added open-ended field name handling

src/internfile/internfile.cpp

@@ -1,5 +1,5 @@
 #ifndef lint
-static char rcsid[] = "@(#$Id: internfile.cpp,v 1.30 2007-05-23 08:29:04 dockes Exp $ (C) 2004 J.F.Dockes";
+static char rcsid[] = "@(#$Id: internfile.cpp,v 1.31 2007-06-19 08:36:24 dockes Exp $ (C) 2004 J.F.Dockes";
 #endif
 /*
  *   This program is free software; you can redistribute it and/or modify
@@ -270,12 +270,12 @@ static const string keyab("abstract");
 static const string keyau("author");
 static const string keycs("charset");
 static const string keyct("content");
+static const string keyds("description");
 static const string keyfn("filename");
 static const string keykw("keywords");
 static const string keymd("modificationdate");
 static const string keymt("mimetype");
 static const string keyoc("origcharset");
-static const string keysm("sample");
 static const string keytt("title");
 
 bool FileInterner::dijontorcl(Rcl::Doc& doc)
@@ -283,15 +283,24 @@ bool FileInterner::dijontorcl(Rcl::Doc& doc)
     Dijon::Filter *df = m_handlers.back();
     const std::map<std::string, std::string>& docdata = df->get_meta_data();
 
-    getKeyValue(docdata, keyau, doc.author);
-    getKeyValue(docdata, keyoc, doc.origcharset);
-    getKeyValue(docdata, keyct, doc.text);    
-    getKeyValue(docdata, keytt, doc.title);
-    getKeyValue(docdata, keykw, doc.keywords);
-    getKeyValue(docdata, keymd, doc.dmtime);
-    if (!getKeyValue(docdata, keyab, doc.abstract))
-	getKeyValue(docdata, keysm, doc.abstract);
-    LOGDEB1(("FILENAME: %s\n", doc.utf8fn.c_str()));
+    for (map<string,string>::const_iterator it = docdata.begin(); 
+	 it != docdata.end(); it++) {
+	if (it->first == keyct) {
+	    doc.text = it->second;
+	} else if (it->first == keymd) {
+	    doc.dmtime = it->second;
+	} else if (it->first == keyoc) {
+	    doc.origcharset = it->second;
+	} else if (it->first == keymt || it->first == keycs) {
+	    // don't need these.
+	} else {
+	    doc.meta[it->first] = it->second;
+	}
+    }
+    if (doc.meta[keyab].empty() && !doc.meta[keyds].empty()) {
+	doc.meta[keyab] = doc.meta[keyds];
+	doc.meta.erase(keyds);
+    }
     return true;
 }
 
@@ -324,7 +333,7 @@ void FileInterner::collectIpathAndMT(Rcl::Doc& doc, string& ipath) const
 	} else {
 	    ipath += isep;
 	}
-	getKeyValue(docdata, keyau, doc.author);
+	getKeyValue(docdata, keyau, doc.meta["author"]);
 	getKeyValue(docdata, keymd, doc.dmtime);
     }
 
@@ -672,7 +681,7 @@ int main(int argc, char **argv)
 	"]]]]\n-----------------------------------------------------\n" <<
 	"doc.keywords [[[[" << doc.keywords <<
 	"]]]]\n-----------------------------------------------------\n" <<
-	"doc.abstract [[[[" << doc.abstract <<
+	"doc.meta["abstract"] [[[[" << doc.meta["abstract"] <<
 	"]]]]\n-----------------------------------------------------\n" <<
 	"doc.text [[[[" << doc.text << "]]]]\n";
 }

src/internfile/mh_html.cpp

@@ -136,15 +136,16 @@ bool MimeHandlerHtml::next_document()
     m_metaData["origcharset"] = m_defcharset;
     m_metaData["content"] = result.dump;
     m_metaData["charset"] = "utf-8";
-    m_metaData["title"] = result.title;
-    m_metaData["keywords"] = result.keywords;
     // Avoid setting empty values which would crush ones possibly inherited
     // from parent (if we're an attachment)
-    if (!result.author.empty())
-	m_metaData["author"] = result.author;
     if (!result.dmtime.empty())
 	m_metaData["modificationdate"] = result.dmtime;
-    m_metaData["sample"] = result.sample;
     m_metaData["mimetype"] = "text/plain";
+
+    for (map<string,string>::const_iterator it = result.meta.begin(); 
+	 it != result.meta.end(); it++) {
+	if (!it->second.empty())
+	    m_metaData[it->first] = it->second;
+    }
     return true;
 }

src/internfile/myhtmlparse.cpp

@@ -144,22 +144,7 @@ MyHtmlParser::opening_tag(const string &tag, const map<string,string> &p)
 		    if ((j = p.find("name")) != p.end()) {
 			string name = j->second;
 			lowercase_term(name);
-			if (name == "description") {
-			    if (sample.empty()) {
-				sample = i->second;
-				decode_entities(sample);
-			    }
-			} else if (name == "keywords") {
-			    if (!keywords.empty()) keywords += ' ';
-			    string tmp = i->second;
-			    decode_entities(tmp);
-			    keywords += tmp;
-			} else if (name == "author") {
-			    if (!author.empty()) author += ' ';
-			    string tmp = i->second;
-			    decode_entities(tmp);
-			    author += tmp;
-			} else if (name == "date") {
+			if (name == "date") {
 			    // Yes this doesnt exist. It's output by filters
 			    // And the format isn't even standard http/html
 			    // FIXME
@@ -172,7 +157,14 @@ MyHtmlParser::opening_tag(const string &tag, const map<string,string> &p)
 				sprintf(ascuxtime, "%ld", (long)mktime(&tm));
 				dmtime = ascuxtime;
 			    }
-			} 
+			} else if (name == "robots") {
+			} else {
+			    if (!meta[name].empty())
+				meta[name] += ' ';
+			    string tmp = i->second;
+			    decode_entities(tmp);
+			    meta[name] += tmp;
+			}
 		    } else if ((j = p.find("http-equiv")) != p.end()) {
 			string hequiv = j->second;
 			lowercase_term(hequiv);
@@ -309,8 +301,8 @@ MyHtmlParser::closing_tag(const string &tag)
 	    break;
 	case 't':
 	    if (tag == "title") {
-		if (title.empty()) {
-		    title = dump;
+		if (meta["title"].empty()) {
+		    meta["title"] = dump;
 		    dump = "";
 		}
 		break;

src/internfile/myhtmlparse.h

@@ -22,6 +22,8 @@
  * USA
  * -----END-LICENCE-----
  */
+#include <map>
+using std::map;
 
 #include "htmlparse.h"
 
@@ -37,7 +39,8 @@ class MyHtmlParser : public HtmlParser {
     bool in_body_tag; 
     bool in_pre_tag;
     bool pending_space;
-    string title, sample, keywords, dump, dmtime, author;
+    map<string,string> meta;
+    string dump, dmtime;
     string ocharset; // This is the charset our user thinks the doc was
     string charset; // This is the charset it was supposedly converted to
     string doccharset; // Set this to value of charset parameter in header