Apache
/
lucene
mirror of https://github.com/apache/lucene.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

fix for bug 27423 

slightly modified version of Daniel?s patch


git-svn-id: https://svn.apache.org/repos/asf/lucene/java/trunk@150297 13f79535-47bb-0310-9956-ffa450edef68
This commit is contained in:
Christoph Goller 2004-04-20 16:27:24 +00:00
parent d00487507e
commit 420d1d7c07
2 changed files with 38 additions and 20 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

35
src/demo/org/apache/lucene/demo/html/HTMLParser.java
View File

@ -10,7 +10,8 @@ public class HTMLParser implements HTMLParserConstants {
StringBuffer title = new StringBuffer(SUMMARY_LENGTH); StringBuffer title = new StringBuffer(SUMMARY_LENGTH);
StringBuffer summary = new StringBuffer(SUMMARY_LENGTH * 2); StringBuffer summary = new StringBuffer(SUMMARY_LENGTH * 2);
Properties metaTags=new Properties(); Properties metaTags=new Properties();
String currentMetaTag=""; String currentMetaTag=null;
String currentMetaContent=null;
int length = 0; int length = 0;
boolean titleComplete = false; boolean titleComplete = false;
boolean inTitle = false; boolean inTitle = false;
@ -120,11 +121,6 @@ InterruptedException {
void addText(String text) throws IOException { void addText(String text) throws IOException {
if (inStyle) if (inStyle)
return; return;
if (inMetaTag)
{
metaTags.setProperty(currentMetaTag, text);
return;
}
if (inTitle) if (inTitle)
title.append(text); title.append(text);
else { else {
@ -143,6 +139,13 @@ InterruptedException {
afterSpace = false; afterSpace = false;
} }
void addMetaTag() throws IOException {
metaTags.setProperty(currentMetaTag, currentMetaContent);
currentMetaTag = null;
currentMetaContent = null;
return;
}
void addSpace() throws IOException { void addSpace() throws IOException {
if (!afterSpace) { if (!afterSpace) {
if (inTitle) if (inTitle)
@ -262,11 +265,17 @@ InterruptedException {
&& t2 != null) && t2 != null)
{ {
currentMetaTag=t2.image.toLowerCase(); currentMetaTag=t2.image.toLowerCase();
if(currentMetaTag != null && currentMetaContent != null) {
addMetaTag();
}
} }
if(inMetaTag && t1.image.equalsIgnoreCase("content") && t2 != if(inMetaTag && t1.image.equalsIgnoreCase("content") && t2 !=
null) null)
{ {
addText(t2.image); currentMetaContent=t2.image.toLowerCase();
if(currentMetaTag != null && currentMetaContent != null) {
addMetaTag();
}
} }
break; break;
default: default:
@ -441,18 +450,18 @@ null)
finally { jj_save(1, xla); } finally { jj_save(1, xla); }
} }
final private boolean jj_3_2() {
if (jj_scan_token(ArgQuote2)) return true;
if (jj_scan_token(CloseQuote2)) return true;
return false;
}
final private boolean jj_3_1() { final private boolean jj_3_1() {
if (jj_scan_token(ArgQuote1)) return true; if (jj_scan_token(ArgQuote1)) return true;
if (jj_scan_token(CloseQuote1)) return true; if (jj_scan_token(CloseQuote1)) return true;
return false; return false;
} }
final private boolean jj_3_2() {
if (jj_scan_token(ArgQuote2)) return true;
if (jj_scan_token(CloseQuote2)) return true;
return false;
}
public HTMLParserTokenManager token_source; public HTMLParserTokenManager token_source;
SimpleCharStream jj_input_stream; SimpleCharStream jj_input_stream;
public Token token, jj_nt; public Token token, jj_nt;

23
src/demo/org/apache/lucene/demo/html/HTMLParser.jj
View File

@ -74,7 +74,8 @@ public class HTMLParser {
StringBuffer title = new StringBuffer(SUMMARY_LENGTH); StringBuffer title = new StringBuffer(SUMMARY_LENGTH);
StringBuffer summary = new StringBuffer(SUMMARY_LENGTH * 2); StringBuffer summary = new StringBuffer(SUMMARY_LENGTH * 2);
Properties metaTags=new Properties(); Properties metaTags=new Properties();
String currentMetaTag=""; String currentMetaTag=null;
String currentMetaContent=null;
int length = 0; int length = 0;
boolean titleComplete = false; boolean titleComplete = false;
boolean inTitle = false; boolean inTitle = false;
@ -184,11 +185,6 @@ InterruptedException {
void addText(String text) throws IOException { void addText(String text) throws IOException {
if (inStyle) if (inStyle)
return; return;
if (inMetaTag)
{
metaTags.setProperty(currentMetaTag, text);
return;
}
if (inTitle) if (inTitle)
title.append(text); title.append(text);
else { else {
@ -207,6 +203,13 @@ InterruptedException {
afterSpace = false; afterSpace = false;
} }
void addMetaTag() throws IOException {
metaTags.setProperty(currentMetaTag, currentMetaContent);
currentMetaTag = null;
currentMetaContent = null;
return;
}
void addSpace() throws IOException { void addSpace() throws IOException {
if (!afterSpace) { if (!afterSpace) {
if (inTitle) if (inTitle)
@ -284,11 +287,17 @@ void Tag() throws IOException :
&& t2 != null) && t2 != null)
{ {
currentMetaTag=t2.image.toLowerCase(); currentMetaTag=t2.image.toLowerCase();
if(currentMetaTag != null && currentMetaContent != null) {
addMetaTag();
}
} }
if(inMetaTag && t1.image.equalsIgnoreCase("content") && t2 != if(inMetaTag && t1.image.equalsIgnoreCase("content") && t2 !=
null) null)
{ {
addText(t2.image); currentMetaContent=t2.image.toLowerCase();
if(currentMetaTag != null && currentMetaContent != null) {
addMetaTag();
}
} }
} }
)? )?