Lost missiles

[Tony]

Just notice air to air missile problem in which after engaging other aircraft with missile remaining missiles would disappear from remaining AC magazines.

Example 4 plane group each with 4BVR and 4WVR missile engage 4 enemy AC. After firing on the enemy AC the total of fired missiles and remaining missiles would vary. Each red group contained 16 BVR missiles. The F15 would fire eight but only 7 or less BVR missile would remain sometime all remaining BVR missiles would simply disappear. They would fire 8BVR missiles and upon inspection of attacking group in unit window, one or more extra missiles would be gone leaving 7 or less. Cntr Alt S show no other than the missiles originally fired existed anymore. It was as if they had been used.

I am including a small scenario to test with. In the scenario the two blue bases will launch their AC against the Red base. Upon launch simply put Blue AC at highest altitude with radar on and attack red base to get attention from formation CAP. The red base has eight different type fighter groups launched in formation.

Expected behavior Red AC engaging with BVR missiles. Each red group had 16 missiles to use on blue AC. Blue player AC are to be used a bait and not to attack as they are to be bait.

Behavior was Red AC groups would intercept and engage blue group firing an initial missile volley. Upon reviewing in unit window red AC groups did appear to have expended more missile than on group and unit window even with Cntr Alt S on. The extra missiles were not found on either window or on the plane and seemed to have disappeared. The was not the case of the Blue player controlled AC if it fired.

Tested with Export.dll,s present and not.

 

Pacmissile test.scq

[Tony]

Forgot to give version and DB used. Was tested with both Version 2025.025 and 2025.024 and using HCDB2. Testing included removing all Export.dll,s while testing to eliminate possible causes as much as possible.

[eeustice]

Issue Verified.

WestPac Battleset

Ran the test scenario and received the same results each time.

In my run of the scenario Red air-group ZAA had 33 Air-to-Air weapons. After launching 8 missiles which Dont's AI Tool lbox verified there were only 17 Air-to-Air weapons left. Missile group ZCM shows 8 missiles in the AI Tool box.

I ran the test multiple  with GE2025.025 and with Export.dll's active.

If any additional info is needed, please let us know.

 

[donaldseadog]

I think it's simulating shody launch rails😁

I have a couple of test ideas and should look at it tonight. I've been in the bush for a couple of days but did a quick run and saw the same thing happen.

It is the AI side that has the problem, initial number minus number fired is greater than number left. 

The only add I have at moment is that store consumed procedure tallys with the number of missile fired and in the attacking missile group.

 

[donaldseadog]

I've done a small revision of the test scenario by deleting one blue base and adding some blue planes of the same type as I set to patrol in the red base.

The scenario and a few saved games are in the zip attached.

The save games are (if I remember correctly)

1/ playing blue I send up two patrolilng ac and flys towards the red base, they are intercepted by the red formation patrol identical AC with identical loadout who fire 4 LR missile at me but have only 2 out of 8 left. I attack and AI suggests I fire 6 missile, I fire 6 and have 2 left.

2/ a little later the red intercepting plains 'loose' their remaining 2 missile but none have been fired.

3/ similar but I send up a single patrolling AC. 

I'll simplify further and try for a game save a second before red fires, but what seems to be happening is that AI is deciding to fire X number of missiles, X is deleted from the planes available missiles but less than X are actually fired.

pacMissTest-A.zip

[donaldseadog]

This might be getting close

save game ...B1 should be 1 or 2 secs before the red group (2 ac unit) fires on blue group (single ac unit).

I believe AI wants to fire 4 missiles but only fires 2, 4 out of 8 are remaining.

save game ...B1.0002 is shortly after and in one or two seconds the red group expends all LR missile but none are fired.

Using DB HCDB2-170909, westpac BS, GE vers 2025.025

pacMissTest-B1.zip

[TonyE]

From pacMissTest-B1.hpq

Combat5.c around line 1050 number of rounds should be 4, appears to be 1024 (but there is a funky 1-byte offset for # of attacks)...  This section of code creates the AttackEvent.

Effects.c around line 559 number of rounds is the correct 4.

Effects.c around line 929 changes number of rounds to 2.

 

Yup, that's the problem, ProcessAttackEvent changes the number of rounds to fire from 4 to 2 but the ammo was previously removed from the mount in AllocateRounds.

Issue confirmed.  This can be can be fixed by overriding ProcessAttackEvent in an ExportDLL.  This isn't exposed at a very granular level, it is at ProcessHarpoonEvents aka dllexEffect18 meaning all Harpoon Event processing needs to be overrideen.

 

[TonyE]

Puzzling thru this with eeustice on Discord...

					case 18: /* event processor */
						if (dllexEffect18() == FALSE)
							ProcessEvents();
						break;

The ExportDLLs are called before the built-in ProcessEvents logic.  To avoid re-creating all of the harpoon event-processing logic (which is a ton), an ExportDLL can cheat by cherry-picking the events it wants to process then returning False so that the built-in ProcessEvents is still called.  The negative side effect is that the order ExportDLLs are called is not nicely sorted by filename (See c++ - How can I get FindFirstFile to sort files - Stack Overflow for a nice explanation) so using this approach is somewhat risky but it would still be a valid approach.