {
         std::cout << e.what();
+        std::cout << std::endl;
         BOOST_CHECK (false);
     }
 }