#Pathfinder Seizures
1 messages Β· Page 1 of 1 (latest)
Well, to have it here, pathfinding is based on ticks passed, as far as I know
Slowing the game down will delay pathfinding in real time.
That was the explicit reason I sped up the game in the test script I gave the other night
wouldn't that effect things like total distance that can be covered?
also why does it eventualy work
https://www.youtube.com/watch?v=i3aFpQyvHTc link to your video from earlier
what? I don't think you understood me. Slowing down the game gets fewer pathfinder requests done. I saw your game running at 1/10th speed
Oh yes okay
I was thinking in terms of a tick limit
- yeah that was silly to think like that now that I re-read
so in the first 25 seconds I zoom on a path that is trying to get around the landfill that has been done - note: not IS being done
so I think we can rule out that theory
Well, uh, it's after 7AM for me right now
I should be sleeping, hopefully I will be more useful tomorrow
Haha I was like, okay early morning for you π
Bruh I wish I could wake up at 7 reliably
yeah me too
I support users that are two hours ahead of me some times :/
I'll catch you after you have a nap I'll keep bashing my head chasing it down in the mean time
π
Okay!
@bright inlet
I rewrote the go_to_position conditions which does deal with re-calling the action which in turn calls the path finder.. In hindsight to doing so, calling the action without returning only for it to continue doing the condition function wasn't great at all.. I was getting pretty sure this was the issue.
To give you an idea.. the go_to_position condition gone from 9 levels to at worst... 4 levels of indentation.
..... I was happy it worked first go and I didn't make any mistakes.
back to the seizure issue!
..... it's STILL happening.
Not only that I am acutely convinced it cannot be the code..
I have poured over the code even more since we last spoke but the main reason why I think this is that there is only one entry point / function calling the path finder and a break point is NOT triggered when I place one or one hundred in there.
I do not know what else to do. π
@meager dove update for you too ^
Pathfinder Seizures
I recorded another video creating paths using the remote..
https://youtu.be/fisKp7d1mx8
1:40 trying to path to the large island took quite a long time to get to attempt 5...
Let me know what you think.. I took away a few thoughts from these test but I don't want to pollute yours.
If you want to know what attempts do what..
https://github.com/ILLISIS/Constructron-Continued/blob/main/script/pathfinder.lua#L140
This looks like the pathfinder itself is struggling to find a path but also doesnβt want to abort and say no path found?
^^^^^^^^^^^^^^^
Have you tried to test the old pathfinder with a remote to see if it behaves the same?
I can..... but it never changed the PRM to 0 to walk over 1 tile wide gaps
so it's not like for like
Ok
The PRM isn't being changed in the seizures yet because of...
this
One thing I saw happen tho that might not be wanted is if you have 2 pathfinder requests for the same ctron running at the same time the one that finishes last will overwrite the result of the other one
I seen that too, not ideal I agree but that's a product of using the remote
jobs wait for event before doing anything
I made my conclusion: https://forums.factorio.com/viewtopic.php?f=7&t=104660
@bright inlet seems like you were right about the landfill
I called it more as a seizure because it looked seizure-inducing. I hope I don't offend anyone.
would you like me to retract that? π
I thought it was both seizure inducing and having a seizure itself heh
If you would, "resembled flashes that cause seizures"? or at least "no offense intended" π¦
I just removed your name
I'm an aussie, how I talk will offend someone at some point lol
On second thought I don't want to put anyone in that position where they need to refer to it so I'll change it completely