[Bug libgcj/16134] New: Memory leak in String.getBytes()

[Bug libgcj/16134] New: Memory leak in String.getBytes()

[hannes at helma dot at]

Looks like there is a memory leak in java.lang.String.getBytes() in libgcj
3.3.3. The following code keeps growing continuously in memory:

//
// Test code for String.getBytes() memory leak
//     
    
public class StringBytesLeak {

    public static void main(String[] args) {
        while (true) {
            for (int i=0; i<10000; i++) {
                // leaks memory
                "xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx".getBytes();
                // using our own bogus implementation fixes the problem
                // getBytes("xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx");
            }
            try {
                Thread.sleep(500);
            } catch (InterruptedException x) {
                System.exit(0);
            }
        }
    }

    // simplistic implementation of String.getBytes() that doesn't leak
    static byte[] getBytes(String str) {
        int length = str.length();
        byte[] bytes = new byte[length];
        for (int i=0; i<length; i++) 
            bytes[i] = (byte) str.charAt(i);
        return bytes;
    }

}

-- 
           Summary: Memory leak in String.getBytes()
           Product: gcc
           Version: 3.3.3
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: libgcj
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: hannes at helma dot at
                CC: gcc-bugs at gcc dot gnu dot org,java-prs at gcc dot gnu
                    dot org
 GCC build triplet: i686-suse-linux
  GCC host triplet: i686-suse-linux
GCC target triplet: i686-suse-linux


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16134

[Bug libgcj/16134] Memory leak in String.getBytes()

[hannes at helma dot at]

------- Additional Comments From hannes at helma dot at  2004-06-22 13:01 -------
OK, I found what's going on. 

The problem is in gnu.gcj.convert.UnicodeToBytesgetEncoder(String encoding)

SuSE 9.1 uses UTF-8 as standard, but the file.encoding System property is set to
"UTF-8" rather than the canonical "UTF8". Calling
UnicodeToBytesgetEncoder(String encoding) with a non-canonical encoding name
causes the lookup in the encoderCache to fail. A new Encoder instance is created
each time and the encoderCache is swamped with instances (at least that's what I
guess is happening).

The fix should be easy - just canonicalize() the encoding name before looking it
up in the cache (it is done afterwards anyway).

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16134

[Bug libgcj/16134] Memory leak in String.getBytes()

[hannes at helma dot at]

------- Additional Comments From hannes at helma dot at  2004-06-22 13:28 -------
That should have been UnicodeToBytes.getEncoder(String encoding) in the previous
comment (copy/paste error)

Looks like the same problem is present in BytesToUnicode.getDecoder(String
encoding).

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16134

[Bug libgcj/16134] Memory leak in String.getBytes()

[mckinlay at redhat dot com]

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|                            |1
   Last reconfirmed|0000-00-00 00:00:00         |2004-06-22 16:31:41
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16134

[Bug libgcj/16134] Memory leak in String.getBytes()

[mckinlay at redhat dot com]

------- Additional Comments From mckinlay at redhat dot com  2004-06-22 16:56 -------
There is certainly a problem with the encoder cache here in that cache lookups
will always fail. I've written the obvious patch and will check it in shortly.
However the leak you see is not due to this code directly because it never keeps
more than 4 encoder instances in the cache. Rather, it is due to the
stack-walking code leaking memory as part of the calling classloader check in
Class.forName(). This was fixed in gcc 3.4, see bug 12475.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16134

[Bug libgcj/16134] Memory leak in String.getBytes()

[cvs-commit at gcc dot gnu dot org]

------- Additional Comments From cvs-commit at gcc dot gnu dot org  2004-06-22 17:21 -------
Subject: Bug 16134

CVSROOT:	/cvs/gcc
Module name:	gcc
Changes by:	bryce@gcc.gnu.org	2004-06-22 17:21:23

Modified files:
	libjava        : ChangeLog 
	libjava/gnu/gcj/convert: UnicodeToBytes.java BytesToUnicode.java 

Log message:
	PR libgcj/16134:
	* gnu/gcj/convert/BytesToUnicode.java: Canonicalize encoding name
	before cache lookup. Thanks to Hannes Wallnoefer.
	* gnu/gcj/convert/UnicodeToBytes.java: Likewise.

Patches:
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/libjava/ChangeLog.diff?cvsroot=gcc&r1=1.2875&r2=1.2876
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/libjava/gnu/gcj/convert/UnicodeToBytes.java.diff?cvsroot=gcc&r1=1.10&r2=1.11
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/libjava/gnu/gcj/convert/BytesToUnicode.java.diff?cvsroot=gcc&r1=1.9&r2=1.10



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16134

[Bug libgcj/16134] Memory leak in String.getBytes()

[mckinlay at redhat dot com]

------- Additional Comments From mckinlay at redhat dot com  2004-06-22 17:23 -------
Fix checked in.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16134

[Bug libgcj/16134] Memory leak in String.getBytes()

[pinskia at gcc dot gnu dot org]

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |3.5.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16134