I went to a backup and could recreate it and sent you the .mfaWell posting this without giving us the ability to take a look to see if it's fixed, or can be fixed is not very helpful.
Sure - but that sort of requires me to be able to recreate it - and it doesn't happen every time. If I send you the whole game - you'd need hours to find it - and you wouldn't know what's supposed to happen and what not.It's not random. Can you send me the MFA?
I had thought it would be easier for you to check the fusion code than my game actually - but I'll see what I can do.
This part: Previously if I said - end the level when there are none of these 5 objects left - it would end the level. Now it won't.
This happened in 2 parts - it happened once on the Windows version that I hadn't changed - but now I can't find it again - (I had picked a random level to test at the time).
The 2nd part I could recreate - .mfa attached - this is the scenario:
- I had deleted some of the objects, but their qualifier stayed behind, although there were no more objects associated with the qualifier. In this case I just need to manually go delete the qualifiers.
- You can see the original frame in frame 1 - and then what happens on frame 2 - where the level doesn't end.
- I guess the issue then is rather if all the objects are removed - the qualifiers in events should also be removed.
When creating this .mfa I also picked up that the command: destroy if too far from the frame didn't work - I previously encountered that and just put my own code in - but there's an example of that too in this .mfa
The random vs non-random collisions not happening
Part 1 was also in that test level on Windows that I can't find again - basically the level didn't end which means either there were objects still there off the frame that hadn't been destroyed - or the test for these objects were failing - but in this case I hadn't changed a thing from when it was working (it had been tested 100s of hours).
Part 2 was random behaviour on the part of this user - which I've fixed (my error in Fusion is fixed - I don't know how to fix random behaviour...)
Part 3 I can spot that happening on my cell phone and tablet but still can't find a pattern or anything to go on - I can't definitively say it's a collision not happening or a sound not happening - I tried to video record a level and then of course it didn't happen at all... I'll keep checking.
Yes, this issue has been reported a couple of times for global events: when an object is not in the frame, the conditions/actions that reference it are removed at build time. If it's a condition, the entire event is removed (as the condition is treated as being false). It's a problem for conditions like "compare number of objects" as you would expect it to compare the expression with 0 instead of being removed. It's impossible to change this in a general case.I guess the issue then is rather if all the objects are removed - the qualifiers in events should also be removed.
We could perhaps make an exception for "compare number of objects" and automatically replace it by a "compare 2 general values" with 0 as first parameter, I can think about that. I guess we would get the same "bug" report for some other conditions though...
I think it's due to the movement you use for those objects. You oddly use a user-controlled 8-direction movement for these objects, any reason why? Apparently this option works only for movements that are not controlled by the user. If you replace the movement by a bouncing ball movement, the destroy option works.destroy if too far from the frame didn't work
The collision detection between 2 non-rotated, non-stretched objects is safe on all platforms. What might happen perhaps is if for some reason you get a frame rate drop and the object moves faster to catch up the delay and then the red box object is skipped by the object. Maybe increase the height of the red box object to avoid this. Though in theory if you replace the movement by a bouncing ball movement you can remove this object.random collision detection
I didn't know that. There's no reason - I just picked the one on top to illustrate the other problem. In my game, the objects use bouncing ball movement except for the user-controlled player. The collision objects are quite large and I have fine detection turned off.I think it's due to the movement you use for those objects. You oddly use a user-controlled 8-direction movement for these objects, any reason why? Apparently this option works only for movements that are not controlled by the user. If you replace the movement by a bouncing ball movement, the destroy option works.
I did manage to capture what was I seeing on video and then play it frame by frame:The collision detection between 2 non-rotated, non-stretched objects is safe on all platforms. What might happen perhaps is if for some reason you get a frame rate drop and the object moves faster to catch up the delay and then the red box object is skipped by the object. Maybe increase the height of the red box object to avoid this. Though in theory if you replace the movement by a bouncing ball movement you can remove this object.
- The sound is failing to play on the collision event - and although Fernando had helped me solve that for all my other games - it still happens in this one - but infrequently - although it happens on my tablet as well - so could be a different issue - the previous sound issue was isolated to my cell phone.
- The animation isn't playing because right at that moment the character moves - changes direction or something and the movement animation overrides the collision animation.
- It definitely looks like the collision event isn't happening because of this, but it probably is.
Pause / Resume Sound issue: I can solve that by ending the one sound and resuming the other sound in the last second before the frame ends. So if it's going to take a lot of time for you to look at the issue - it would probably take me an hour or two to change everything my side. Just let me know.
Info about back buffers:
When opening my games in Windows - that were written a few years ago - it puts in a default Number of Back Buffers: 2 (I think these games were written before that feature was introduced.)
This results in:
1. Jumping from full screen to window mode as part of a pause action (i.e. switch to window mode, pause game) returns a full black screen that stays black - until you press any key to end the pause - in which case the screen then displays correctly.
2. Minimising the game in window mode and then opening it again results in a partial pause - the icons I made invisible as part of the pause action are now visible - the pause message is gone - but the game is still paused.
Changing the back buffers to 1 - solved both these issues.
Thank you for the report, I'll check it.
The new runtime option in 2.5+ "Set anti-aliasing when resizing" does not work in DX11. Is this a side effect of using DX11 or is it a bug?