"502 Bad Gateway" when rendering on an ajax callback

Hi guys,

I am not sure if the problem I have now started with Seaside 3.1.4.1 or not... but I think it does because I don't remember seeing this before. The thing is... when I have an error in a callback of a normal request, then I guess the normal walkback screen with Remote Debug, Full Stack, etc. If I have an error at the RENDERING phase of seaside, then I get the error dialog embedded in the screen in the component that throw the error. 

My problem is now with ajax. It seems that if I have an error at the ajax callabck, then I get an "502 Bad Gateway". Even worst, if the error was in seaside rendering phase (called from the #script:  via a #replaceWith: or similar), then the whole screen is with the nginx "502 Bad Gateway".  Here I would expect the same as the normal request...that is, the component that go an error get rendered with the error description.

The gems do not go down and they seem to be alive. But why I have above behavior? 

Any idea? I don't know even were to start to debug the problem....

The error is logged in the stack AND in the object log. Just the screen is not rendering accordly.

thanks in advance for any tip.

--

Mariano
http://marianopeck.wordpress.com

Just in case it was not clear what I meant...here is attached an example of when I properly get the error that happened in a rendering phase but with a normal request (not ajax).  A rendering problem (via #replaceWith: for example from an ajax #script:) yields a "502 Bad Gateway" from nginx...

Maybe the problem could be around WARemoteDebuggingWalkbackErrorHandler >> open: 

   self requestContext request isXmlHttpRequest

            ifTrue: [ 

              wb logContinuation: anException description.

              (self requestContext responseGenerator internalError: anException)

                respond ]

       

?

Thanks in advance,

On Fri, Jun 12, 2015 at 8:50 PM, Mariano Martinez Peck <[hidden email]> wrote:

Hi guys,

I am not sure if the problem I have now started with Seaside 3.1.4.1 or not... but I think it does because I don't remember seeing this before. The thing is... when I have an error in a callback of a normal request, then I guess the normal walkback screen with Remote Debug, Full Stack, etc. If I have an error at the RENDERING phase of seaside, then I get the error dialog embedded in the screen in the component that throw the error. 

My problem is now with ajax. It seems that if I have an error at the ajax callabck, then I get an "502 Bad Gateway". Even worst, if the error was in seaside rendering phase (called from the #script:  via a #replaceWith: or similar), then the whole screen is with the nginx "502 Bad Gateway".  Here I would expect the same as the normal request...that is, the component that go an error get rendered with the error description.

The gems do not go down and they seem to be alive. But why I have above behavior? 

Any idea? I don't know even were to start to debug the problem....

The error is logged in the stack AND in the object log. Just the screen is not rendering accordly.

thanks in advance for any tip.

--

Mariano
http://marianopeck.wordpress.com

--

Mariano
http://marianopeck.wordpress.com

mmmm that could is surrounded with an error handler:

 do: [ :err | 

          Transcript

            cr;

            show: 'Error creating WAWalkback: ' , err description. "error depends upon whether we're using #call:onAnswer: or #call:"

       

I do a 

 grep -r "Error creating WAWalkback" *.log 

and I get plenty of entries like this one:

FastCGI_server-9013.log:--transcript--'Error creating WAWalkback: WAGsInvalidCallbackContext: You can only #call: and #answer: from within a callback or a Task.'

wtf....

 

On Fri, Jun 12, 2015 at 10:31 PM, Mariano Martinez Peck <[hidden email]> wrote:

Just in case it was not clear what I meant...here is attached an example of when I properly get the error that happened in a rendering phase but with a normal request (not ajax).  A rendering problem (via #replaceWith: for example from an ajax #script:) yields a "502 Bad Gateway" from nginx...

Maybe the problem could be around WARemoteDebuggingWalkbackErrorHandler >> open: 

   self requestContext request isXmlHttpRequest

            ifTrue: [ 

              wb logContinuation: anException description.

              (self requestContext responseGenerator internalError: anException)

                respond ]

       

?

Thanks in advance,

On Fri, Jun 12, 2015 at 8:50 PM, Mariano Martinez Peck <[hidden email]> wrote:

Hi guys,

I am not sure if the problem I have now started with Seaside 3.1.4.1 or not... but I think it does because I don't remember seeing this before. The thing is... when I have an error in a callback of a normal request, then I guess the normal walkback screen with Remote Debug, Full Stack, etc. If I have an error at the RENDERING phase of seaside, then I get the error dialog embedded in the screen in the component that throw the error. 

My problem is now with ajax. It seems that if I have an error at the ajax callabck, then I get an "502 Bad Gateway". Even worst, if the error was in seaside rendering phase (called from the #script:  via a #replaceWith: or similar), then the whole screen is with the nginx "502 Bad Gateway".  Here I would expect the same as the normal request...that is, the component that go an error get rendered with the error description.

The gems do not go down and they seem to be alive. But why I have above behavior? 

Any idea? I don't know even were to start to debug the problem....

The error is logged in the stack AND in the object log. Just the screen is not rendering accordly.

thanks in advance for any tip.

--

Mariano
http://marianopeck.wordpress.com

--

Mariano
http://marianopeck.wordpress.com

--

Mariano
http://marianopeck.wordpress.com