Re: spawn II--the Sequel

Richard Schoen wrote:

So when I call spawn(), is the job spawning as part of the same iSeries
job or separately ?

spawn() is similar, conceptually, to the SBMJOB command. It starts a new job, and runs a program in that new job.

However, with spawn(), the new job is a "child job" of the current one, and will report it's status to the current job. Consider looking up spawn() in the Information Center for more details.

What makes it really attractive in your situation is that you can start the job with file descriptors open that map to pipes, and you can read the pipes from your current job. You can't easily do that with SBMJOB.

But, it definitely starts a separate job.

I'm actually calling a CL program that calls the QP2SHELL to call the
SAMBA API's to transfer a file. I need to capture the STDOUT responses
so I'm thinking your RPG spawn example should work as a starting point
and I want to get the data from the pipe that captures STDOUT, correct ?

I've written quite a few RPG spawn examples -- so I'm not sure which one you're referencing -- but I guess it doesn't matter... spawn() should be able to run PASE programs and use it to capture the output. waitpid() should be able to give you the exit status, as well.

That assumes you're running a Samba program -- like smbclient -- as opposed to an actual API (calling a function in a shared library).