| How to wrap existing runnables with agents to create a heirarchical setup [message #83686] |
Sun, 17 September 2006 07:15  |
Eclipse User |
|
|
|
Hi,
I would like to test some already existing C++ programs which communicate
over the network with TPTP, but can not find good examples.
I have 2 main considerations:
1)
How can I wrap runnables into agents?
I would like the agents to start and configure the programs ,and in the
end kill them.
On this way a wouldn't have to worry about the distribution of the
programs.
I was reading the help pages and googling, but I couldn't find anything
about this matter. But it can also happen that I just overlooked some
pages.
2)
How can I build a hierarchical setup of Agents?
As far as I understood agents work in a master - slave fashion (1 master
controlls lots of slaves).
But I would like to use them in a herarchical manner. So for example the
root of the hieararchy would start up some nodes, which when they are
ready would also start up other nodes.
Is it possible for an Agent to be a master and a slave at the same time?
I know these questions sound silly, I just couldn't find any information
in the manuals and examples I have searched.
Thank you advance for your help.
Kristof
|
|
|
| Re: How to wrap existing runnables with agents to create a heirarchical setup [message #84060 is a reply to message #83686] |
Mon, 25 September 2006 07:21  |
| Eclipse User |
|
|
|
Hi Kristof,
These are all good questions. However, I need some more information to
fully understand your requirements:
1) Are you using TPTP as your test tool?
2) Do you want to write your own test tool using the TPTP framework?
3) What other tasks would your agents be handling other than
configuring/starting/stopping executables for testing?
Paul
"Kristof Szabados" <szabadosk@yahoo.com> wrote in message
news:fda1563905aaacaa846eb86c7ccfca8e$1@www.eclipse.org...
> Hi,
>
> I would like to test some already existing C++ programs which communicate
> over the network with TPTP, but can not find good examples.
>
> I have 2 main considerations:
> 1) How can I wrap runnables into agents?
> I would like the agents to start and configure the programs ,and in the
> end kill them.
> On this way a wouldn't have to worry about the distribution of the
> programs.
> I was reading the help pages and googling, but I couldn't find anything
> about this matter. But it can also happen that I just overlooked some
> pages.
>
> 2)
> How can I build a hierarchical setup of Agents?
> As far as I understood agents work in a master - slave fashion (1 master
> controlls lots of slaves).
> But I would like to use them in a herarchical manner. So for example the
> root of the hieararchy would start up some nodes, which when they are
> ready would also start up other nodes.
> Is it possible for an Agent to be a master and a slave at the same time?
>
> I know these questions sound silly, I just couldn't find any information
> in the manuals and examples I have searched.
>
> Thank you advance for your help.
>
> Kristof
>
|
|
|
Powered by
FUDForum. Page generated in 0.32497 seconds
Home