First snippet of result, needs more work.  Note the AIR_PATROL event nested in the LAUNCH_EVENT.  Those are our suspects.  Hehe, nice Lat/Lon too.

	<EVENT group_id="YRb" unit_id="255" event_type="LAUNCH_EVENT" time="1" time_type="TimeAbs" dummy_1="204" dummy_2="204">
		<DATA>
			<LAUNCH unit_id="0" plane_type="11581" loadout_idx="32" qty="1">
				<DESTINATION dest_lat="2348.823532" dest_lon="-2258.823532"/>
				<AIR_PATROL plane_annex_id="11581" loadout_idx="32" patrol_type="ASW_PATROL" qty="1" patrol_status="AAW_PATROL">
					<RINGS>
						<RING>ASW_RING</RING>
					</RINGS>
					<SECTORS>
						<SECTOR>FS_045_090</SECTOR>
					</SECTORS>
				</AIR_PATROL>
			</LAUNCH>
		</DATA>
	</EVENT>

 

2022.017 posted.  It would be great if someone could re-do the https://harpgamer.com/harpforum/topic/32501-air-patrols-crashing-unable-to-deleteait-patrols/?do=findComment&comment=62127 testing since I fixed an off by one error in the debug logging.  The writing of XML files for events is in place as well when I32501 debugging is enabled, feel free to play with that.  I haven't been able to glean any nuggets with that logging yet, especially since it doesn't log the group id of the AirPatrol event embedded in the LaunchAir event (yet). 

 

2022.018 posted.  2022.018 testing situation is to start the scenario fresh and play it as long as possible without loading from a saved game.  This is to bolster the theory that the corruption is happening during game saving and/or loading of a saved game.

Eric thinks he will give it a try today.

And 2022.019 posted which includes the group id of the AirPatrol event associated with a LaunchAir event.

 

<EVENT group_id="ZPC" unit_id="255" event_type="LAUNCH_EVENT" time="2" time_type="TimeAbs" dummy_1="204" dummy_2="204">
		<DATA>
			<LAUNCH unit_id="109" plane_type="10467" loadout_idx="32" qty="1">
				<DESTINATION dest_lat="2348.823532" dest_lon="-2258.823532"/>
				<AIR_PATROL group_id="ZPC" plane_annex_id="10467" loadout_idx="32" patrol_type="AEW_PATROL" qty="1" patrol_status="PATROL_TRANSIT">
					<RINGS>
						<RING>AAW_RING</RING>
					</RINGS>
					<SECTORS>
						<SECTOR>FS_045_090</SECTOR>
					</SECTORS>
				</AIR_PATROL>
			</LAUNCH>
		</DATA>
	</EVENT>

 

I used 2022.019 to load Eric's scenario (not saved game, I loaded the .SCq file) at zero time compression.  I immediately saved the game.  Then I closed HC, started it up again and loaded the saved game I had just created.  Here is the comparison of LaunchAir events with associated AirPatrol events so obviously there is something wrong in the saving and loading as no gametime elapsed.

https://extendsclass.com/xmldiff/4271838

1 minute ago, TonyE said:

I used 2022.019 to load Eric's scenario (not saved game, I loaded the .SCq file) at zero time compression.  I immediately saved the game.  Then I closed HC, started it up again and loaded the saved game I had just created.  Here is the comparison of LaunchAir events with associated AirPatrol events so obviously there is something wrong in the saving and loading as no gametime elapsed.

https://extendsclass.com/xmldiff/4271838

I should note that every LaunchAir event with an associated AirPatrol event upon loading the saved game, has the wrong AirPatrol event associated with each LaunchAir event.

2022.020 posted and I think it will fix this particular issue.  To test it out, start playing the scenario anew from the .SCq file (scenario) with the I32501 logging enabled.  Save now and again then load the saved game and keep going along as usual.  If you get the assertion errors, report as usual ge.log, saved games zipped, etc.

 

On 9/27/2022 at 11:30 PM, TonyE said:

Oh, but you are on to something.  You rightly want to do a comparison of the state of the HarpoonEvents when they are dumped into a saved game file with the HarpoonEvents as loaded.  Thanks to the XML scenario file format work (that I haven't finished), I have code that can save the HarpoonEvents in XML format.  One could then in theory take that XML file of HarpoonEvents at save and compare it to the XML file of HarpoonEvents after load.  One online comparison tool is https://extendsclass.com/xml-diff.html .  Does this approach seem like a place to start, essentially duct taping together existing functionality with a limited amount of code to aid in the troubleshooting?

For this problem it seems good to me.

Quote

HCE - 2022.020 - 2022-09-30
===========================
- Chg:32501 GE Internal testing shows the Issue is solved.  Any existing
                corruption will not be fixed by this build, the player will
                need to start the scenario fresh.
 

I'm working on a much more comprehensive revamp of the HarpoonEvent structure for 2022.021.  No promises that I won't just scrap it for what works in 2022.020 but if it does emerge, here is what is going into it so far:

1. Leverages the uuids I've been adding to key game structures.  

    a) Adds a uuid to each HarpoonEvent so it can be uniquely referenced instead of depending upon a Group Id and Unit Id combination or order of saving in the saved game file to indicate a correct reference.  This fixes I32501 again and better.

  b) References the associated unit, if there is one, by that unit's uuid instead of its Unit Id.

  c) References the associated group by its uuid instead of its Group Id in Scenario files.

2. Separates event type from the time trigger type to allow for more and easier additions of new event types