avoid File.deleteOnExit() on temp files

JVM registers each file in a list regardless of already deleted and never
cleans up the list during runtime.
This accumulates to a considerable amount of mem during large crawls and/or
long uptime.
To tackle this, all temp files are now created in a subdir of java.io.tmpdir 
and the jvm tmpdir property is set to this subdir, which is deleted by
code on shutdown.
Additionally let pdfParser use this tmp subdir too.

source/net/yacy/document/parser/apkParser.java

@@ -74,7 +74,6 @@ public class apkParser extends AbstractParser implements Parser  {
         Document[] docs = null;
         try {
             File tempFile = File.createTempFile("apk" + System.currentTimeMillis(), "jar");
-            tempFile.deleteOnExit();
             final FileOutputStream out = new FileOutputStream(tempFile);
             int read = 0;
             final byte[] data = new byte[1024];

source/net/yacy/document/parser/audioTagParser.java

@@ -109,8 +109,7 @@ public class audioTagParser extends AbstractParser implements Parser {
         		f = AudioFileIO.read(location.getFSFile());
         	} else {
             	// create a temporary file, as jaudiotagger requires a file rather than an input stream 
-            	tempFile = File.createTempFile(filename,fileext);  
+            	tempFile = File.createTempFile(filename,fileext);              
-                tempFile.deleteOnExit();              
                 fout = new BufferedOutputStream(new FileOutputStream(tempFile));  
                 int c;  
                 while ((c = source.read()) != -1) {  

source/net/yacy/document/parser/bzipParser.java

@@ -82,7 +82,6 @@ public class bzipParser extends AbstractParser implements Parser {
             final BZip2CompressorInputStream zippedContent = new BZip2CompressorInputStream(source);
 
             tempFile = File.createTempFile("bunzip","tmp");
-            tempFile.deleteOnExit();
 
             // creating a temp file to store the uncompressed data
             final FileOutputStream out = new FileOutputStream(tempFile);

source/net/yacy/document/parser/gzipParser.java

@@ -80,7 +80,6 @@ public class gzipParser extends AbstractParser implements Parser {
             final GZIPInputStream zippedContent = new GZIPInputStream(source);
 
             tempFile = File.createTempFile("gunzip","tmp");
-            tempFile.deleteOnExit();
 
             // creating a temp file to store the uncompressed data
             final FileOutputStream out = new FileOutputStream(tempFile);

source/net/yacy/document/parser/odtParser.java

@@ -229,7 +229,6 @@ public class odtParser extends AbstractParser implements Parser {
         try {
             // creating a tempfile
             dest = File.createTempFile("OpenDocument", ".odt");
-            dest.deleteOnExit();
 
             // copying the stream into a file
             FileUtils.copy(source, dest);

source/net/yacy/document/parser/ooxmlParser.java

@@ -215,7 +215,6 @@ public class ooxmlParser extends AbstractParser implements Parser {
         try {
             // creating a tempfile
             dest = File.createTempFile("OpenDocument", ".odt");
-            dest.deleteOnExit();
 
             // copying the stream into a file
             FileUtils.copy(source, dest);

source/net/yacy/document/parser/pdfParser.java

@@ -63,6 +63,7 @@ import net.yacy.document.VocabularyScraper;
 import net.yacy.kelondro.io.CharBuffer;
 import net.yacy.kelondro.util.FileUtils;
 import net.yacy.kelondro.util.MemoryControl;
+import org.apache.pdfbox.pdfparser.PDFParser;
 
 
 public class pdfParser extends AbstractParser implements Parser {
@@ -100,13 +101,13 @@ public class pdfParser extends AbstractParser implements Parser {
 
         // create a pdf parser
         PDDocument pdfDoc;
-        //final PDFParser pdfParser;
         try {
             Thread.currentThread().setPriority(Thread.MIN_PRIORITY); // the pdfparser is a big pain
-            pdfDoc = PDDocument.load(source);
+            //pdfDoc = PDDocument.load(source);
-            //PDFParser pdfParser = new PDFParser(source);
+            final PDFParser pdfParser = new PDFParser(source);
-            //pdfParser.parse();
+            pdfParser.setTempDirectory(new File(System.getProperty("java.io.tmpdir")));
-            //pdfDoc = pdfParser.getPDDocument();
+            pdfParser.parse();
+            pdfDoc = pdfParser.getPDDocument();
         } catch (final IOException e) {
             throw new Parser.Failure(e.getMessage(), location);
         } finally {

source/net/yacy/kelondro/util/FileUtils.java

@@ -819,7 +819,6 @@ public final class FileUtils {
             File.createTempFile(
                 parserClassName + "_" + ((idx > -1) ? fileName.substring(0, idx) : fileName),
                 (!fileExt.isEmpty()) ? "." + fileExt : fileExt);
-        tempFile.deleteOnExit();
         return tempFile;
     }
     

source/net/yacy/peers/SeedDB.java

@@ -832,7 +832,6 @@ public final class SeedDB implements AlternativeDomainNames {
         try {
             // create a seed file which for uploading ...
             seedFile = File.createTempFile("seedFile",".txt", seedDB.myOwnSeedFile.getParentFile());
-            seedFile.deleteOnExit();
             if (Network.log.isFine()) Network.log.fine("SaveSeedList: Storing seedlist into tempfile " + seedFile.toString());
             final ArrayList<String> uv = storeSeedList(seedFile, true);
 

source/net/yacy/yacy.java

@@ -128,6 +128,7 @@ public final class yacy {
     * @param startupFree free memory at startup time, to be used later for statistics
     */
     private static void startup(final File dataHome, final File appHome, final long startupMemFree, final long startupMemTotal, final boolean gui) {
+        String tmpdir=null;
         try {
             // start up
             System.out.println(copyright);
@@ -202,6 +203,12 @@ public final class yacy {
             	lock = channel.tryLock(); // lock yacy.running
             } catch (final Exception e) { }
 
+            // set jvm tmpdir to a subdir for easy cleanup (as extensive use file.deleteonexit waists memory during long runs, as todelete files names are collected and never cleaned up during runtime)
+            try { 
+                tmpdir = java.nio.file.Files.createTempDirectory("yacy-tmp-").toString(); // creates sub dir in jvm's temp (see System.property "java.io.tempdir")
+                System.setProperty("java.io.tmpdir", tmpdir);
+            } catch (IOException ex) { }
+
             try {
                 sb = new Switchboard(dataHome, appHome, "defaults/yacy.init".replace("/", File.separator), conf);
             } catch (final RuntimeException e) {
@@ -386,6 +393,8 @@ public final class yacy {
         } finally {
         }
 
+        if (tmpdir != null) FileUtils.deletedelete(new File(tmpdir)); // clean up temp dir (set at startup as subdir of system.propery "java.io.tmpdir")
+
         ConcurrentLog.config("SHUTDOWN", "goodbye. (this is the last line)");
         ConcurrentLog.shutdown();
         shutdownSemaphore.release(1000);
@@ -618,7 +627,7 @@ public final class yacy {
 	        if (OS.isWindows) headless = false;
 	        if (args.length >= 1 && args[0].toLowerCase().equals("-gui")) headless = false;
 	        System.setProperty("java.awt.headless", headless ? "true" : "false");
-            
+
 	        String s = ""; for (final String a: args) s += a + " ";
 	        yacyRelease.startParameter = s.trim();