Mailing-List: contact commits-help@harmony.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@harmony.apache.org
Message-ID: <16008796.1170957186422.JavaMail.jira@brutus>
Date: Thu, 8 Feb 2007 09:53:06 -0800 (PST)
From: "Alexei Zakharov (JIRA)" <jira@apache.org>
To: commits@harmony.apache.org
Subject: [jira] Commented: (HARMONY-2467) [classlib][nio charset]
 java.nio.charset.CharsetDecoder(cs, Float.MIN_VALUE,Float.MIN_VALUE) does
 not throw IllegalArgumentException
In-Reply-To: <25232761.1165328182641.JavaMail.jira@brutus>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit


    [ https://issues.apache.org/jira/browse/HARMONY-2467?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12471422 ] 

Alexei Zakharov commented on HARMONY-2467:
------------------------------------------

I can't say I completely understand the RI's logic here. Why maxCharPerByte shouldn't be less than 1? If we take pure 16-bit unicode representation then maximum number of characters per byte should be 0.5, because it requires minimum two bytes to store any UTF-16 character (see [1] for example). Right? 

However, if you run the following test on RI
---
import java.nio.charset.*;

public class CharsetTest {

    public static void main(String argv[]) {

        String cnames[] = {"UTF-16", "UTF-8", "ISO-8859-1", "US-ASCII", "UTF-16BE"}; 
        for (String cname : cnames) {
            Charset cs = Charset.forName(cname);

            System.out.println(cname + ": " + cs.newDecoder().maxCharsPerByte());
        }
 
    }
}
---
You'll see:
UTF-16: 1.0
UTF-8: 2.0
ISO-8859-1: 1.0
US-ASCII: 1.0
UTF-16BE: 1.0

I miss something very important probably. How can we produce two chars with just one byte of input with UTF-8? Any comments?

[1] http://www.unicode.org/faq/utf_bom.html#37


> [classlib][nio charset] java.nio.charset.CharsetDecoder(cs, Float.MIN_VALUE,Float.MIN_VALUE) does not throw IllegalArgumentException
> ------------------------------------------------------------------------------------------------------------------------------------
>
>                 Key: HARMONY-2467
>                 URL: https://issues.apache.org/jira/browse/HARMONY-2467
>             Project: Harmony
>          Issue Type: Bug
>          Components: Classlib
>            Reporter: Anton Ivanov
>         Assigned To: Alexei Zakharov
>            Priority: Minor
>         Attachments: H-2467.patch
>
>
> According to the specification constructor must throw IllegalArgumentException if the preconditions on the parameters do not hold.
> java.nio.charset.CharsetDecoder(cs, Float.MIN_VALUE,Float.MIN_VALUE) does not throw IllegalArgumentException while RI does with message:
>    java.lang.IllegalArgumentException: Replacement too long
> Test to reproduce the problem:
> import junit.framework.TestCase;
> import java.nio.charset.*;
> import java.nio.*;
> public class test10036 extends TestCase {     
>     public void test1 () { 
>         try {                              
>             CharsetDecoderImpl obj = new CharsetDecoderImpl(Charset.forName("UTF-8"),
>                     Float.MIN_VALUE , Float.MIN_VALUE );   
>             fail("IllegalArgumentException should be thrown");                 
>         } catch (IllegalArgumentException e) { 
>             //expected
>         }        
>     }
> }
> class CharsetDecoderImpl extends CharsetDecoder{
>     public CharsetDecoderImpl(Charset cs, float averageBytesPerChar,
>             float maxBytesPerChar){
>         super(cs, averageBytesPerChar, maxBytesPerChar);
>     }
>     public CoderResult decodeLoop(ByteBuffer out, CharBuffer in) {
>                 return CoderResult.UNDERFLOW;
>     }
> }

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.