On Sep 5, 2008, at 8:09 AM, David Avendasora wrote:  
  On Sep 5, 2008, at 10:53 AM, Florijan Stamenkovic wrote:

[...]

Doh! Of course. I forgot about the oh-so-intuitive "? extends Class"  
syntax.

[...]

Once again, I have _no_ idea what you are doing / trying to do. [...]

[...]

In Dave's defense, his WO app makes delicious cookies, so he's doing  
something right. 

 ms  
 

Yeah, Dave, this sounds kind of evil. I mean, I see what you're  
trying to do, but man, it sounds fishy. And you for sure lose a lot  
of type safety. Though I *guess* if you're careful it could work. 

 One thing's for sure though: you have a talent for bizarre design :)  
 F   [...]

[...]

You get those messages too?  
 --  
Chuck Hill             Senior Consultant / VP Development 

Practical WebObjects - for developers who want to increase their  
overall knowledge of WebObjects or who are trying to solve specific  
problems. 
http://www.global-village. [...]

[...]

Please refrain from judging my parts :)  
 ms  
 

Whaaaat?!?! You don't trust me? I always do things the most efficient  
and supported way. I am offended! 

I mean really! My projects, utilizing Direct to Java Client with  
Vertical Inheritance on Microsoft SQL Server deployed to Tomcat  
running on Windows is a veritable tour-de-force of . . . [...]

[...]

Which is why I'm hesitant doing any crazy changes to EM here ... I'm  
still not sure there isn't something highly suspicious going on here. 

 ms  
 

http://cappuccino.org/

[...]

This isn't really a reliable way to implement this feature anyway,  
though.  You would need to make your session lifetime way to long for  
this to be useful.  A better way would be for you app to set a custom  
auth token cookie (that really should only be set over https, btw .. [...]

[...]

Doh! Of course. I forgot about the oh-so-intuitive "? extends Class"  
syntax.

[...]

Yep. After you make the relationship on the super, go to the sub and  
simply make the same relationship (same name) but pointing to a  
different Entity. In my case the the way it will be is: 

 In Part: [...]

Hi,  We're using WOMailDelivery with Java 1.5, mac intel, wo 5.4.2 and I'm seeing the following:  Anything changed with WOMailDelivery since WO 5.2.4? The app runs, however, it does not send the email its supposed to.  Thanks, Mersida  [2008-9-5 10:50:13 EDT] <Thread-2> com.webobjects.foundation. [...]

Uh, to be more pricisely, my problem is not because WO won't parse
empty cookie properly.

My problem is, cookie containing wosid and woinst will be deleted
after quitting browser. Hence, there's no way to implement function
like "login automatically. [...]

But you can override a method that returns NSArray<? extends

[...]

done this way you could have the relationships declared in all the  
Java classes. 

Still trying to figure out how Dave actually modeled this... Had no  
idea you could "override" relationship properties (destination) in   [...]

Peter:  
That was it. I never would have found that because for SQL Server 2000  
I never had to populate a value for "driver" in my connection  
dictionary. It always worked just with the url. I don't really  
understand why that changed or, better, why the adaptor was looking   [...]

[...]

I didn't realize that you could pass values in through the templates.  
That would make it really easy. I saw the getRelationshipNamed(String)  
in the code, but didn't think I could pass in the relationship name  
from the template. You know, pretty soon Veogen is going to catch  
on... ; [...]

[...]

Never tried, but does  
$ 
entity.parent.getRelationshipNamed(relationship.name).destination.name  
work? 

oh nevermind .. I just committed  
relationship.getParentRelationship ... It's only in the new branch,  
though, so you'll need to wait to switch to get it (I'm sort of done   [...]

Okay, but the relationshipOnParent doesn't exist in the current  
EORelationship api. 

I was just using it as an example of the type of addition that would  
need to be made. Or am I missing some existing api that will give me  
the EORelationship from the Parent? [...]

[...]

Given that nobody else has ever asked for this, I'm leaning towards  
not addressing it with some custom extension in the model ... You can  
already do this with other API, right (like the eogen sample code you  
showed)? 

Incidentally, your line: 
#if (($relationship. [...]

[...]

Or we could just make EORelationship.isInherited() return false if the  
destination of a subclass's to-one relationship is different than the  
superclass's relationship destination. 

This just bit me recently as well. Not only has the package changed,  
but the URL format for the driver has changed as well. It is no longer: 

 jdbc:microsoft:sqlserver://YourServer:1433;databaseName=YourDatabase  
 now it is:  
 jdbc:sqlserver://YourServer:1433; [...]

This just bit me recently as well. Not only has the package changed,  
but the URL format for the driver has changed as well. It is no longer: 

  jdbc:microsoft:sqlserver://YourServer:1433;databaseName=YourDatabase  
 now it is:  
  jdbc:sqlserver://YourServer:1433; [...]

