🚀 go-pugleaf

In article <BxytAB.8KB@newcastle.ac.uk> unregistered@newcastle.ac.uk (Jason O'Broin) writes:
>Is there anyway of intercepting the RiscOs 3.1 Desktop shutdown mechanism? I need
>to be able to do some processing before this is allowed to continue.
>In addition to this, I need to be able to distinguish between this and Exiting
>the desktop (ie. closing down the desktop is not the same as starting the
>shutdown procedure). Does anyone know the actual operations performed by
>shutdown. Ideally, I'd like to replace it completely.

I believe there is a service call (Service_Shutdown?) which is issued
when a shutdown occurs. I think one of its parameters is a flag to say
whether it is an 'exit' or a 'shutdown'. I suspect that this call is
only issued after the Wimp and OS have done most of their shutdown
processing.

I'll have a look in my PRMs tonight and flesh out the details for you.

Tom.

============================================================================
Tom Hughes, CS3

thughes@uk.ac.warwick.dcs            57 Shakleton Road
csugk@uk.ac.warwick.csv              Earlsdon
                                     Coventry
(0203) 673584                        CV5 6HT
============================================================================

In article <BxytAB.8KB@newcastle.ac.uk>, unregistered@newcastle.ac.uk (Jason O'Broin) writes:
>
> In addition to this, I need to be able to distinguish between this and Exiting
> the desktop (ie. closing down the desktop is not the same as starting the
> shutdown procedure). Does anyone know the actual operations performed by
> shutdown. Ideally, I'd like to replace it completely.

I believe the mechanism is almost identical between RiscOS3.1 and previous
versions, in terms of its external interface. The difference between the
shutdown dbox and merely leaving the desktop is slight.

Basically, all tasks are first sent a PreQuit message (number 8). If none of
them respond, they're all sent Quit (0) messages. So - the former should be
handled if you want the opportunity to object to the shutdown, or the latter
can be handled if you just want to do something.

Under RiscOS3, the PreQuit message *may* have a flag word at address 20 in
the message block (if the block is large enough). The bottom bit of this is set
if it's your application being quit, or clear if a shutdown is happening.

--Clive.

In article <1992Nov20.020639.10146@infodev.cam.ac.uk> crj10@phx.cam.ac.uk (Clive Jones) writes:
>In article <BxytAB.8KB@newcastle.ac.uk>, unregistered@newcastle.ac.uk (Jason O'Broin) writes:
>> In addition to this, I need to be able to distinguish between this and Exiting
>> the desktop (ie. closing down the desktop is not the same as starting the
>> shutdown procedure). Does anyone know the actual operations performed by
>> shutdown. Ideally, I'd like to replace it completely.
>
>I believe the mechanism is almost identical between RiscOS3.1 and previous
>versions, in terms of its external interface. The difference between the
>shutdown dbox and merely leaving the desktop is slight.
>
>Basically, all tasks are first sent a PreQuit message (number 8). If none of
>them respond, they're all sent Quit (0) messages. So - the former should be
>handled if you want the opportunity to object to the shutdown, or the latter
>can be handled if you just want to do something.
>
>Under RiscOS3, the PreQuit message *may* have a flag word at address 20 in
>the message block (if the block is large enough). The bottom bit of this is set
>if it's your application being quit, or clear if a shutdown is happening.

Right. I looked this up in the RiscOS 3 PRMs last night, and whilst
the above info is correct, there are some new calls.

Firstly, there is a new messagem, Message_Shutdown (message 14) which
I think is only issued on the portable when an enforced shutdown
occurs (due to battery failure). I suspect this replaces the
Message_PreQuit/Quit sequence, as applciations should try to save and
exit gracefully when they receive it (accoring to the manuals).

Secondly, there are two service calls. These are Service_Shutdown
(&7E) which is issued before a shutdown, and may be claimed to block
the shutdown, and Service_ShutdownComplete (&80) which is issued after
a shutdown is completed.

One other clever (?) way of replacing the shutdown sequence that I
thought of is to install a post-filter on the task manager, watch for
Shutdown/Exit being selected from the menu of ctrl-shift-F12 being
pressed, and then mask these out from the task manager and do your own
thing. This is however truly sick.

Tom.

============================================================================
Tom Hughes, CS3

thughes@uk.ac.warwick.dcs            57 Shakleton Road
csugk@uk.ac.warwick.csv              Earlsdon
                                     Coventry
(0203) 673584                        CV5 6HT
============================================================================