From f897f087aa0f530d564e6f3f042af19d2fe943ab Mon Sep 17 00:00:00 2001
From: Jean-Francois Dockes <jfd@recoll.org>
Date: Sat, 12 Jan 2013 14:06:40 +0100
Subject: [PATCH] HTML: do not concatenate text found before body tag with the
 title. Fixes issue #125

---
 src/internfile/myhtmlparse.cpp | 26 ++++++++++++++++++++------
 src/internfile/myhtmlparse.h   |  3 ++-
 tests/html/html.txt            |  2 +-
 3 files changed, 23 insertions(+), 8 deletions(-)

diff --git a/src/internfile/myhtmlparse.cpp b/src/internfile/myhtmlparse.cpp
index b47fa260..d449f9c1 100644
--- a/src/internfile/myhtmlparse.cpp
+++ b/src/internfile/myhtmlparse.cpp
@@ -177,6 +177,7 @@ MyHtmlParser::MyHtmlParser()
     : in_script_tag(false),
       in_style_tag(false),
       in_pre_tag(false),
+      in_title_tag(false),
       pending_space(false),
       indexing_allowed(true)
 {
@@ -256,12 +257,20 @@ void MyHtmlParser::decode_entities(string &s)
 void
 MyHtmlParser::process_text(const string &text)
 {
-    LOGDEB2(("process_text: pending_space %d txt [%s]\n", pending_space,
-	    text.c_str()));
+    LOGDEB2(("process_text: title %d script %d style %d pre %d "
+	     "pending_space %d txt [%s]\n", 
+	     in_title_tag,
+	     in_script_tag,
+	     in_style_tag,
+	     in_pre_tag,
+	     pending_space,
+	     text.c_str()));
     CancelCheck::instance().checkCancel();
 
     if (!in_script_tag && !in_style_tag) {
-	if (!in_pre_tag) {
+	if (in_title_tag) {
+	    titledump += text;
+	} else if (!in_pre_tag) {
 	    string::size_type b = 0;
 	    bool only_space = true;
 	    while ((b = text.find_first_not_of(WHITESPACE, b)) != string::npos) {
@@ -461,7 +470,11 @@ MyHtmlParser::opening_tag(const string &tag)
 	    break;
 	case 't':
 	    if (tag == "table" || tag == "td" || tag == "textarea" ||
-		tag == "th") pending_space = true;
+		tag == "th") {
+		pending_space = true;
+	    } else if (tag == "title") {
+		in_title_tag = true;
+	    }
 	    break;
 	case 'u':
 	    if (tag == "ul") pending_space = true;
@@ -542,9 +555,10 @@ MyHtmlParser::closing_tag(const string &tag)
 	    break;
 	case 't':
 	    if (tag == "title") {
+		in_title_tag = false;
 		if (meta.find("title") == meta.end()|| meta["title"].empty()) {
-		    meta["title"] = dump;
-		    dump.clear();
+		    meta["title"] = titledump;
+		    titledump.clear();
 		}
 		break;
 	    }
diff --git a/src/internfile/myhtmlparse.h b/src/internfile/myhtmlparse.h
index 38b3e8f6..eba0785e 100644
--- a/src/internfile/myhtmlparse.h
+++ b/src/internfile/myhtmlparse.h
@@ -37,9 +37,10 @@ class MyHtmlParser : public HtmlParser {
     bool in_script_tag;
     bool in_style_tag;
     bool in_pre_tag;
+    bool in_title_tag;
     bool pending_space;
     map<string,string> meta;
-    string dump, dmtime;
+    string dump, dmtime, titledump;
     // This is the charset our caller thinks the doc used (initially
     // comes from the environment/configuration, used as source for
     // conversion to utf-8)
diff --git a/tests/html/html.txt b/tests/html/html.txt
index cdf4c9f4..e0b8e113 100644
--- a/tests/html/html.txt
+++ b/tests/html/html.txt
@@ -6,7 +6,7 @@ text/html	[file:///home/dockes/projets/fulltext/testrecoll/html/qtextedit.html]
 text/html	[file:///home/dockes/projets/fulltext/testrecoll/html/mysqlmanualSmall.html]	[MySQL 3.23, 4.0, 4.1 Reference Manual]	1480354	bytes	
 text/html	[file:///home/dockes/projets/fulltext/testrecoll/html/mysqlmanual.html]	[MySQL 3.23, 4.0, 4.1 Reference Manual]	8136414	bytes	
 1 results
-text/html	[file:///home/dockes/projets/fulltext/testrecoll/html/badhtml.html]	[bla bla "bad quotes]	236	bytes	
+text/html	[file:///home/dockes/projets/fulltext/testrecoll/html/badhtml.html]	["bad quotes]	236	bytes	
 1 results
 text/html	[file:///home/dockes/projets/fulltext/testrecoll/html/htmlfield.html]	[htmlfield.html]	137	bytes	
 testfield = testfieldvalue