Second Life Forums Archive - land_collision not registering normally

Al Bravo

Retired

Join date: 29 Jun 2004

Posts: 373

10-21-2004 02:23

Is anyone else having problems with land_collision()? Especially when used within attachments. Since the last patch (or possibly the previous one), I seem to be receiving considerably fewer (sometimes none at all) land_collision() events.

Jillian Callahan

Rotary-winged Neko Girl

Join date: 24 Jun 2004

Posts: 3,766

10-21-2004 03:00

Yes, either verly slow to fire (up to nintey seconds) or just never happens.

Al Bravo

Retired

Join date: 29 Jun 2004

Posts: 373

10-21-2004 07:15

OK, glad to see it's not just me. Unfortunately that has pretty much broken every parachute in SL. I bugged it. Hopefully LL will fix it in the next dot rev. So anybody having problems with their parachutes not closing when then hit the ground, this is the reason why... (bug it with LL)

Ardith Mifflin

Mecha Fiend

Join date: 5 Jun 2004

Posts: 1,416

10-21-2004 07:23

Do you really need a land collision event for such a thing? Couldn't the same be accomplished using a call of llGround to determine the height of the ground, and then derez the parachute when the av's position and the ground are relatively close? By throwing in llGetAgentSize, it would be just as effective without requiring collision events. Wouldn't it? Maybe I'm missing something.

Al Bravo

Retired

Join date: 29 Jun 2004

Posts: 373

10-21-2004 07:32

It could be done that way. But not really the point here. It was working for a long time and then LL made a change somewhere that broke it. And it isn't just the chutes, I have vehicles that rely on land_collision() to deploy landing gear properly.

Tiger Crossing

The Prim Maker

Join date: 18 Aug 2003

Posts: 1,560

10-21-2004 11:17

A collision detection would be better, since it's an even handler. It does nothing until the event fires. To use the ground height, you would have to be continuously polling the ground height in a time to see how far you are from it. If a parachute was already doing that to display altitude from the ground on a altimiter, then tapping into it wouldn't cause any further slowness in the script.

_____________________

~ Tiger Crossing
~ (Nonsanity)

Al Bravo

Retired

Join date: 29 Jun 2004

Posts: 373

10-21-2004 11:24

Right, but the thing is - I have literally hundreds of these in the field. land_collision() needs to be fixed.

Cubey Terra

Aircraft Builder

Join date: 6 Sep 2003

Posts: 1,725

10-21-2004 13:48

Yep. I've had problems with land_collision too. It usually doesn't detect a collision, even at high speed nose-down into the ground. Other times, it detects a collision even though it's dozens of meters above the ground. It's b0rked.

I tried to use land_collision in my latest plane, but had to drop the feature b/c it just didn't ever work.

_____________________

C U B E Y · T E R R A
planes · helicopters · blimps · balloons · skydiving · submarines
Available at Abbotts Aerodrome and XstreetSL.com

Moleculor Satyr

Fireflies!

Join date: 5 Jan 2004

Posts: 2,650

10-21-2004 16:06

Has anyone actually bug reported this recently?

Rickard Roentgen

Renaissance Punk

Join date: 4 Apr 2004

Posts: 1,869

10-21-2004 16:33

I did when I noticed it say a month ago. not since. land_collision_start still seams to work. only difference is it doesn't keep firing while the object is in contact with the ground. just when it starts being in contact with the ground.

_____________________

Carnildo Greenacre

Flight Engineer

Join date: 15 Nov 2003

Posts: 1,044

10-22-2004 00:01

I fired off several dozen clay pigeons last night set to delete themselves when a land_collision_start() event occurred. I think all of them deleted properly.

_____________________

perl -le '$_ = 1; (1 x $_) !~ /^(11+)\1+$/ && print while $_++;'

Catfart Grayson

Registered User

Join date: 16 May 2004

Posts: 264

10-22-2004 07:10

I posted a bug report on this at the begining of october I think. It start after the latest version.

_____________________

Cat

Al Bravo

Retired

Join date: 29 Jun 2004

Posts: 373

12-09-2004 05:28

This is still broke 3 revisions later. Bugs have been filed. It is not in 1.5.8. What now?

Jillian Callahan

Rotary-winged Neko Girl

Join date: 24 Jun 2004

Posts: 3,766

12-13-2004 03:36

I don't get why this keeps breaking.

Anotther thing I've noticed is that llGround() lies in some places. Notably in the sandbox sims (I've got some pics of a tests I did I'll post later if there's interest) but also in other places around SL as well.

