Doug -

That actually a good idea.. Still interested in knowning why the error code
not going up though...

On Sun, Aug 23, 2009 at 9:21 AM, Douglas Knudsen
<[email protected]>wrote:

> How are you exposing the REST call?  A Servlet?  Throwing a error in this
> case should trickle up to a HTTP code in the headers somehow, right?  The
> AMF goodness handles this, encapsulates it, for you thus a error thrown in a
> RemoteObject call trickles up and all.  But HTTPService may need more there
> by returning a particular HTTP status cod.  Another approach is to return a
> standard HTTP response with erro info in the body.  Your resultHamdler
> parses the response and decides to continue or call the faultHandler() if
> need be.  I've had to do this with WebServices before with a client.
>
> HTH!
> Douglas Knudsen
> http://www.cubicleman.com
> this is my signature, like it?
>
>
>
> On Fri, Aug 21, 2009 at 5:01 PM, Dale Bronk <[email protected]> wrote:
>
>>  I had thought about it, but I really don’t like the idea of managing
>> session timeout on the client and the server.  That is a server job and no
>> matter how careful you are, it is bound to get out of sync.  Also, our app
>> will change the session timeout depending on the time of day and that would
>> mean then having to manage that on the client as well.  I’m really very
>> shocked that my exception is getting eaten somewhere and not being brought
>> back to me.  This tells me that if you do HTTPService then fault handling is
>> useless as you will never know what happened.  Right now I’m going down the
>> path of creating a servlet filter that will hopefully be able to change the
>> http status code and use that.  I’ll let you know if it works.  Of course it
>> is a java solution.
>>
>>
>>
>> *From:* [email protected] [mailto:[email protected]] *On Behalf Of *eric
>> decoff
>> *Sent:* Friday, August 21, 2009 4:36 PM
>>
>> *To:* [email protected]
>> *Subject:* Re: [AFFUG Discuss] Session timeout
>>
>>
>>
>> Hey Dale,
>>
>>
>>
>>      You might want to consider using a timer
>>
>>
>>
>>      I put a sample of what I am talking about on my new blog
>>
>>
>>
>>      http://ericdecoff.blogspot.com
>>
>>
>>
>>
>>
>>
>>
>>
>>
>> On Fri, Aug 21, 2009 at 1:26 PM, Dale Bronk <[email protected]> wrote:
>>
>> I have looked via debugger and basically expanded every piece of data.  I
>> will look at the codes closer though.  I gave them a look the first time,
>> but didn’t seem to be any different from any other fault.  But a second look
>> certainly is not going to hurt.
>>
>>
>>
>> *From:* [email protected] [mailto:[email protected]] *On Behalf Of *Marty
>> Blood
>> *Sent:* Friday, August 21, 2009 1:15 PM
>>
>>
>> *To:* [email protected]
>> *Subject:* Re: [AFFUG Discuss] Session timeout
>>
>>
>>
>> Have you tried running in debug and inspecting the FaultEvent manually?
>> Since it's a HTTPService call it might return a FaultEvent with a
>> HTTPResponse code (instead of a 200 maybe something more meaningful) or some
>> such that's populated differently than the AMF call.
>>
>> Looking at the FaultEvent API you might look at the statusCode field or
>> the content Object in the FaultEvent's Fault object to see if it has the
>> HTTPResponse data you need to see that it was a session timeout (or if it
>> was an exception the 500 or whatever server error response is coming back
>> down the pipe).  There's also the headers Object on the FaultEvent and the
>> ErrorMessage which might contain the data you need.
>>
>> If you're feeling really fancy you can use WebScarab or some such to
>> intercept the HTTPService response and inspect the HTTP headers manually to
>> see if those give you any more insight.
>>
>> Anyway, good luck with that.
>>
>> -Marty
>>
>> On Fri, Aug 21, 2009 at 12:53 PM, Dale Bronk <[email protected]>
>> wrote:
>>
>> Right, I'm not trying to catch it on the client.  The server is catching
>> it
>> and throwing an exception as per our design.  That causes a FaultEvent,
>> not
>> a ResultEvent.  When we make AMF calls this is handled perfectly.  Server
>> side throws an exception.  Our AMF fault handler interrogates the
>> FaultEvent
>> and sees that it is a session time out and handles it.  But, when we make
>> the HTTPService call the server still does what it is suppose to and
>> throws
