Carl Spiby
New member
- Feb 28, 2012
- 1,756
- 0
Ball lost in the scoop, probably fell through the geometry. Why is this engine so hopeless when it comes to collision detection?
http://db.tt/yYwyIAlC
http://db.tt/yYwyIAlC
-
Welcome Back to Digital Pinball Fans - please read this first
Bug Sigh, why do I bother?
- Thread starter Carl Spiby
- Start date
superballs
Active member
- Apr 12, 2012
- 2,654
- 2
It's definitely a big flaw in the engine that needs to be address
K
Korven
Guest
Just lost a ball during Atlantis. Ended up just like the screenshot you posted. Played this table a lot lately and nudging or call attendant always worked up until now...
Edit: http://digitalpinballfans.com/showthread.php?p=36235 looks like this is the bug I got, doing well on Atlantis.
Edit: http://digitalpinballfans.com/showthread.php?p=36235 looks like this is the bug I got, doing well on Atlantis.
Last edited:
Pete
New member
- Jul 16, 2012
- 564
- 1
I think farsight really needs to stop releasing tables so fast and perfect the awesome one's that are already out. I am a die hard fan of this game I (i love it! best game ever) and will put up with these bugs but what about the new comers that are like "what the hell just happened?" and quickly give up on it as a quality game. I hear allot of the guys in the forum say who cares about the bugs and quality of the game just release more and more tables as quickly as possible, and this is really making the game suffer. Almost all of the most popular tables have been released already and those are the one's that should really be given great care. How about we make a compromise and have a kickstarter to hire more employee's so they can actually handle the work load, I'll donate to that.
N
netizen
Guest
http://digitalpinballfans.com/showthread.php/5254-Hey-Farsight-Take-a-Break!
there's a whole thread for that line of thought.
Carl Spiby
New member
- Feb 28, 2012
- 1,756
- 0
- Thread starter
- #7
They need a constant revenue stream, fixing past tables doesn't pay the bills.
However, one can hope, that when they release the PC version that can act as a buffer whilst they spend time fixing the tables they have already released.
However, one can hope, that when they release the PC version that can act as a buffer whilst they spend time fixing the tables they have already released.
superballs
Active member
- Apr 12, 2012
- 2,654
- 2
I don't think that they (we i guess) "don't care" about the bugs, but we want more of a balance between releasing and fixing.
There are a lot of issues that have been fixed over time, though a lot of glaring issues have not been touched and they are being used as representative to "bugs never being fixed".
Both trains of thought are a bit mental when taken to the extreme. Bugs are being fixed, especially when well documented and reproducible. Other bugs are certainly being looked at but attempted fixes may have introduced new or worse bugs into the fray and are put on the back burner for things that they can fix a bit quicker and easier. A lot of the bugs seem to affect the engine universally, especially collision between the ball and the table/toys/rails. The "through the skull" bug on T2 could also be called the "through the left saucer when a ball is locked" bug on black hole.
Art fixes would be nice as well occasionally as at least it's unlikely (in my imagination anyway) to introduce other issues. But art really doesn't affect gameplay that much, unless of course it's something important missing like an insert (or the inability to tell if an insert is lit or not). Speaking of which, i need to check out Gorgar again since i think they did fix that issue finally.
xAzatothx
New member
- Sep 22, 2012
- 824
- 0
most of these bugs they can't reproduce. they don't play test these tables long enough or in enough detail. Many of these bugs occur after lengthy play sessions.
they need a test rig they can leave running with unlimited balls detecting errors with logging........
they need a test rig they can leave running with unlimited balls detecting errors with logging........
superballs
Active member
- Apr 12, 2012
- 2,654
- 2
They could pay me minimum wage to play their tables all day long and live in Cali.
I would leave a salaried job to do so, as long as it comes with a hut and 3 cots. Food would be nice too, or at least some fruit bearing trees.
I would leave a salaried job to do so, as long as it comes with a hut and 3 cots. Food would be nice too, or at least some fruit bearing trees.
I believe the issue is that the engine doesn't interpolate between frames. this explains how the ball can just warp through objects when moving fast. It happens more ofte on smartphones because they tend to have frame rate hiccups when you get text messaged and when other apps phone home and stuff.
SO this is why a test under ideal conditions doesn't reveal the bug.
I think there are serious errors with the code to determine if you can or can't call an attendant. Here's how it should work.
THe location of all balls should be tracked by the engine itself at all times. If any ball is outside the playable area, it shoudl get poofed and returned to plunger.
If all balls are touching switches, then they are considered in play, and not stuck.
Captive balls are a special case. they are considered in play and ineligible for movement if they are in their designated zones. if a captive ball it out of that zone it should be returned to it.
if there is exactly one ball free floating, attendent call should take 15 seconds. after then without a switch tripped, the ball should be relocated to plunger.
If multiple balls are free, then require long enough for the game to go into ball search.
Now it coudl be that all balls are "taken" and yet the game is stuck. THis means the emulation missed an event. in this case the right thing to do is most likely to move the ball that's not in the trough that tripped the last switch to the plunger lane, if it's not a captive ball. If it happens again, or if no ball was outside of the trough, then move the last ball to enter the trough to the plunger lane. I believe this shoudl cover any case.
THere is one more case. it could be that a ball is stuck in a place it can score forever. the software should catch this. if the same switch is tripped a certain number of times in a row with no intervening switches, then this is a "beneficial malfunction" and your score is invalidated. "sorry, machine out of order" , and the game is force quit. (and yes this can happen in real life)
SO this is why a test under ideal conditions doesn't reveal the bug.
I think there are serious errors with the code to determine if you can or can't call an attendant. Here's how it should work.
THe location of all balls should be tracked by the engine itself at all times. If any ball is outside the playable area, it shoudl get poofed and returned to plunger.
If all balls are touching switches, then they are considered in play, and not stuck.
Captive balls are a special case. they are considered in play and ineligible for movement if they are in their designated zones. if a captive ball it out of that zone it should be returned to it.
if there is exactly one ball free floating, attendent call should take 15 seconds. after then without a switch tripped, the ball should be relocated to plunger.
If multiple balls are free, then require long enough for the game to go into ball search.
Now it coudl be that all balls are "taken" and yet the game is stuck. THis means the emulation missed an event. in this case the right thing to do is most likely to move the ball that's not in the trough that tripped the last switch to the plunger lane, if it's not a captive ball. If it happens again, or if no ball was outside of the trough, then move the last ball to enter the trough to the plunger lane. I believe this shoudl cover any case.
THere is one more case. it could be that a ball is stuck in a place it can score forever. the software should catch this. if the same switch is tripped a certain number of times in a row with no intervening switches, then this is a "beneficial malfunction" and your score is invalidated. "sorry, machine out of order" , and the game is force quit. (and yes this can happen in real life)
Last edited:
