PSSI.01 error on Initialize (OpenInsight 64-bit)
At 14 NOV 2023 10:57:11PM Matthew Crozier wrote:
It's possible that as part of the application initialisation, the CREATE event of our main MDI frame can send a CLOSE event to logout under certain conditions. In OI10 this can lead to a PSSI.01 error during PS_SYSINITIALIZE:
How can I signal during the CREATE event that the session should be closed cleanly (allowing the CLOSE event to do a lot of tidy up) without upsetting the application initialization?
I've tried the following to no avail:
- using Post_Event( 'CLOSE' ), to allow the CREATE event to finish completely before closing. It seems this is still yielded before initialization is complete.
- Call Set_EventStatus( 1), in the hope of avoiding any further processing
- moving this code to be called in the Finalize section of the CREATE Quick Event
Cheers, M@
At 15 NOV 2023 09:30AM Carl Pates wrote:
HI M@,
OK - the system isn't really set up for that scenario (killing it _silently_ during bootup) - That's by design - it checks to make sure that there _is_ a visible form after startup to make sure that we don't have orphaned invisible instances of OI hanging around.
It would take a modification to add an option to suppress that message and at this stage in the game I don't see it making 10.2.1 because that really really does have to get released - we've been holding it back so long now to try and fit everything in that we can, but we're at cutoff.
So, I can add an option (cmdline/rxi switch and SYSTEM property) for 10.3.
However, in the meantime a possible workarounds could be:
1) Display a dialog saying "Computer says No", and when they click on OK then kill the system. I assume you're not keen on this for reasons.
2) Display a tiny dialog with 99% translucency, and a posted omnievent or IDLEPROC that checks the return stack for PS_SYSINITIALIZE - if it's not there kill the system.
Cheers
At 15 NOV 2023 06:31PM Matthew Crozier wrote:
I can add an option (cmdline/rxi switch and SYSTEM property) for 10.3.
However, in the meantime a possible workarounds could be:
1) Display a dialog saying "Computer says No", and when they click on OK then kill the system. I assume you're not keen on this for reasons.
2) Display a tiny dialog with 99% translucency, and a posted omnievent or IDLEPROC that checks the return stack for PS_SYSINITIALIZE - if it's not there kill the system.
Thanks Carl,
We're still a long way off releasing an OI10 version, so can wait for a SYSTEM property in 10.3 :) .
Yes, killing the system straight out would mean reorganising a few steps in the initialisation so things don't get left hanging. It's a bit awkward as there are a few cross-dependencies. There's a few things I can move to IDLEPROC though.
Cheers, M@