[Info-vax] Looking into C-include files on VMS

[Bob Eager]

On Thu, 29 Oct 2009 08:04:25 -0600, Bob Koehler wrote:

> In article <7krs6vF3af5v3U10 at mid.individual.net>, Bob Eager
> <rde42 at spamcop.net> writes:
>> 
>> You miss the point. They are talking about VMS SPAWN, which is unlike
>> vfork()+exec(). It is nearer to creating a new (empty) process and
>> loading a program into it.
> 
>    Doesn't that end up with the same thing?  I would expect vfork() to
>    start a new process and exec() to load a new program into it.

No, it is *much* more expensive (no criticism of VMS, which has a 
completely different philosophy), and one would not want to do it too 
often.

vfork() doesn't start a new process. fork() does, and clones the existing 
process into it. vfork() effectively starts a new thread in the current 
process, and on the next exec(), the new process is created. But the UNIX 
philosophy is that this is a cheap operation (it doesn't have F11ACP 
mapped into it (if that still happens), nor DCL).

>    In fact, the C RTL does have a pair of routines (or macros in the
>    approriate header file) called vfork() and exec() that end up
>    spawning a new process running a new program.  Just don't expect the
>    implementation to look like those in UNIX.

Exactly. It 'costs' quite a bit more. Fair enough, but it means that a 
proper port ought to consider those issues and not just do them 'the UNIX 
way'.



-- 
Use the BIG mirror service in the UK:
 http://www.mirrorservice.org