Hey Arkaitz,
If all you will be committing is the diff you sent, plus a new "rm_pbs.c",
then go ahead and check it in. I'll be working from the same code base, but
the changes seem pretty independent of my stuff. Make sure and 'svn update'
before you commit, as I have been making changes recently that you need to
reflect in your version, so you don't overwrite my changes. :-)
For the code change, can we call it "resource manager endpoint" rather than
"queue node"? This is a bit more generic for potential other resource
managers that might ask for a config file, or http URL, or something like
that to be provided to their RM backend. So basically it becomes '-r' rather
than '-q' and 'rm_endpoint' rather than 'queue_node'. I know it seems
cosmetic, but if it's not too much trouble.....
-- Chris
On 7/11/08 08:56 , "Arkaitz Ruiz-Alvarez" <ar...@cs...> wrote:
> Hi Chris,
>
> I'm ready to commit the first version of the bespp's besserver with
> PBS code. Changes to the common code are minimal, I just added a new
> command line parameter called -q queue_node. That is, I usually run
> besserver on a machine different than the queue I want to submit jobs to
> so I need to pass the machine name as a command line argument. The
> string is passed to rm_initialize as the second argument. Since
> rm_initialize in rm_lsf.c ignores that argument it should not affect
> your code. I have attached a path for besserver.c
>
> However, since Supercomputing is so close, I don't know if you prefer
> me not to touch the current repository. Are you going to use the code in
> Supercomputing? If so, do you want me to wait to commit my changes?
|