Its so very strange that the event fails to trigger when the object is happily rolling around on the ground - meaning the sim does seem aware of the contact. Vewwy stwange.

_____________________

Carnildo Greenacre

Flight Engineer

Join date: 15 Nov 2003

Posts: 1,044

12-13-2004 23:40

How is llGround() lying? I've been getting some anomolous results in a recent mapping project, but I think they're just an artifact from my communications system.

_____________________

perl -le '$_ = 1; (1 x $_) !~ /^(11+)\1+$/ && print while $_++;'

Jillian Callahan

Rotary-winged Neko Girl

Join date: 24 Jun 2004

Posts: 3,766

12-14-2004 07:32

I was going nuts trying to understand what I'd done wrong on a robot script. On occasion it would pop up way into the air as part of its ground avoidance routine. After a bit I figured out that llGround() was giving me bad results. Here's some pics of the tests I did. I had one that placed every twenty meters on a grid over the whole sim, and that prim would change it's Z value to llGround() + 0.25 meters.
I had another that would rez prims on a 0.5 meter grid, and the prims would place themselves right where llGround() said to go.

Here's the resuts:

_____________________

Rickard Roentgen

Renaissance Punk

Join date: 4 Apr 2004

Posts: 1,869

12-14-2004 10:03

remember, llground takes an offset vector.

_____________________

Jillian Callahan

Rotary-winged Neko Girl

Join date: 24 Jun 2004

Posts: 3,766

12-14-2004 15:54

From: Rickard Roentgen

remember, llground takes an offset vector.

Right, so that's <0,0,0> to get the ground height under the prim, yes?

_____________________

Rickard Roentgen

Renaissance Punk

Join date: 4 Apr 2004

Posts: 1,869

12-14-2004 17:37

yep

_____________________

Jillian Callahan

Rotary-winged Neko Girl

Join date: 24 Jun 2004

Posts: 3,766

12-14-2004 21:07

Well, I ran the same test again today and the prims did as they should have. Either whatever the problem was is fixed, or my original script (which I could not find, so I wrote a new one) was giving llGround a nonzero offset or I had done something else equally wierd.

Either way, my point's now totally moot.

_____________________

Al Bravo

Retired

Join date: 29 Jun 2004

Posts: 373

12-15-2004 00:36

Have you bugged the llGround() issue? It sounds like it may be linked to the land_collision() problem. Thanks.

