The patch for addEntry method is available here:

http://nagoya.apache.org/bugzilla/showattachment.cgi?attach_id=830

As far as the remove problem is concerned, when I remove a portlet registry
entry, am I not removing it from registry stored in memory? I can actually
query the registry immediately after removal and entry is not there.
However, when I restart, it comes back. It's been a while since I played
with this. When I get some time, I'll try to verify whether the workaround I
suggested even works. If you do it sooner, I wouldn't mind knowing the
result.

Best regards,

Mark C. Orciuch
Next Generation Solutions, Ltd.
Voice: 219-365-0691
e-Mail: [EMAIL PROTECTED]
web: http://www.ngsltd.com


-----Original Message-----
From: Meik Arends [mailto:[EMAIL PROTECTED]]
Sent: Tuesday, December 18, 2001 8:36 AM
To: Jetspeed Users List
Subject: Re: Update portlets.xreg


Mark Orciuch wrote:

>Meik,
>
>Be aware of some outstanding problems with the current
CastorRegistryService
>when registry is changed programatically. I also have an application which
>manipulates portlet registry entries at runtime. It was orignally developed
>using 1.3a1 code but when ported to 1.3a2, I started experiencing some
>problems. One of the problems has to do with adding new entries (see
>Bugzilla Bug 5091). I sumbitted a patch but the fix seems to be a lower
>priority right now. Another problem that I ran into has to do with removing
>registry entries - they seem to mysteriously come back after the portal is
>restarted.
>
OK. I'm not alone :-) Can you send me your patch ?
I think your other problem is related to the things Paul Spencer
explains in his reply to my question.
 He said:

"Also Jetspeed will write the .xreg files at shutdown using the entries
it has  in memory.  If you make a change that is not read by Jetspeed
BEFORE shutdown, then the change will be overwritten at shutdown. "

>I have inquired about registry refresh capability as well. If I remember
>correctly, calling existing refresh method will make any programatic
changes
>disappear. Quick workaround to your problem would be to work with specific
>fragments and explicitly call saveFragment and then refresh immediately.
>However, this approach is not recommened by Raphael Luta since the
>implementation may change and could become database-based.
>
I have to try it, because there is no time to wait for DB-support ;-)

>
>
>I'd like to get involved in getting these problems or shortcomings resolved
>but there's a lot of questions I have about how this stuff is intended to
>work.
>
>Hope this is related to your issue. Otherwise, kindly disregard it.
>
>Best regards,
>
Regards

Meik

--
----------------------------------------------------------------
                          Meik Arends
                        [EMAIL PROTECTED]
     WAP-Mail: [EMAIL PROTECTED] ( f�r wirklich dringendes ;-)
                         ICQ: 42486557
----------------------------------------------------------------




--
To unsubscribe, e-mail:
<mailto:[EMAIL PROTECTED]>
For additional commands, e-mail:
<mailto:[EMAIL PROTECTED]>


--
To unsubscribe, e-mail:   <mailto:[EMAIL PROTECTED]>
For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>

Reply via email to