Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!usc!wuarchive!uunet!overload!dillon
From: dillon@overload.Berkeley.CA.US (Matthew Dillon)
Newsgroups: comp.sys.amiga.programmer
Subject: Re:  CShell 5.10, arp stuff, A3000 w/2.0; should they go together
Message-ID: <dillon.8133@overload.Berkeley.CA.US>
Date: 29 May 91 20:42:01 GMT
References: <1845@contex.contex.com> <37100008@hpfcdc.HP.COM>
Organization: Not an Organization
Lines: 66

In article <37100008@hpfcdc.HP.COM> koren@hpfcdc.HP.COM (Steve Koren) writes:
>
>>	RunCommand(seglist,stacksize,arguments,argument length).  This is what
>>is used by the system shell to run commands on your process.
>
>Well, I admit I don't have the docs for RunCommand, but from the arguments
>to it listed above, it appears there is no way to redirect input and
>output to the command.  If true, that would make it unusable for a shell.
>
>  - steve

    There are THREE main 2.0 calls associated with running programs:

    (1) RunCommand()

	RunCommand() runs a seglist under THE CALLERS CONTEXT, using the
	caller's task.  THIS IS THE COMMAND A SHELL WRITER WOULD USE TO RUN
	COMMANDS.  The caller must LoadSeg() or FindResident() (and bump
	the resident tag count as appropriate) to get a segment which may
	be used in a call to RunCommand().  The argument line must be
	terminated with \n\0 instead of just \0 to be compatible with
	resident C: commands.

	All other setup is up to the calling program (in terms of setting
	up input, output, etc...)

	Since no new process is created and the seglist may be pre-cached
	or found in a resident list, this call can be made to run VERY
	quickly.

    (2) System()

	System() will start up a shell, executable your command (like
	Execute()), then return the exit code.  It has options similar to
	CreateNewProc() below in terms of directing input and output and
	choosing between the system shell or a custom shell.

	Commands run under System() run in a separate shell context,
	meaning that ^C-^F will not be propogated.  However, full shell
	command lines are allowed, including redirection on the command
	line, and use of internal shell commands.

	Running a command via System() is slower than RunCommand() since
	a separate process will be created, a shell/cli run, and the
	command then interpreted by the shell/cli.

    (3) CreateNewProc()

	Create a new process specifying either an entry point or seglist,
	with options to set input, output, current dir, stack, name,
	priority, console task, windowptr, CLI, etc etc etc....

	This call is faster than System() but slower than RunCommand().
	However, it is an excellent method to start up asynchronous
	processes or tasks (such as servers, handlers, co-processes...)


				    -Matt


--

    Matthew Dillon	    dillon@Overload.Berkeley.CA.US
    891 Regal Rd.	    uunet.uu.net!overload!dillon
    Berkeley, Ca. 94708
    USA