land_collision not registering normally
Al Bravo Retired Join date: 29 Jun 2004 Posts: 373	10-21-2004 02:23 Is anyone else having problems with land_collision()? Especially when used within attachments. Since the last patch (or possibly the previous one), I seem to be receiving considerably fewer (sometimes none at all) land_collision() events.
Jillian Callahan Rotary-winged Neko Girl Join date: 24 Jun 2004 Posts: 3,766	10-21-2004 03:00 Yes, either verly slow to fire (up to nintey seconds) or just never happens.
Al Bravo Retired Join date: 29 Jun 2004 Posts: 373	10-21-2004 07:15 OK, glad to see it's not just me. Unfortunately that has pretty much broken every parachute in SL. I bugged it. Hopefully LL will fix it in the next dot rev. So anybody having problems with their parachutes not closing when then hit the ground, this is the reason why... (bug it with LL)
Ardith Mifflin Mecha Fiend Join date: 5 Jun 2004 Posts: 1,416	10-21-2004 07:23 Do you really need a land collision event for such a thing? Couldn't the same be accomplished using a call of llGround to determine the height of the ground, and then derez the parachute when the av's position and the ground are relatively close? By throwing in llGetAgentSize, it would be just as effective without requiring collision events. Wouldn't it? Maybe I'm missing something.
Al Bravo Retired Join date: 29 Jun 2004 Posts: 373	10-21-2004 07:32 It could be done that way. But not really the point here. It was working for a long time and then LL made a change somewhere that broke it. And it isn't just the chutes, I have vehicles that rely on land_collision() to deploy landing gear properly.
Tiger Crossing The Prim Maker Join date: 18 Aug 2003 Posts: 1,560	10-21-2004 11:17 A collision detection would be better, since it's an even handler. It does nothing until the event fires. To use the ground height, you would have to be continuously polling the ground height in a time to see how far you are from it. If a parachute was already doing that to display altitude from the ground on a altimiter, then tapping into it wouldn't cause any further slowness in the script. _____________________ ~ Tiger Crossing ~ (Nonsanity)
Al Bravo Retired Join date: 29 Jun 2004 Posts: 373	10-21-2004 11:24 Right, but the thing is - I have literally hundreds of these in the field. land_collision() needs to be fixed.
Cubey Terra Aircraft Builder Join date: 6 Sep 2003 Posts: 1,725	10-21-2004 13:48 Yep. I've had problems with land_collision too. It usually doesn't detect a collision, even at high speed nose-down into the ground. Other times, it detects a collision even though it's dozens of meters above the ground. It's b0rked. I tried to use land_collision in my latest plane, but had to drop the feature b/c it just didn't ever work. _____________________ C U B E Y · T E R R A planes · helicopters · blimps · balloons · skydiving · submarines Available at Abbotts Aerodrome and XstreetSL.com
Moleculor Satyr Fireflies! Join date: 5 Jan 2004 Posts: 2,650	10-21-2004 16:06 Has anyone actually bug reported this recently?
Rickard Roentgen Renaissance Punk Join date: 4 Apr 2004 Posts: 1,869	10-21-2004 16:33 I did when I noticed it say a month ago. not since. land_collision_start still seams to work. only difference is it doesn't keep firing while the object is in contact with the ground. just when it starts being in contact with the ground. _____________________
Carnildo Greenacre Flight Engineer Join date: 15 Nov 2003 Posts: 1,044	10-22-2004 00:01 I fired off several dozen clay pigeons last night set to delete themselves when a land_collision_start() event occurred. I think all of them deleted properly. _____________________ perl -le '$_ = 1; (1 x $_) !~ /^(11+)\1+$/ && print while $_++;'
Catfart Grayson Registered User Join date: 16 May 2004 Posts: 264	10-22-2004 07:10 I posted a bug report on this at the begining of october I think. It start after the latest version. _____________________ Cat
Al Bravo Retired Join date: 29 Jun 2004 Posts: 373	12-09-2004 05:28 This is still broke 3 revisions later. Bugs have been filed. It is not in 1.5.8. What now?
Jillian Callahan Rotary-winged Neko Girl Join date: 24 Jun 2004 Posts: 3,766	12-13-2004 03:36 I don't get why this keeps breaking. Anotther thing I've noticed is that llGround() lies in some places. Notably in the sandbox sims (I've got some pics of a tests I did I'll post later if there's interest) but also in other places around SL as well. Its so very strange that the event fails to trigger when the object is happily rolling around on the ground - meaning the sim does seem aware of the contact. Vewwy stwange. _____________________
Carnildo Greenacre Flight Engineer Join date: 15 Nov 2003 Posts: 1,044	12-13-2004 23:40 How is llGround() lying? I've been getting some anomolous results in a recent mapping project, but I think they're just an artifact from my communications system. _____________________ perl -le '$_ = 1; (1 x $_) !~ /^(11+)\1+$/ && print while $_++;'
Jillian Callahan Rotary-winged Neko Girl Join date: 24 Jun 2004 Posts: 3,766	12-14-2004 07:32 I was going nuts trying to understand what I'd done wrong on a robot script. On occasion it would pop up way into the air as part of its ground avoidance routine. After a bit I figured out that llGround() was giving me bad results. Here's some pics of the tests I did. I had one that placed every twenty meters on a grid over the whole sim, and that prim would change it's Z value to llGround() + 0.25 meters. I had another that would rez prims on a 0.5 meter grid, and the prims would place themselves right where llGround() said to go. Here's the resuts: _____________________
Rickard Roentgen Renaissance Punk Join date: 4 Apr 2004 Posts: 1,869	12-14-2004 10:03 remember, llground takes an offset vector. _____________________
Jillian Callahan Rotary-winged Neko Girl Join date: 24 Jun 2004 Posts: 3,766	12-14-2004 15:54 From: Rickard Roentgen remember, llground takes an offset vector. Right, so that's <0,0,0> to get the ground height under the prim, yes? _____________________
Rickard Roentgen Renaissance Punk Join date: 4 Apr 2004 Posts: 1,869	12-14-2004 17:37 yep _____________________
Jillian Callahan Rotary-winged Neko Girl Join date: 24 Jun 2004 Posts: 3,766	12-14-2004 21:07 Well, I ran the same test again today and the prims did as they should have. Either whatever the problem was is fixed, or my original script (which I could not find, so I wrote a new one) was giving llGround a nonzero offset or I had done something else equally wierd. Either way, my point's now totally moot. _____________________
Al Bravo Retired Join date: 29 Jun 2004 Posts: 373	12-15-2004 00:36 Have you bugged the llGround() issue? It sounds like it may be linked to the land_collision() problem. Thanks.

Welcome to the Second Life Forums Archive

land_collision not registering normally