cocoon-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jeremy Quinn <jer...@media.demon.co.uk>
Subject Re: Serious FlowScript problem
Date Thu, 24 Jun 2004 18:06:03 GMT

On 24 Jun 2004, at 14:21, Bertrand Delacretaz wrote:

> Le 23 juin 04, à 17:07, Jeremy Quinn a écrit :
>
>> ...in the first case it is loaded like this:
>> 	<map:flow language="javascript">
>> 		<map:script src="context://samples/blocks/lucene/query/query.js"/>
>> 	</map:flow>
>>
>> in the second case like this:
>> 	<map:flow language="javascript">
>> 		<map:script src="query.js"/>
>> 	</map:flow>
>>
>> I go back to the original url and it is still broken.
>>
>> What is going on ??????????
>
> Wild-guess mode on: could it be that the flowscript engine considers  
> these to be two different animals (due to the different way of  
> declaring the file) yet gets confused because they declare the same  
> stuff?

Thing is, it is random.
A FlowScript works one time and not another. This was only one scenario  
in which it happens. Normally my flowscripts are loaded as relative  
paths.

Gianugo very kindly spent some time with me with iChat and Subethaedit.

We looked at reducing the store size to 1 for the transient and store  
settings in cocoon.xconf, to see if caching might be a cause. The  
problem continued.

Then we discovered that different clients were getting different  
results.
I could access a URL that used FlowScript via 'localhost' and see the  
Function not found error.
Gianugo could access the same URL (externally) at the same time and see  
the page work.
I could run a different Browser.app and see the page work.
We could access a different URL, and maybe it would work on localhost  
but not remotely.

So we started thinking SESSION !!!!!

Does this trigger any synapses ?

As  
org.apache.cocoon.components.flow.javascript.fom.FOM_JavaScriptInterpret 
er.callFunction is the last method in the stacktrace I tried rolling my  
version of FOM_JavaScriptInterpreter back. The problem still existed  
all the way back (from 1.30) to 1.26, so I don't think this is the  
likely cause.
I am not sure in which specific commit this problem started occurring,  
we had been using a release version of 2.1.4 for a while, and had not  
kept up with 2.1.5-dev very closely.

> Maybe activating logs or debugging the flowscript engine to find out  
> how/when both versions of the file are compiled/interpreted (I don't  
> even exactly know how they are handled) would help?

Thanks mate.


regards Jeremy



--------------------------------------------------------

                   If email from this address is not signed
                                 IT IS NOT FROM ME

                         Always check the label, folks !!!!!
--------------------------------------------------------


Mime
View raw message