Hello Yung-Luen;  
I think these empty cookies will parse in WO 5.4.2 now, but the  
application server will log an error each time which is a bit  
annoying.  I get my error log lines emailed to me so it was actually  
very annoying for a short while there! [...]

Hi,

Is this issue fixed in 5.4.2? I still encountered this problem (Not
only Safari 3, but also FireFox 3 delete my cookie on quit) in 5.4.2.

However, with Andrew's code inside my WOApplication.createRequest(), I
still got the same problem.
What could be wrong in my code? [...]

Hi Tim,  
If the version of the JDBC driver supports SQL Server 2005, then the  
driver class in the eomodel connection dictionary should be  
"com.microsoft.sqlserver.jdbc.SQLServerDriver" not  
"com.microsoft.jdbc.sqlserver.SQLServerDriver" as the error states  
below.

[...] [...]

All:  
I'm stumped. I have a working model and app that accesses a remote db  
server running SQL Server 2000. This service is being upgraded to  
SQLServer 2005 so I have test server to, um, test against. 

The drivers for the old server are embedded in my model and plist   [...]

[...]

As usual my comment comes a month too late. Oh well...  
In 2003 one of the most respected names in our industry - Martin  
Fowler - wrote a book with the dramatic title "Patterns of Enterprise  
Application Architecture". When I read it, I was shocked. [...]

Good to know if I have a need to use something like that again.

Thanks,
Guido

On 04.09.08 22:43, "Andrew Lindesay" <email@hidden> wrote:

[...]

Hello Guido;

[...]

Actually I should say that "very good" is a bit of an understatement   cheers.  
 ___
Andrew Lindesay
www.lindesay.co.nz  
 

Hello Guido;  
Yes OpenJMS is terrible under load.  ActiveMQ is a bit better, but  
SwiftMQ in my experience has been very good despite being a bit of a  
performance to configure. 

 	http://www.swiftmq.com/  
 cheers.

[...]

www.lindesay.co.nz  
 

Thanks for all your collective feedback.  
I'm going to experiment with a couple implementations to see what  
works best. 

 On Sep 4, 2008, at 9:30 PM, Guido Neitzer wrote:  
 On 04.09.08 22:23, "Andrew Lindesay" <email@hidden> wrote:

[...]

Hello Stephane,

Thanks for your prompt reply. I am exploring this issue in the manner you have advised.

Meanwhile, below are my findings from app:

The EmpCls entity in problem has couple of flattened attributes; flattened from a to-many relationship. [...]

[...]

Taught by bad experience with OpenJMS never to trust a JMS tool I still
prefer the database for that.

OpenJMS locks up under load. From my experience it is not a matter if it
locks up but when.

cug


 

Hello;  
I tend to use a JMS topic to distribute this information around  
instead.  This works very well. 

 cheers.

[...]

www.lindesay.co.nz  
 

[...]

Depends on how frequently you are updating that. Normally I don't think a
quick update and select would cause trouble.

But I have a feeling you're doing something weird ... I normally use either
the AjaxLongResponse or an ERXLongResponseTask and there the calling page [...]

That's an approach I thought about, and am considering.  
Do you have any concerns about read/write to the db being  
overwhelming? (We're using the status as a percentage of process  
completion.) 

 On Sep 4, 2008, at 8:24 PM, Guido Neitzer wrote:   [...]

What I normally do in these cases is store the status information in a table
in the database - as I really don't want to touch my session from a DA that
is called from a thread parallel to the RR loop.

I only have to make sure to always get fresh information from the DB, that's
about it. [...]

I'm not storing the id in cookies, as this will ultimately wind up as  
a web service. 

I have moved the thread, which seems to have helped (as there's no new  
session being created). Now I just need to figure out why the 'status'  
isn't being set (or read) correctly. [...]

On Sep 4, 2008, at 6:34 PM, Owen McKerrow wrote:  
 Hi Chuck,  
 On 05/09/2008, at 10:39 AM, Chuck Hill wrote:

[...]

No, savePage.  I grabbed the 5.4 API by mistake.

[...]

I don't use that.

[...]

Just a matter of me being in a hurry, clicking the wrong link and  
pasting the wrong method. [...]

So, I need a method on EORlationship so EOGenerator can check a  
Relationship to see not only if it is inherited, but if it's  
destination entity is different than the inherited method's  
destination Entity. 

 Currently I have:  
 #if (!$relationship. [...]

[...]

Okay - creating these rules is quite simple.
You have two properties lists:  
 er.extensions.ERXGenericRecord.relationshipCacheWhiteList = {\
	_ALL = "any,relationship,named,from,any,entity";\
	_MISFITS = "Entity,Names,That,Don't,apply,to,_ALL"; [...]

Hi Chuck,  
 On 05/09/2008, at 10:39 AM, Chuck Hill wrote:

[...]

