Mailing-List: contact abdera-user-help@incubator.apache.org; run by ezmlm
Precedence: bulk
Reply-To: abdera-user@incubator.apache.org
Received-SPF: pass (athena.apache.org: domain of jasnell@gmail.com designates
 209.85.198.191 as permitted sender)
DomainKey-Signature: a=rsa-sha1; c=nofws;
        d=gmail.com; s=gamma;
        h=received:message-id:date:from:user-agent:mime-version:to:subject:references:in-reply-to:x-enigmail-version:content-type:content-transfer-encoding;
        b=rzsjoLTFIXvjsezwrNuVGz4BMqTlJhznRa4o/IkPqpw64OyqXdw4YAgrR73Jc64/IIXJqKjHT84UloOABNEI2TNjdqZ3n+vFblH26pdvB3cTd0ua6qYST6mgfjEpPCqhycJX4La1qPAXjUinIqLjP8QvJlX9dfFZHOXxe2QG+mY=
Message-ID: <475348DF.3070805@gmail.com>
Date: Sun, 02 Dec 2007 16:07:59 -0800
From: James M Snell <jasnell@gmail.com>
User-Agent: Thunderbird 2.0.0.6 (X11/20071022)
MIME-Version: 1.0
To: abdera-user@incubator.apache.org
Subject: Re: AtomDate bug
References: <30DFA387-25CC-439F-83C4-5D7AB27000C3@gmail.com>
In-Reply-To: <30DFA387-25CC-439F-83C4-5D7AB27000C3@gmail.com>
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit

In the trunk, AtomDate has been changed to use a regex pattern to parse
dates in order to improve performance.  If you would, please try your
tests with the trunk and let me know what problems you're seeing.  We
can tweak the regex accordingly.

- James

Chris Berry wrote:
> Greetings,
> I have found a subtle but dangerous bug in AtomDate (v0.3.0)
> AtomDate walks thru a series of masks, trying each, until it either
> finds a match, or, if not, throws an Exception.
> The problem is that, when a date string with a bad "time section" is
> used, SimpleDateFormat will go on to match the "date section"
> And thus, give the user something entirely different from what they
> wanted, rather than throw an Exception
> 
> Here is an example:
> 
>         String ddd = "2007-11-30T23:59:59:000Z";
>         Date upmin = AtomDate.parse( ddd );
>         log.debug( "********************* ddd= " + ddd );
>         log.debug( "********************* upmin= " + upmin );
>         log.debug( "********************* upmin GMT= " +
> upmin.toGMTString() );
> 
> NOTE: The correct string should be "2007-11-30T23:59:59.000Z";  <=== 
> 59dot000 (59.000) instead of 59colon000 (59:000) -- a pretty easy
> mistake, and hard to see
> 
> AtomDate then sees this as::
> 
> AtomDate: parsing:: 2007-11-30T23:59:59:000Z
> trying mask= yyyy-MM-dd'T'HH:mm:ss.SSSz
> trying mask= yyyy-MM-dd't'HH:mm:ss.SSSz
> trying mask= yyyy-MM-dd'T'HH:mm:ss.SSS'Z'
> trying mask= yyyy-MM-dd't'HH:mm:ss.SSS'z'
> trying mask= yyyy-MM-dd'T'HH:mm:ssz
> trying mask= yyyy-MM-dd't'HH:mm:ssz
> trying mask= yyyy-MM-dd'T'HH:mm:ss'Z'
> trying mask= yyyy-MM-dd't'HH:mm:ss'z'
> trying mask= yyyy-MM-dd'T'HH:mmz
> trying mask= yyyy-MM-dd't'HH:mmz
> trying mask= yyyy-MM-dd'T'HH:mm'Z'
> trying mask= yyyy-MM-dd't'HH:mm'z'
> trying mask= yyyy-MM-dd
> [12/02/07 17:11:48:683] DEBUG - EntryURIHelperTest         -
> ********************* ddd= 2007-11-30T23:59:59:000Z
> [12/02/07 17:11:48:684] DEBUG - EntryURIHelperTest         -
> ********************* upmin= Thu Nov 29 18:00:00 CST 2007
> [12/02/07 17:11:48:685] DEBUG - EntryURIHelperTest         -
> ********************* upmin GMT= 30 Nov 2007 00:00:00 GMT
> 
> NOTE: AtomDate has matched "yyyy-MM-dd"
> And thus, changed the value from   "30 Nov 2007 23:59:59 GMT"  to  "30
> Nov 2007 00:00:00 GMT" -- 24 hrs earlier !!!
> 
> I'm not too sure how this should be fixed?? (Using setLenient(false)
> will NOT fix it)
> I think just checking String length before switching to "date only" will
> result in the correct behavior.
> 
> Something like this works (but isn't too pretty ;-)
> 
>    private static final int checkLengthStartingIndex = 12;
> 
>   /**
>    * Parse the serialized string form into a java.util.Date
>    * @param date The serialized string form of the date
>    * @return The created java.util.Date
>    */
>   public static Date parse(String date) {
> 
>     .....
> 
>     Date d = null;
>     SimpleDateFormat sdf = new SimpleDateFormat();
>     for (int n = 0; n < masks.length; n++) {
>       try {
>         sdf.applyPattern(masks[n]);
>         sdf.setTimeZone(TimeZone.getTimeZone("GMT"));
> 
>         //sdf.setLenient(true);
>         // cwb >>>
>         sdf.setLenient(false);
>         if ( n >= checkLengthStartingIndex ) {
>             if ( date.length() != masks[n].length() )
>                 continue;
>         }
>         // cwb <<<
> 
> Cheers,
> -- Chris
> S'all good  ---   chriswberry at gmail dot com
> 
> 
> 
>