Re: Fetch Specifications defined in the .eomodel

2008-06-25 Thread David Avendasora
I've gone back and looked more at this problem. I do believe that it is a bug in WO that calling editingContext().objectsWithFetchSpecification(fs) from the client generates an error on the server. I wasn't seeing the error on the server before, but now I am. Here it is: [2008-6-25 16:49

Client security bug (WAS: Fetch Specifications defined in the .eomodel)

2008-02-17 Thread Florijan Stamenkovic
Dave, On Feb 16, 2008, at 12:12, David Avendasora wrote: So, if I understand you well, now that you've removed the SQL from the spec, you can use also stateless RMI to get it? Yep. I didn't change anything else. And before that you could only retrieve it with EO based RMI? Exactly! Thi

Re: Fetch Specifications defined in the .eomodel *FIXED*

2008-02-16 Thread David Avendasora
On Feb 16, 2008, at 10:22 AM, Florijan Stamenkovic wrote: [2008-02-15 16:53:14 EST] Server exception: The fetchSpecification was not allowed to execute on the server. If your application needs to execute this method, the security needs to be relaxed by implementing the delegate method

Re: Fetch Specifications defined in the .eomodel *FIXED*

2008-02-16 Thread Florijan Stamenkovic
[2008-02-15 16:53:14 EST] Server exception: The fetchSpecification was not allowed to execute on the server. If your application needs to execute this method, the security needs to be relaxed by implementing the delegate method distributionContextShouldFetchObjectsWithFetchSpecification

Re: Fetch Specifications defined in the .eomodel *FIXED*

2008-02-15 Thread David Avendasora
Turns out all the logic was right, but I had SQL in my Fetch Specification on the model. WO thinks this is less secure (which it is, I guess) so it was kicking out this error: [2008-02-15 16:53:14 EST] Server exception: The fetchSpecification was not allowed to execute on the server. I

Re: Fetch Specifications defined in the .eomodel

2008-02-15 Thread David Avendasora
On Feb 15, 2008, at 1:12 PM, Florijan Stamenkovic wrote: It seems like it's making the trip to the server, and making it as far as executing getFetchSpecification, but then fails and prints the exception above. What makes you say that? Seems to me that the stack trace stops on the client

Re: Fetch Specifications defined in the .eomodel

2008-02-15 Thread Florijan Stamenkovic
By the looks of it the connection between the client and the server got broken. As to why that has happened or where, you gotta figure that out. Skimmed through the code, looks OK. I got a Broken pipe exception some time in the past, but I totally forgot when, where, or what I did to fix i

Re: Fetch Specifications defined in the .eomodel

2008-02-15 Thread Florijan Stamenkovic
Ah, sorry Dave, just saw that I copied the same method twice... :) Here's the server side: public EOFetchSpecification clientSideRequestGetFetchSpecification (String spec, String entity){ try{ return EOFetchSpecification.fetchSpecificationNamed(spec, e

Re: Fetch Specifications defined in the .eomodel

2008-02-15 Thread David Avendasora
Hi Flor, What does the clientSideRequestGetFetchSpecification method that goes along with this in the Session class look like? Thanks, Dave On Jan 14, 2008, at 3:04 PM, Florijan Stamenkovic wrote: Server code, in the Session class: /** * Loads and returns an EOFetchSpecification object

Re: Fetch Specifications defined in the .eomodel

2008-01-14 Thread Florijan Stamenkovic
Stateless Remote Method Invocation is the only way I was able to get model defined fetch specs on the client. Client code: private static final EODistributedObjectStore dos = (EODistributedObjectStore)(EOEditingContext.defaultParentObjectStore()); /** * Invokes the remote method on sessio

Fetch Specifications defined in the .eomodel

2008-01-10 Thread David Avendasora
Hi all, especially you Java Client people out there. While I can easily build a FS in my client-side Java code and execute it, does anyone know if it is possible, and if so, how to, call a FS that is defined in the .eomodel file from a client-side class? Right now, if I try calling it, I ju