Server crashes after linking in to Gahreesen #4 cloth

[diafero]

Don't bother with this... blockers are still crashing Uru... Will have to find why.

What exactly does that mean? Colliding with a blocker crashes the client?

[diafero]

I just noticed that, when actually entering the wall, the server kicks the client due to an unknown message (0x044C). I will add support for that message to Alcugs later today... hopefully that should fix the remaining wall online issues

[tachzusamm]

What exactly does that mean? Colliding with a blocker crashes the client?

In my humble experience, colliding with a blocker usually crashes my nose.
*SCNR*

[Sirius]

In my humble experience, colliding with a blocker usually crashes my nose.

ROTFL !

Ok, I managed to make all the responders, behaviors, etc not netforced. The only exception is the sound responder for the panel buttons (it doesn't send a lot of notifies and all the Python checks (for example the blockers) are skipped). I also fixed some responders called twice or physics enabled/disabled twice, and removed some useless SDL vars.
There are still things I would like to know:
- are ptSceneObject.runAttachedResponder(state) and ptSceneObject.fastForwardAttachedResponder(state) netforced ?
- is the clickable for a SittingModifier disabled for everyone in an Age when someone uses it ?
- are exclude regions netforced when controlled by a Python file ?

[diafero]

I was just able to setup up the wall, enter the wall area, climb it up, collide with a blocker, and then reach the top, on the DI Shard. So it seems those issues are fixed by now - the actual multi-player testing can start as soon as I got the Python updates from you, Sirius

The person most likely to answer your questions is, I think, D'Lanor.

[D'Lanor]

The person most likely to answer your questions is, I think, D'Lanor.

OK, I'll give it a try.

There are still things I would like to know:
- are ptSceneObject.runAttachedResponder(state) and ptSceneObject.fastForwardAttachedResponder(state) netforced ?

No. Scene object attributes are not netforced by default. They are only netforced when you set a hard netforce flag.

Code: Select all

ptSceneObject.netForce(1)

Once this flag is set anything you do with the object will be netforced until you unset that flag with:

Code: Select all

ptSceneObject.netForce(0)

- is the clickable for a SittingModifier disabled for everyone in an Age when someone uses it ?

I never managed to check because the sitting avatar's fat butt is always in the way. I doubt a programmer would automatically disable something that is out of reach anyway.

- are exclude regions netforced when controlled by a Python file ?

PlasmaTypes.py does not give us any clues about netforcing or netpropagating of exclude regions. However, in order to do their job properly I suspect that they need to be either netpropagated or netforced. I'm guessing netpropagation because netforcing seems a bit strong for this.

[Sirius]

I was just able to setup up the wall, enter the wall area, climb it up, collide with a blocker, and then reach the top, on the DI Shard. So it seems those issues are fixed by now

Good news ! Do you know what was crashing it, or was it a random bug ?


Thank you D'Lanor for these helpful answers.

There are still things I would like to know:
- are ptSceneObject.runAttachedResponder(state) and ptSceneObject.fastForwardAttachedResponder(state) netforced ?

No. Scene object attributes are not netforced by default. They are only netforced when you set a hard netforce flag.

Code: Select all

ptSceneObject.netForce(1)

Once this flag is set anything you do with the object will be netforced until you unset that flag with:

Code: Select all

ptSceneObject.netForce(0)

Ok, great - these are used to run layer anims of blockers, and are played by every client. That would be really bad if they were netforced.

- is the clickable for a SittingModifier disabled for everyone in an Age when someone uses it ?

I never managed to check because the sitting avatar's fat butt is always in the way. I doubt a programmer would automatically disable something that is out of reach anyway.

Hehe. For some reason Cyan did for the panel chairs... they even added an SDL var to check it isn't currently used . Looks really useless.

- are exclude regions netforced when controlled by a Python file ?

PlasmaTypes.py does not give us any clues about netforcing or netpropagating of exclude regions. However, in order to do their job properly I suspect that they need to be either netpropagated or netforced. I'm guessing netpropagation because netforcing seems a bit strong for this.

Yes, there is only one other python file using exclude regions without responders, and there is no avatar check for it. So I guess the default value is not netpropagated. That's better, because on link-in a player automatically "clears" the region for the tubes.


I corrected other errors in the Python file (some sounds for the panels were not played correctly), there is just a problem with the blocker counter lights. Should be easy to fix.
I just hope multiplayer won't behave wrong.

[diafero]

I was just able to setup up the wall, enter the wall area, climb it up, collide with a blocker, and then reach the top, on the DI Shard. So it seems those issues are fixed by now

Good news ! Do you know what was crashing it, or was it a random bug ?

It wasn't crashing, the server just kicked the client because it sent an unknown message (plClimbingMsg).

[D'Lanor]

- are exclude regions netforced when controlled by a Python file ?

PlasmaTypes.py does not give us any clues about netforcing or netpropagating of exclude regions. However, in order to do their job properly I suspect that they need to be either netpropagated or netforced. I'm guessing netpropagation because netforcing seems a bit strong for this.

Yes, there is only one other python file using exclude regions without responders, and there is no avatar check for it. So I guess the default value is not netpropagated. That's better, because on link-in a player automatically "clears" the region for the tubes.

Actually that only happens when the player is alone. Note the return under "if len(PtGetPlayerList()):". Weird enough this also skips the ageSDL initialization in multiplayer mode. That does not seem right to me. I would move this section...

Code: Select all

        ageSDL = PtGetAgeSDL()
        ageSDL.setFlags('nChairOccupant', 0, 1)
        ageSDL.setFlags('sChairOccupant', 0, 1)
        ageSDL.setNotify(self.key, 'nChairOccupant', 0.0)
        ageSDL.setNotify(self.key, 'sChairOccupant', 0.0)
        ageSDL.sendToClients('nChairOccupant')
        ageSDL.sendToClients('sChairOccupant')

...to the beginning of OnServerInitComplete().

These SDL variables are not so much for the chair itself but to make sure that only one person can use the control panel.

[Charura]

With all that has been done I am able to maintain continuity in Gahreesen..and that includes any aspect of the age..Now, as far as figuring out the wall, the fun begins? Great job folks!!!