|
2008-04-26 22:58
|
crazyvasudandrivers.jpg (136,039 bytes) |
|
2008-04-26 22:58
|
crazyvasudandrivers2.jpg (125,003 bytes) |
|
2008-04-26 22:58
|
crazyvasudandrivers3.jpg (95,245 bytes) |
|
|
After a quick look in FRED: -The Sobek should be dodging the Orion. -There appears to be a bug with the fighter count. Rapier wing consists of only four fighters yet when the second wave arrives there are five, which leaves a positive count of one remaining even when all ships present in the mission other than the Orion have been destroyed. No idea what could be causing this. |
|
|
After testing with 3.6.9 final. -The Sobek should follow a waypoint course away from the Orion. It does so in 3.6.9. It doesn't not do so in this latest build. Reason presently unknown as all waypoint are present in FRED and necessary events are also present. -No fifth Myrmidon listed in directives list in 3.6.9, but it is always listed in 3.6.10. The fifth fighter does not exist in the FRED'd mission. With this in mind, while collision detection does appear to be a problem (based upon the screenshots), the larger issue is: 1. Why is the Sobek not following its waypointed course? 2. Why is the game including a non-existent fighter? |
|
|
More testing. This time with Taylor's XT0314 build. This build plays out the mission just fine, the same as 3.6.9. Whatever changes have caused the bug appear to have been entered into SVN sometime between when taylor created that build and April, when Chief checked it out. |
|
|
Yes indeed, latest trunk build shows this strange behavior. We've run into this issue when a lot of ships are in the mission. Usually just hostile ships are affected. |
|
|
Bumping to high since this looks like a regression that should be checked. Wondering if this is a combination of bugs as well, collision as well as waypoint ignoring. |
|
|
The Orion seems to like being non-collision violated by both the Sobek and the Hatshepsut "shovel" as both will clip right through it. |
|
|
Someone just needs to go through and narrow down when the bug was introduced since there is over 2 months worth of commits between the Xt0314 build and rev 4608. I gave the logs a quick look, but there are there has got to be nearly 130 commits between those builds which makes it a pain to go though. |
|
|
When exactly in the mission was this supposed to happen? I was played the 'The Sting' mission on easy difficulty level for four times and couldn't see anything like this happening using r4870. If possible can some one who sees this effect make a mission where it repeatedly appears? |
|
|
Guys, this was marked as a need to fix for 3.6.10, but unless we get some better back and forth going I have no problem bumping it back down in priority, seeing as not many people are going to be affected by it. If anyone concerned about getting it fixed helped narrow down where it was introduced (there's plenty of builds between those dates to try out) I'm sure it would help too. |
|
|
Taking out of the fix for next release filter. |
|
|
I'm not sure if this is the same issue I encountered... but... I know a way how I was able to pretty much always break collision detection: Put a waypoint somewhere, then create an every-time (true) SEXP node and move the waypoint.. For example make the SEXP keep the waypoint always 500 meters on the z-axis behind a certain ship or something. From what I know it doesn't even matter if that ship is moving.. as soon as a waypoint is moved on regular intervals collision detection is broken for some reason. From what I could see it seems like colission detection is "offset" somehow, meaning one can fly through the front part of a ship but collides in empty space in the rear and stuff like that. I don't really see a pattern anywhere, but this is what caused it for Saga. We tried to have a repositioned waypoint in some mission and later solved the mission differently, but I was able to narrow down the issue with the collision detection to this. So maybe this will give someone a starting point. if this bug is caused not just by this but by other stuff as well, this should at least provide a fool proof way to reproduce it. |
|
|
I don't know if the waypoints are driven by the autopilot/AI (my guess is that they are), but line 1061 in autopilot.cpp is assuming that the wing number and ship number are the same. This: ai_goal *aigp = &Wings[i].ai_goals[j]; should probably be this: ai_goal *aigp = &Wings[Ships[i].wingnum].ai_goals[j]; (cheers kara) EDIT: Sorry, wrong issue! |
|
|
Well since this already got bumped I'm going to tweak it. Putting it back to normal and unassigned, since with KK's info someone might be able to figure out what's going on now. |
|
|
Still prevalent? Could very well be fixed, even through new mediavps |
|
|
Normally if someone abandons a bug, I'd say close it, but in this case I'd still prefer a coder took one more peek at it based on Keldor's last statements about it. |
|
|
|
|
|
Confirmed with the attached mission based on Keldor's suggestion. |
|
|
Sorry, I'm not sure anything in that mission jumped out at me as "wrong". I see the Orion and the Poseidon, and a few craft defending the Orion by flying around it. Could you highlight what is being reproduced by this mission? |
|
|
Oh, sorry... Try to collide with the Orion. Given this is a bug about collision, I had assume people would try to collide with things. |
|
|
Closing this one. If anyone else finds collision problems, please open a new issue report. |
- Anonymous
