[omniORB] About IORs

Poilpret Jean Francois jfpoilpret@hn.vnn.vn
Tue, 21 Sep 1999 23:15:30 +0700


Hi Basma,

I'm not sure you actually answered Cristi's question, which was not =
about IORs for bound objects, but about IORs for name contexts, which is =
not the same, since contexts are recreated by omniNames when it is =
started (I mean, the second and next times), thus, object =
implementations do not need to recreate those context, at least they =
just should rebind their implementation objects.

Looking at the code in omniNames, it seems that its logfile contains =
only the object keys (not the complete IORs, which seems normal =
behavior) for contexts.
When it is launched the second and next times, it reads the logfile, and =
whenever it should (re-)create a context, it uses the saved object-key.
So the answer to Cristi's question is "yes", PROVIDED THAT omniNames is =
always run on the same host (in fact I'm not sure you can change =
omniNames' running location having it use a previous logfile from =
another host, ie, without having it restarting -with option -start - =
from the beginning).

Hope this helps,

    Jean-Francois

-----Original Message-----
From: Basma Driss Elkaleh <driss@enst.fr>
To: Cristi <bcristi@cti.ro>
Cc: OmniORB-List <omniorb-list@uk.research.att.com>
Date: mardi 21 septembre 1999 22:37
Subject: Re: [omniORB] About IORs


>
>No the IOR changes everytime you run the application. Consider the =
exemple eg2
>in  src/examples/echo, the IOR returned by eg2_impl does not remain the =
same.
>
>
>Cristi wrote:
>
>> Hi everybody.
>>
>> Can anyone answer this question?
>>
>> If nobody alters the logfile of omninames then the IORs of the =
context names
>> are the same everytime?
>>
>> Thanks a lot.
>
>--
>*******************************************************************
>Basma Driss Elkaleh
>Etudiante a l'universite de Stuttgart, Stagiere a l'ENST-Paris
>Bur: C214,  ++33-1-45 81 80 86
>
>Adr: 212, rue de Tolbiac, Ch. 404,F-75013 Paris
>Tel: ++33-1-45 81 79 35
>*******************************************************************
>
>
>