HeroEngine Forums
Welcome, Guest. Please login or Register for HeroCloud Account.

Author Topic: another probable teleport robustness bug  (Read 1103 times)

AWW_boss

  • General Accounts
  • *
  • Posts: 846
  • gdfgd
    • View Profile
another probable teleport robustness bug
« on: Jun 07, 14, 04:33:01 PM »

Every time i run into one of those i feel sick to stomach,  because i can't do much about it. And it takes a lot of energy from my side to describe it in a manner that it would be easily understood.
Can you please check this scenario and see if you can reproduce it? The problem is even more complicated because it is not 100% reproducible. The frequency of fault is i'd say 30 out of 70 teleports.
So here goes:

The problem in this specific scenario is teleportation from area B to area A, where red circle represents my player character called Scripto(and his awareness range), while blue circle represents remote player character called Fafko(and his awareness range).



This teleport results in one of two outcomes: correct and incorrect

1.
We have a correct outcome which is described with this set of debug traces;
[hsl_debug] 9223372088885664364,2: SCRIPT[E_playerAccountClassMethods]: entered | entity: scripto | subject: fafko | 5910.366210938
[hsl_debug] 9223372088885664364,2: SCRIPT[E_playerAccountClassMethods]: entered | entity: fafko | subject: scripto | 5910.366210938
[hsl_debug] 9223372088885664364,2: SCRIPT[E_playerAccountClassMethods]: departed | entity: fafko | subject: scripto | 5910.366210938
[hsl_debug] 9223372088885664364,2: SCRIPT[E_playerAccountClassMethods]: departed | entity: scripto | subject: fafko | 5910.366210938


Note that in this outcome departed event is called on both characters

2.
We also have an incorrect outcome which is described with this set of debug traces:

[hsl_debug] 9223372088885664364,2: SCRIPT[E_playerAccountClassMethods]: departed | entity: fafko | subject: scripto [proxy] | 5910.366210938


Note that in this outcome departed  event is called only on one character (the remote player) and it should be called on both characters <-----  THIS IS THE PROBLEM



thanks
 
Logged

HE-Cooper

  • *****
  • Posts: 2221
    • View Profile
Re: another probable teleport robustness bug
« Reply #1 on: Jun 07, 14, 07:02:23 PM »

I'm going to add a clean test world to your group, because that will cost me less money than having engineers try to reproduce your teleport issues. :-)

That way you can create a reproduction scenario that is independant of your world's codebase.
Logged

AWW_boss

  • General Accounts
  • *
  • Posts: 846
  • gdfgd
    • View Profile
Re: another probable teleport robustness bug
« Reply #2 on: Jun 08, 14, 01:08:06 AM »

I already have one clean world. I will try to reproduce the fault on that one, although i might have to introduce some code to setup a test scenario.

But taken into consideration the changes you did to SAS/seamless2 in quartz (which were great btw), i can come to conclusion that some of the teleportation issues i reported were not introduced by my code.

As SAS and seamless2 are fundamental parts of my game, could you somehow teach us how to debug it better, especially SAS? perhaps introduce some cli command that would turn on more debugging info which would tell us what caused some SAS event to fire or misfire?

Logged

HE-SCOTT

  • HeroEngine
  • *****
  • Posts: 158
    • View Profile
Re: another probable teleport robustness bug
« Reply #3 on: Jun 09, 14, 03:25:22 PM »

The situation you have described was believed to have been fixed in a Quartz update. We will double check to make sure that is the case and we did not miss anything. Also we will make sure the fix was correctly applied to your world. Setting up a test case in a clean world like Cooper suggested would be useful if we do not find anything. Try and keep the test case as simple as possible since we want to know exactly what is happening if it fails.

The SAS system already provides a good set of debugging tools through the /hesas chat command, there is more on it here http://wiki.heroengine.com/wiki/Spatial_Awareness_System#Debugging. Also there are many things it can do that you can find out more about by running the command with no parameters. In HSL there is an error callback when a SAS event can not be called for one reason or another.

Code: [Select]
method _SAS_FailedEventNotificationForNodeID(entity_ID as ID, nodeID as ID)
This happens most commonly when a node has been deleted but not removed from the SAS for example.
Logged

AWW_boss

  • General Accounts
  • *
  • Posts: 846
  • gdfgd
    • View Profile
Re: another probable teleport robustness bug
« Reply #4 on: Jun 11, 14, 10:05:39 AM »

That's very useful info.
All right, let's put this on hold, until i find a way to reproduce it in clean engine and check the debugging tools you suggested.
Thanks
Logged