encapsulate potential expensive objects in TextSnippet to allow GC them asap

this reduces chance of OOMs at massive search & snippet-fetching

git-svn-id: https://svn.berlios.de/svnroot/repos/yacy/trunk@7865 6c8d7289-2bf4-0310-a012-ef5d649a1542

2 files changed, 40 insertions, 34 deletions

diff --git a/source/de/anomic/search/TextSnippet.java b/source/de/anomic/search/TextSnippet.java
index d68b7dcc1..8ee4164c1 100644
--- a/source/de/anomic/search/TextSnippet.java
+++ b/source/de/anomic/search/TextSnippet.java
@@ -165,39 +165,46 @@ public class TextSnippet implements Comparable<TextSnippet>, Comparator<TextSnip
             init(url.hash(), snippetLine, source, null);

             return;

         }

-

-        Document document = loadDocument(loader, comp, queryhashes, cacheStrategy, url, reindexing, source);

-        if (document == null) {

-            if (this.error == null) {

-            	init(url.hash(), null, ResultClass.ERROR_PARSER_FAILED, "parser error/failed"); // cannot be parsed

-            }

-            return;

-        }

-

-        /* ===========================================================================

-         * COMPUTE SNIPPET

-         * =========================================================================== */

-        // we have found a parseable non-empty file: use the lines

-

-        // compute snippet from text

-        final Collection<StringBuilder> sentences = document.getSentences(pre);

-        if (sentences == null) {

-            init(url.hash(), null, ResultClass.ERROR_PARSER_NO_LINES, "parser returned no sentences");

-            return;

-        }

-        final SnippetExtractor tsr;

+        

         String textline = null;

         HandleSet remainingHashes = queryhashes;

-        try {

-            tsr = new SnippetExtractor(sentences, queryhashes, snippetMaxLength);

-            textline = tsr.getSnippet();

-            remainingHashes =  tsr.getRemainingWords();

-        } catch (final UnsupportedOperationException e) {

-            init(url.hash(), null, ResultClass.ERROR_NO_MATCH, "no matching snippet found");

-            return;

-        }

-

-        // compute snippet from media

+        { //encapsulate potential expensive sentences

+	        final Collection<StringBuilder> sentences;

+	        { //encapsulate potential expensive document 

+		        final Document document = loadDocument(loader, comp, queryhashes, cacheStrategy, url, reindexing, source);

+		        if (document == null) {

+		            if (this.error == null) {

+		            	init(url.hash(), null, ResultClass.ERROR_PARSER_FAILED, "parser error/failed"); // cannot be parsed

+		            }

+		            return;

+		        }

+		

+		        /* ===========================================================================

+		         * COMPUTE SNIPPET

+		         * =========================================================================== */

+		        // we have found a parseable non-empty file: use the lines

+		

+		        // compute snippet from text

+		        sentences = document.getSentences(pre);

+		        document.close();

+	        } //encapsulate potential expensive document END

+	        

+	        if (sentences == null) {

+	            init(url.hash(), null, ResultClass.ERROR_PARSER_NO_LINES, "parser returned no sentences");

+	            return;

+	        }

+	        

+	        try {

+	        	final SnippetExtractor tsr = new SnippetExtractor(sentences, queryhashes, snippetMaxLength);

+	            textline = tsr.getSnippet();

+	            remainingHashes =  tsr.getRemainingWords();

+	        } catch (final UnsupportedOperationException e) {

+	            init(url.hash(), null, ResultClass.ERROR_NO_MATCH, "no matching snippet found");

+	            return;

+	        }

+        } //encapsulate potential expensive sentences END

+

+        // compute snippet from media - attention document closed above!

         //String audioline = computeMediaSnippet(document.getAudiolinks(), queryhashes);

         //String videoline = computeMediaSnippet(document.getVideolinks(), queryhashes);

         //String appline = computeMediaSnippet(document.getApplinks(), queryhashes);

@@ -220,7 +227,7 @@ public class TextSnippet implements Comparable<TextSnippet>, Comparator<TextSnip
         // finally store this snippet in our own cache

         snippetsCache.put(wordhashes, urls, snippetLine);

 

-        document.close();

+//        document.close();

         init(url.hash(), snippetLine, source, null);

     }

     

diff --git a/source/net/yacy/document/WordTokenizer.java b/source/net/yacy/document/WordTokenizer.java
index a482cb9ad..fb26699a2 100644
--- a/source/net/yacy/document/WordTokenizer.java
+++ b/source/net/yacy/document/WordTokenizer.java
@@ -172,8 +172,7 @@ public class WordTokenizer implements Enumeration<String> {
      */
     public static SortedMap<byte[], Integer> hashSentence(final String sentence, final WordCache meaningLib) {
         final SortedMap<byte[], Integer> map = new TreeMap<byte[], Integer>(Base64Order.enhancedCoder);
-        Enumeration<String> words = null;
-        words = new WordTokenizer(new ByteArrayInputStream(UTF8.getBytes(sentence)), meaningLib);
+        final Enumeration<String> words = new WordTokenizer(new ByteArrayInputStream(UTF8.getBytes(sentence)), meaningLib);
         int pos = 0;
         String word;
         byte[] hash;