I can see a savePage(WOComponent aPage) and a savePageInPermanentCache 
(WOComponent aPageComponent) in session, but no savePageInPageCache 
(WOComponent aPage). Im assuming you mean savePageInPermanentCache ? [...]

[...]

That would give you deadlock or unlocked access, not create a new  
session. 

 Are you storing the session ID in cookies too?  
  Chuck  

 On Sep 4, 2008, at 3:41 PM, Guido Neitzer wrote:  
Only thought I have is that it might not be a good idea to fire  
that DA from  [...]

I'm not swallowing exceptions anywhere, and if I am they're printing  
to stdout. 

I do think the session is locked or not checked in. I need to  
investigate further. 

 Thanks.  
 On Sep 4, 2008, at 3:41 PM, Guido Neitzer wrote:  
Only thought I have is that it might not be a good idea to fire that  
DA from 
a spawned thread from the RR loop. You might end up trying to check  
out a 
session from the session store that is already checked out. 

Not sure what would cause the creation of a new session though ...  
maybe 
that it is throwing an exception while trying to access the old (but  
locked) 
session and this could create a new session. 

 Are you swallowing exceptions somewhere?  
 cug  
 On 04.09.08 16:33, "Josh Paul" <email@hidden> wrote:  
 100%.  
 I left out the fact that this is related to a Thread, as that's where
the status is being created. So, I have a DA that fires the process
and another DA that checks in on it. The issue is that when checking,
a new Session keeps getting created.  
 Thoughts?  
 On Sep 4, 2008, at 3:26 PM, Guido Neitzer wrote:  
 And you are sure, your session is not terminated somewhere else?  
 cug  
 On 04.09.08 15:59, "Josh Paul" <email@hidden> wrote:  
 I'm using WO 5.3 with a version of Wonder about 1 week old...  
 On Sep 4, 2008, at 2:50 PM, Guido Neitzer wrote:  
 On 04.09.08 15:42, "Chuck Hill" <email@hidden> wrote:

[...]

[...]

Does that not greatly increase memory usage?  
  Chuck  

which is no where near as nice as what Wonder does, but does work.  
I've never had memory problems. 
Using this, you can instantiate the components in your DA, set them  
up however, then get the result string from generateResponse.... [...]

[...]

All pages that get created go into the page cache.

[...]

The stuff in Wonder is very cool...

[...]

Does the result returned from the DA have any component actions in  
it?  If not, you can disable all caching of pages returned by direct  
actions by adding this to Session: 
 [...]

I'm pretty sure that whenever you instance a new component using  
pageWithName... it gets added to the page cache. I get around this  
with the hackish method of adding this line in the components  
constructor: 

session().savePageInPermanentCache(this); [...]

[...]

 We've been using the ERXJGroups notification in our non-WOnder app without issues for a while now. The only issue we've been having is with JGroups itself and concurrent restart of some of our app instances.     I do remember though that I had to dive deep into WOnder to figure out all the dependencies and load order to make it work.     I suspect the reason why we haven't seen any issues yet is because we pretty much have all the editing context "fixes" that WOnder's ERXEC does and we don't have conflicts with delegates since don't setup our own delegates for anything.     The code that is used to initialise ERXJGroups is:      EODatabaseContext.setContextClassToRegister( ERXDatabaseContext.class );     /* initialise the synchronizer */  ERXObjectStoreCoordinatorSynchronizer.initialize();     This code is executed early on in the application loading (actually in the framework's principal class) before any model is loaded. Also, it uses an old version of WOnder (the latest one at the time I did this).     Regards        Peter Vandoros Software Engineer Etech Group Pty Ltd Level 3/21 Victoria St Melbourne VIC 3000 Australia  Ph: +61 3 9639 9677 Fax: +61 3 9639 9577 ---------------------------------- IMPORTANT: This e-mail message and any attachments are confidential and may be privileged. If received in error, please reply to this message and destroy all copies and any attachments. You should check this message and any attachments for viruses or defects. Our liability is limited to resupplying any affected message or attachments. For more information about Etech Group, please visit us at http://www.etechgroup.com.au.                

Hi All,  
We have an app that has some small hand made AJAX on it  (not the  
cool stuff from project wonder) and we're having an issue with page  
cacheing on the Applications side. 

I had set the AJAX action up as a call to a DA which I had though did  
not add to/effect the sessions page cache. [...]

Only thought I have is that it might not be a good idea to fire that DA from
a spawned thread from the RR loop. You might end up trying to check out a
session from the session store that is already checked out.

Not sure what would cause the creation of a new session though ... [...]

100%.  
I left out the fact that this is related to a Thread, as that's where  
the status is being created. So, I have a DA that fires the process  
and another DA that checks in on it. The issue is that when checking,  
a new Session keeps getting created. 

 Thoughts? [...]