Re: [Help-bash] Getting PID

[Danny]

O.k ... I give up ... lol ...

On Jan 01 15, Dave Rutherford :
> To: help-bash <address@hidden>
> Date: Fri, 1 Jan 2016 13:11:59 -0500
> From: Dave Rutherford <address@hidden>
> Subject: Re: [Help-bash] Getting PID
> X-BeenThere: address@hidden
> 
> On Fri, Jan 1, 2016 at 12:22 AM, Danny <address@hidden> wrote:
> 
> > So I was thinking ... :
> >
> [...]
> 
> > -get gtkterm's PID and pipe it into /var/lock/LCK..ttyS1
> > -send the required data to /dev/ttyS1
> 
> 
> You appear not to understand the role of the lockfile. The contents
> thereof are not used to determine who has access to the port;
> they are merely advisory. The lockfile would work just as well if
> it were always empty. It indicates "a process has obtained
> exclusive access to this comm port." That exclusive access is
> enforced by the kernel, not by something in the filesystem.
> 
> So, what you are thinking will not work. You are better off
> with the method you're using now, awkward as it is. You
> might be able to do something more elegant using ckermit,
> though I can't promise anything there.
> 
> > What makes you think the last PID is the one you want to use in your
> > > script?  How would your script act in this case:
> > >
> > > address@hidden: pidof application
> > > 14 32765 32767
> >
> 
> Also pay more attention to that observation. If pids wrap around
> while you are starting your terminal, is it still the "last one"
> which is listed in the lockfile? What would happen running your
> program on a system that uses random PID order?