Mailing-List: contact issues-help@commons.apache.org; run by ezmlm
Precedence: bulk
Reply-To: issues@commons.apache.org
Message-ID: <12511223.123461295681144909.JavaMail.jira@thor>
Date: Sat, 22 Jan 2011 02:25:44 -0500 (EST)
From: "Julius Davies (JIRA)" <jira@apache.org>
To: issues@commons.apache.org
Subject: [jira] Updated: (CODEC-105) ArrayIndexOutOfBoundsException when
 doing multiple reads() on encoding b64InputStream
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit


     [ https://issues.apache.org/jira/browse/CODEC-105?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Julius Davies updated CODEC-105:
--------------------------------

    Attachment: codec-105.patch

This fixes the bug by removing the "little optimization".   The fix is actually 99% code removal, and it makes the overall base64 code a bit simpler, and should result in higher code-coverage percentages, since I think this "little optimization" was hard to cover.

Mea culpa:  the "little optimization" was my own foolish idea from the beginning (it's there in my original CODEC-69 patch), and I've felt funny about it for a long time.  Now that I see it's even causing a real bug I'm very much +1 for purging it!!!   Sorry!

> ArrayIndexOutOfBoundsException when doing multiple reads() on encoding b64InputStream
> -------------------------------------------------------------------------------------
>
>                 Key: CODEC-105
>                 URL: https://issues.apache.org/jira/browse/CODEC-105
>             Project: Commons Codec
>          Issue Type: Bug
>    Affects Versions: 1.4
>         Environment: 64-bit Linux
>            Reporter: ZAK Magnus
>            Assignee: Julius Davies
>         Attachments: codec-105-junit-including-ASL2-license.patch, codec-105.patch
>
>
> When encoding a sizable stream byte by byte (so, just calling Base64InputStream.read()), after 10920 successful read()s, this happens: 
> java.lang.ArrayIndexOutOfBoundsException: 2
>         at org.apache.commons.codec.binary.Base64.encode(Base64.java:502)
>         at org.apache.commons.codec.binary.Base64InputStream.read(Base64InputStream.java:157)
>         at org.apache.commons.codec.binary.Base64InputStream.read(Base64InputStream.java:109)
> Based on this, the necessary conditions seem to be that buffer = null and modulus = 2. Then, if a read() is done, a single-byte buffer is used, whose length is doubled by resizeBuffer(), but that still doesn't make it big enough to hold the 4 bytes written to it because modulus was just incremented to 0. 
> Here's some sample code:
> import org.apache.commons.codec.binary.Base64InputStream;
> public class TestReads {
>     public static void main(String[] args) {
>         Base64InputStream b64stream = new Base64InputStream(System.in, true, 0, null);
>         int n = 0;
>         try {
>             while (b64stream.read() != -1) n++;
>         } catch (Exception x) {
>             System.out.println(n);
>             x.printStackTrace();
>         }
>     }
> }

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