Platform Behavior bug (gliding over platforms)

Bugs will be moved here once resolved.

Post » Thu Jan 10, 2013 8:52 pm

I'm happy for the new release, but I don't see any changes for the platform behaviour... have you received my capx file with the bugs? Maybe you'll fix in the next release? I know you have to do a lot of cool things and many upgrades, but this issue there wasn't before latest releases...
so if you could return in the original status.
B
16
S
4
Posts: 35
Reputation: 2,292

Post » Tue Feb 05, 2013 10:20 pm

@Ashley - just wanted to mention this is in r117 (which i know is a beta ). it only seems to happen at higher ground speeds in my project (player's x is set every tick to remain stationary while the platforms travel at the player)

edit: it seems that it happens when the corner of the player connects with the corner of the platform, almost like a falling through the map glitch/bug.wretchedshark2013-02-05 22:35:10
B
14
S
6
G
2
Posts: 136
Reputation: 3,210

Post » Wed Feb 06, 2013 7:57 am

I have this happen to me as well. I am using tiled blocked tiles, that could be the issue. I am using r117.

Example of error here.

Properties of the object here.

I may be using it wrong, but I find it extremely useful for spreading tiles out easily.Helladen2013-02-06 08:00:42
Image
B
20
S
4
Posts: 105
Reputation: 2,315

Post » Thu Feb 07, 2013 3:41 pm

We cannot fix bugs from screenshots: we need .capx projects made from scratch. Can you provide this?
Scirra Founder
B
359
S
214
G
72
Posts: 22,952
Reputation: 178,630

Post » Fri Feb 08, 2013 8:45 am

@Ashley the problem is the collision is getting stuck with the edges of the solid tiles. Gravity should cause any collisions to have friction, but not make you clip to another object. For example, if I have a diagonally blocked sprite (something that is not blocked completely on the same X value all the way from top to bottom), it will get stuck on it. The greater the difference of the angle of the block, the more it will clip.

I've read you say to use movement with a 32x32 square, but it seems like a bug to me. Developers have to do a hack around it. I personally do not want to design my game around a 32x32 block square. It defeats the purpose of collisions in the first place.Helladen2013-02-08 16:03:55
Image
B
20
S
4
Posts: 105
Reputation: 2,315

Post » Fri Feb 08, 2013 10:12 am

@Helladen That's not a bug. You need to use a separate collision mask object (usually a rectangle) and use that as your player instead. The sprites are purely cosmetic.

@Ashley It's very difficult to make a .capx to reproduce this but there is a bug with the platform behavior and corners. Instead of sliding along the top like the OP was about, the "on landed" condition isn't triggered. This is especially bad for platform characters with landing animations that affect the x velocity as the whole thing is skipped.
Image
B
225
S
27
G
13
Posts: 1,774
Reputation: 18,024

Post » Fri Feb 08, 2013 12:23 pm

@Tokinsom I suppose so. I just figured blocking would be better if it were done on the cosmetic sprite rather than an invisible one. Never mind about it, then.
Image
B
20
S
4
Posts: 105
Reputation: 2,315

Post » Fri Feb 08, 2013 2:08 pm

@Helladen - we only support the Platform behavior when used with a bounding box collision mask, so we don't fix bugs unless they occur with that. Using a rectangle collision for Platform with a cosmetic sprite on top is a totally standard and far more reliable way of designing platform games.

@Tokinsom - you're the only one who's provided an actual .capx reproducing the issue so far, and the link in the OP is now a 404. I can't work on this any further unless I have a .capx!
Scirra Founder
B
359
S
214
G
72
Posts: 22,952
Reputation: 178,630

Post » Fri Feb 08, 2013 6:21 pm

@Ashley - The bug was no longer reproducible in that .capx so I took it down. Here is a new one showing 2 ways to reproduce it with slightly different results. The 2nd is kind of tricky, though. Thanks for looking into it still.Tokinsom2013-02-08 18:29:05
Image
B
225
S
27
G
13
Posts: 1,774
Reputation: 18,024

Post » Mon Feb 11, 2013 7:52 pm

OK, I think I *might* have finally got to the bottom of this. It seems on rare occasions, if you pixel-perfectly jump sideways on to a platform, sometimes the platform behavior would accidentally detect that as an upward slope that you're walking up. By tweaking the handling of slope detection, I think I have fixed it. You can try the new platform runtime.js here and let me know if this resolves it for you:

https://dl.dropbox.com/u/15217362/platformruntimer119.zip

I was hoping to do a stable release next, but I would really prefer this to have thorough testing in beta before putting the fix in a stable update. We've had small tweaks like this break the platform movement for some people before, because it's really hard to get right for all rotating gravity/jumpthrus/etc. that people use. So even if this works I might *not* release it in the next (stable) update, but put it in the first beta after that. (In the mean time you can use the runtime.js above.)

Let me know if this works for you, and especially if you notice anything working at all differently with slope handling.
Scirra Founder
B
359
S
214
G
72
Posts: 22,952
Reputation: 178,630

PreviousNext

Return to Closed bugs

Who is online

Users browsing this forum: No registered users and 2 guests