Anyone else still need a SUR generator?

Forsaken wrote:
ST, no offense to your abilities or anything… but are you absolutely sure it isn’t just a SUR issue?

No offence taken at all, I need all the help I can get to find the cause of this problem. If it’s a method problem then I need to find it.

I’ve been struggling with surs for a long time, and many work right away when I use the Sur Splicer method. Others crash right away on launch and are easy to remake. That can be strange, as remaking the same sur from the same parts often works!

I don’t have the problem others have with some parts of a sur not registering hits, this I think is naming problems that I solve by careful naming during the process.

Other methods and also some of my efforts with Sur Splicer run for a while but just refuse to run clean for a long period no matter how many times I optimise the ship groups and remake them.

By optimise I mean I remove thin bits and booms from ships to ensure only the big chunky sections remain for sur building.

So I have been doubting myself, and this is why I have asked others to test the same surs, and Quarks made a new set for me which also failed in the same way mine did - so either we both have bad environments or there is a problem with sur generation.

I have used two PCs for sur building and testing, one is my laptop and the other is a desktop. Same results, a good sur is good on both and a bad one is bad on both.

This particular problem is odd because it takes a long time for a sur to fail, the FLServer.exe crashes. I test them one at a time, there are only these test ships used by NPCs (DevilRay and Manta) and a base ship that kills them. I have some other Nephilim ships that I have made surs for (Moray, Lamprey, TripleSkate) and those fail in the same way, so I am just using these two for testing now.

When I put vanilla surs on the test ships too then the game runs for 12 hours plus with no crashes.

If we can isolate the cause of the problem between us then I’m hoping LS might melt a little and fix it for us, or another expert might generate a new tool or exporter.

But without isolating the precise cause of the problems, if it was me I would not be interested either.

LS hinted at internal lines in the sur, and Bejaymac many years ago fiound the same problem when using the exporter v1.1 (and v1.2 was worse). This is connections between lower-order vertices, which causes the sur shape to be concave and not convex throughout as it should be.

So - if a couple of others would please test the same two surs that I attached and confirm they are clean or crash, and perhaps if you would also test the new surs that Quarks made for me, we can take it from there.

I have used the manual method with Sur Splicer with the best success, other methods are faulty including Sur Builder, which is so close that it is a sad thing.

What is my ultimate goal? A guaranteed method or utility that everyone can use to make good surs quickly without the sweat and tears that I have to go through each time.

All testing and confirmation of good/bad is very welcome, fixing these difficulties will benefit all modders.

Sushi’s 1.5 hours looks promising that the surs are in fact good and it might be my PCs and Quarks’ PC setups.

Sushi, many thanks.

There is another angle I am trying to isolate if you have the time please:-

My crashes happen by a single NPC faction attacking the base ship, and adoxa has narrowed the crash to a possible munition collision.

This doesn’t sound completely convincing to me.

My theory is hull to hull collision, maybe when an NPC with the sur crashes into the base ship hull. It does not crash when I use the same base ship and use any vanilla sur for the NPC faction with the same DevilRay or Manta ship.

Could you please use one faction with the DevilRay or Manta, one at a time, and make them hostile to your base, let them attack it and be killed by the base please?

I really would appreciate you testing this method, it is such an intermittent problem - if it fails for you in this way then we can be almost certain it is a ship to ship collision problem that happens so rarely.

By the way - I am assuming you are not using any error-preventer hack?

Thank you.

Thanks adoxa, will do asap.

Sushi - I ran it your way last night, I set the test-sur NPCs wth the DevilRay to neutral to the base so it would not attack them, and had my Lib Navy (vanilla ships and surs) attacking them.

Guess what - it ran clean for 2 hours! I didn’t have time to test longer, but normally it will fail within 10-20 minutes every time when they are attacking the base ship.

So I think to be conclusive I would like to see what happens on your machine when you set them up (one faction only) to attack the base ship/station so it kills them.

A base ship with a custom sur would be more accurate, as that is what I’m using, but mine is huge - 1800 metres long, so if it is a hull collision it may still not happen on yours if your base ship is small.

I will set up adoxa’s traps to see what collisions we catch in any case.

Thanks again.

Quarks wrote:
Since I have this problem, too…I just jeplaced the li_fighter’s model with Startrader’s ship and edited the loadouts, then edited the manhattan pop zone, nothing else. So usually there should not be any problem.

Thanks Quarks.

So you get the same problem just by having the two factions kill each other?

Not by attacking a base, right?

So my hull to hull collision theory is wrong, I will do more extended combat testing too.

Many thanks guys, it will be great to identify the cause of yet another rare crash problem.

adoxa wrote:
Tried Quarks mod. Crashed reasonably quickly (the last packet sent was FLPACKET_SERVER_DAMAGEOBJECT, with different parameters to all other such packets). Applied my two plugins, which caught two more (another damage object; the other one’s time didn’t correspond exactly, but FLPACKET_SERVER_DESTROYOBJECT was likely), but none after another hour. Restarted, but still none after more than an hour. Restarted, and still none after almost an hour. So StarTrader, don’t bother running Packets; GetCollisions probably isn’t worth it, either; GetRoot will prevent the crash, though. Once I finish off tweaks to Console, I’ll get to work on the MilkShape sur exporter.

adoxa - Well I got 3 crash logs, the common factor is use object 50 to the DevilRay, which is where the NPC’s Countermeasure Dropper is mounted - indicating a missile is onto him.

I am wondering that the NPC is hit by the missile and then that hit collision calc is going wrong? I attach the 3 crash logs below anyway, hope it will give you something to get us on the track.

I shortened the GetCollision-bad-2.txt in the middle to fit it into the 1MB limit but all of the other files are complete. Thanks adoxa

Sushi - if your test ships are not being chased by missiles this might also explain why your tests aren’t failing.

Hi Mirkha.

Do you mean when a missile launcher is inside the hitbox (the hitbox is bigger than the ship or above the ship surface)?

I understand that would be a problem.

But in this case the problem is with a small fighter sur, and we are trying to isolate if the FL Server crash is happening because of a missile hitting the fighter.

Quarks wrote:
If you are interested, this is the mod I was using, Junkers and Hogosha using the bad ships and fighting in front of Manhattan.

Quarks, this fits - your DevilRay loadout included missiles! Perhaps Sushi’s loadout did not have missiles. I’ll go isolate!

The base ship has a lot of cannons and these are destroying most of the test fighters.

I will replace all of the cannons with missile launchers, to see if I can aggravate the problem. Then I will replace all missile launchers with cannons, to try to isolate it to missile hit detections as the cause.

If it is, we then have to understand what part of a sur is needed to detect missile hits correctly. I have a feeling this is the purpose of the Type 5 sur parts but I do not understand how to make those fully yet.

Maybe together we can work it out, I am farily good at inventing ways to isolate problems, with help from LS and adoxa and some others to test maybe we can crack it.

Then when it is solved and we can produce good surs, I can drop FL altogether, install Linux on my laptop and PC, and go chew carrots full time.

StarTrader wrote:

adoxa - Well I got 3 crash logs, the common factor is use object 50 to the DevilRay, which is where the NPC’s Countermeasure Dropper is mounted - indicating a missile is onto him.

Adoxa - Sorry, that was incorrect, hpid 50 is the shield batteries.

But you got your own results anyway.

In the meantime, I changed the missiles to cannons and it still crashed.

I will go through the rest of the isolation anyway in case it gives us any more.