correct way of running a sub-process

[Jp Calderone]

On Thu, Feb 12, 2004 at 06:10:01AM -0800, Daniel Timothy Bentley wrote:
> What is the (hopefully unique) obvious way of runnign a sub-process if I
> want to get the exit code and input without resorting to multi-threading?
> 
> It seems like I should be able to do the following:
> 
> foo = popen2.Popen3(cmd)
> foo.wait()
> foo.fromchild.read()
> 
> But, it seems like on my system (a Sun Ultra 30 running Solaris 8) this
> will hang if the input happens to have more than about 12k of text.  I'm
> guessing this is a buffersize issue, but it would be a mistake to assume
> at compile time an output size for the child process.
> 
> Couldn't wait be modified to actually wait until a process exits by
> buffering?
> 
> While we're at it, is there any way to read/wait with a time-out?  I.e.,
> read more than 0 bytes, unless it takes more than a time-out value, so I
> can run sub-processes that terminate?
> 

  (Untested)

    from twisted.internet import protocol, reactor

    class PrintyProtocol(protocol.ProcessProtocol):
        bytes = ''
        errBytes = ''

        def outReceived(self, bytes):
            self.bytes += bytes

        def errReceived(self, bytes):
            self.errBytes += bytes

        def processEnded(self, reason):
            print 'Process done.  Got:', repr(self.bytes), repr(self.errBytes)
            reactor.stop()

    reactor.spawnProcess(PrintyProtocol(), cmd, args=(cmd,))
    reactor.run()

  Jp