>> an exception and that still causes a FaultEvent, but in our fault handler
>> for the HTTPService call I do not see anywhere in the FaultEvent anything
>> that tells me what the issue was.
>>
>> Our code for handling the AMF faults is:
>>
>>        // Some non important to this issue code
>>
>>        var errorMessage : ErrorMessage = ErrorMessage(faultEvent.message);
>>        var rootCause : Object = errorMessage.rootCause;
>>        if ( rootCause != null )
>>        {
>>                // We have our own well known exception with properties
>> that
>> tell us what went wrong.
>>                var errorType : String = ( rootCause.hasOwnProperty("type")
>> ? String(rootCause.type) : null );
>>                if ( errorType == "INVALID_OR_EXPIRED_CREDENTIALS" )
>>                {
>>                        // Code to publish a session time out to the
>> application
>>                }
>>
>>                // Other non important to this issue code
>>        }
>>
>> The problem is that even though we throw the same exception, when it is a
>> HTTPService call this information is not in the FaultEvent.  I see no
>> information from our backend processing at all anywhere in the FaultEvent.
>>
>> Thanks,
>> Dale
>>
>>
>>
>> -----Original Message-----
>> From: [email protected] [mailto:[email protected]] On Behalf Of Cameron
>> Childress
>> Sent: Friday, August 21, 2009 12:11 PM
>> To: [email protected]
>> Subject: Re: [AFFUG Discuss] Session timeout
>>
>> The session timeout occurs because the client hasn't connected in a
>> long time.  This occurs on the server, while not connected to the
>> client.  So....  I may be wrong but I don't think you are ever going
>> to get a client side Flex Event since this all occurs because the
>> client hasn't been around in awhile.
>>
>> You can always catch it on the server and then have the Client check
>> for a flag when it reconnects to see when the server session timed
>> out.
>>
>> Someone please correct me if I am wrong.
>>
>> -Cameron
>>
>> On Fri, Aug 21, 2009 at 10:40 AM, Dale Bronk<[email protected]> wrote:
>> > Anyone have a good pattern for detecting a session timeout for an
>> > HTTPService call?  We use Java on the backend and while we through a
>> good
>> > exception on the backend, it doesn’t seem to make in the FaultEvent.
>> I’ve
>> > looked everywhere in the FaultEvent and just can’t find any mention of
>> the
>> > true exception that was thrown.  We have one small portion of the app
>> that
>> > is HTTPService calls and the rest is AMF using GraniteDS.  We can handle
>> the
>> > AMF calls no problem as we can access the Java Exception that was thrown
>> but
>> > I’m stumped for the HTTPService calls.
>>
>>
>> -------------------------------------------------------------
>> To unsubscribe from this list, simply email the list with unsubscribe in
>> the
>> subject line
>>
>> For more info, see http://www.affug.com
>> Archive @ http://www.mail-archive.com/discussion%40affug.com/
>> List hosted by http://www.fusionlink.com
>> -------------------------------------------------------------
>>
>>
>>
>>
>>
>> -------------------------------------------------------------
>> To unsubscribe from this list, simply email the list with unsubscribe in
>> the subject line
>>
>> For more info, see http://www.affug.com
>> Archive @ http://www.mail-archive.com/discussion%40affug.com/
>> List hosted by http://www.fusionlink.com
>> -------------------------------------------------------------
>>
>>
>>
>>
>> -------------------------------------------------------------
>>
>> To unsubscribe from this list, simply email the list with unsubscribe in
>> the subject line
>>
>> For more info, see http://www.affug.com
>> Archive @ http://www.mail-archive.com/discussion%40affug.com/
>>
>> List hosted by FusionLink <http://www.fusionlink.com/>
>> -------------------------------------------------------------
>>
>>
>>
>> -------------------------------------------------------------
>> To unsubscribe from this list, simply email the list with unsubscribe in
>> the subject line
>>
>> For more info, see http://www.affug.com
>> Archive @ http://www.mail-archive.com/discussion%40affug.com/
>> List hosted by FusionLink <http://www.fusionlink.com/>
>> -------------------------------------------------------------
>>
>
>

Reply via email to