Re: Until loop

In article <[email protected]>,
   Richard Torrens (RiscOS) <[email protected]> wrote:
> In article <[email protected]>,
>    Richard Torrens (RiscOS) <[email protected]> wrote:



> It's  clearly not easy. If both Blackholes are reset together then it's
> indeterminate when the network's going to be ready again, but if I
> reset sequentially, then the one which starts first should be able to
> tell the parent process when it's finished - as now the network will be
> live again.

I'm not clear whether the problem lies in the notification method you are
using or in getting the Blackhole to do something at the relevant point.

My experience on setting up signals on multitasking environments (big
mainframe systems) tells me that finding a shared directory (actually
they are all shared on the mainframes) and creating a flagfile with a
meaningful name is the easiest way of doing something that another system
can detect.  After all, system variables are not in sync across all
systems, but existence/otherwise of files is.  In some cases such files
have fixed names, in others I've created names like: 

    system.application.stopped.for.backup.at.<date>.<time>

so that an interrogating system can see the status and date/time info. 
of course if you do this you need to make sure that the code (obey file?)
that starts an app wipes out the "I'm stopped" file.  You might do that
by having a 

    *wipe system.app.**  

(or whatever command) followed by a 

    *create system.app.up.at.<systime>

or something...

-- 
Jeremy C B Nicoll - my opinions are my own.