[HADOOP-3454] Text.find incorrectly searches beyond the end of the buffer

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 0.17.0
• Fix Version/s: 0.18.0
• Hadoop Flags: Reviewed
• Assignee: Chad Whipkey
• Reporter: Chad Whipkey
• Created: 27/May/08 22:35
• Updated: 22/Aug/08 19:50
• Resolved: 03/Jun/08 07:04

Description

Text.find() does not pay attention to the length field. So, this code:

    public void testTextFind()
    {
        Text t = new Text("FIND AN I");
        t.set(new byte[] { (byte) 'F' });

        assert t.getLength() == 1 : "Length should be 1";
        assert t.find( "F") == 0 : "Found F at " + t.find("F");
        assert t.find( "I") == -1 : "Found I at " + t.find("I");
    }

incorrectly throws an assertion because it finds the I at position 1, even though the Text is only one byte long.

I think to fix this it is enough to change this line in Text.find:

ByteBuffer src = ByteBuffer.wrap(this.bytes);

to

ByteBuffer src = ByteBuffer.wrap(this.bytes,0,this.length);