• All submissions to this site are governed by Second Life Project Contribution Agreement. By submitting patches and other information using this site, you acknowledge that you have read, understood, and agreed to those terms.
Issue Details (XML | Word | Printable)

Key: SVC-1239
Type: Bug Bug
Status: Closed Closed
Resolution: Fixed
Priority: Major Major
Assignee: Simon Linden
Reporter: Creem Pye
Votes: 22
Watchers: 10
Operations

If you were logged in you would be able to see more operations.
2. Second Life Service - SVC

Avatar sitting on a prim unable to have position manipulated with llSetLinkPrimitiveParams

Created: 23/Jan/08 09:01 AM   Updated: 06/Jun/08 02:12 AM
Return to search
Component/s: Scripts
Affects Version/s: Havok4 Beta
Fix Version/s: Havok4 Beta

Environment: Linux with 1.18.5.3 viewer, MacOS with 1.18.3.70780 viewer, Windows 1.18.5 (3) viewer
Issue Links:
Duplicate
 
Relates

Linden Lab Issue ID: DEV-9386
Linden Lab Internal Branch: havok4-3


 Description  « Hide
When an avatar sits on an object with more than 1 prim, a script on the object he's sitting on is unable to change his position using llSetLinkPrimitiveParam's PRIM_POSITION control. This only appears to be the case with Havok4, as Havok1 seems perform the operation correctly.

To duplicate:
1) Rez two boxes and link them together.
2) Have your avatar sit on the new linkset
3) Create and run the following script:
default
{
state_entry()

{ llSay(0, llGetLinkName(3)); // verify that link 3 is indeed the avatar llSetLinkPrimitiveParams(3,[PRIM_POSITION,<0,0,5>]); }

}

In my tests, the avatar's local position will not move to <0,0,5> as expected when the linkset contains more than 1 non-avatar prims. This bug appears to be related to SVC-750, but applies only to Havok4. Interestingly enough, setting PRIM_ROTATION on avatars works in all tested cases, so the problem is purely about avatar position-setting in multiple-prim objects.



 All   Comments   Change History      Sort Order: Ascending order - Click to sort in descending order
Creem Pye added a comment - 24/Jan/08 05:07 PM
This issue is approximately the same as SVC-750, except that this issue only occurs in Havok4 with PRIM_POSITION, in objects with >1 prims.

Rifkin Habsburg added a comment - 24/Jan/08 07:48 PM
Should we just re-open SVC-750? I've verified it breaks my En Garde game, and will probably break all the content of all the people who complained before.

Harleen Gretzky added a comment - 24/Jan/08 07:58 PM
This only applies to Havok 4, SVC-750 applied to Havok 1.

Rifkin Habsburg added a comment - 24/Jan/08 08:12 PM
Yes, I meant we update svc-750 to read "affects version havok4". This seems like a regression of the same bug.

Creem Pye added a comment - 24/Jan/08 08:14 PM
Admittedly, I don't know the proper protocol for closing one ticket in favor of another. I won't be offended if somebody closes this bug in favor of reopening SVC750, although this one seems slightly different in nature. =)

Lex Neva added a comment - 24/Jan/08 08:24 PM
from the way I've seen Linden devs talk in issue comments, I think they would prefer that a new issue be opened even in the case of a regression of an old issue that was previously resolved. I think we should leave this one as it is, with the "relates to VWR-750" link.

Valentine Janus added a comment - 25/Jan/08 04:43 AM
The idea that we're sort of repeating a critical bug from the past annoys me. But response to that one was pretty quick and perhaps the earlier experience will help. I look forward to En Garde! returning to my stadium soon . . . and do not look forward to finding what else breaks. : (

serpy palen added a comment - 25/Jan/08 09:32 AM - edited
Very bad today my favorit land switch to Havok 4 and i confirm, the llSetLinkPrimitiveParams are silently ignored on a multiprime object ((((
Also added Windows viewer in the main descrption.

Lex Neva added a comment - 25/Jan/08 09:36 AM
Are you saying that it's ignored even when the target of the llSetLinkPrimitiveParams() call is not an avatar? That'd mean this is a bigger bug than we thought.

serpy palen added a comment - 25/Jan/08 10:21 AM
Sorry i never be detailed in my previous comment.
For me is ignored when the target of call is an avatar sitting on a multiprime object, thus becoming a child of it .
Is a big bug enough for me anyways

dibbs Dovgal added a comment - 27/Jan/08 10:02 PM
I agree this is a critical bug for many of my products. I look forward to the speedy resolution of this, or the delay of Havok on the main grid until it is resolved.

Boss Spectre added a comment - 31/Jan/08 02:31 PM - edited
This appears to be a different bug. In my testing on Beta, I have also seen that PRIM_ROTATION works on avatars, but PRIM_POSITION fails. I have also noticed that in Havok4 the small internal avatar-size-based "offset" in sit targets appears to be ZERO_VECTOR now, this will affect nearly all content with a sit target, which will sit slightly too high in Havok4.

Simon Linden added a comment - 01/Feb/08 11:53 AM
This fix was done 2/1/2008 - it won't be in the next immediate Havok4 server version we push (that's already in QA) but should be in the following one.

Owen Oyen added a comment - 13/Feb/08 04:39 PM
Simon, an Issue I created about this (1273) was resolved into this Issue as a duplicate. As of today, Feb 13, I have observed that avatars now seem to be moving properly among sit positions on boats as they change tacks and heel. A more expert boat builder more familiar with this matter is following up my observation and I will let you know – here – if she states the fix is not correct.

Otherwise, thanks for the hard work.


Creem Pye added a comment - 13/Feb/08 04:59 PM
Just logged into the beta grid, and confirmed that this bug has been fixed in H4 beta 1.19.0.79818.

Jennifer Brennon added a comment - 22/Feb/08 04:24 PM - edited
Seems that some progress has been made on this issue. However, I wouldn't call this fixed just yet. I have spent most of the day going over my scripts on both the main grid as well as the beta test grid and I can assure you the problem continues. While it is true that the avatars position can once again be changed by llSetLinkPrimitiveParams, this only works for a time. Upon Rezzing a fresh copy of a device from inventory preloaded with scripts that use this function, the device works functions well. Once you begin editing anything in the linked set, this function becomes in jeopardy of no longer functioning. I've found that the more prims in the linked set, the more probable this is to happen. Where I had been using this function in a linked set of 7 objects I could go days without seeing problems, where as in a linked set of say 55 to 60, this fails after just a few edits. The function looses more than the ability to set the relative position of avatars on the prim, including the ability to move the prim itself and/or rescale the prim in the same command. Reading threw the related reported issues, there was some mention given to the possibility that reorganizing the commands may help. After a number of tests, I was not able to improve the reliability of this function. The only way I have been able to restore the function to working status is to take the object back into inventory and re-rez the linked set. As long as no edits are made to the system, the function seems to work without fail. Scripts compiled under both LSL and Mono yield the same results. Scripts under LSL will reset automatically (given the command to reinitialize via on "on_rez" event while all those under Mono all need to be restarted by the tools menu.
If any Linden would like a personal demonstration of this function still failing, I would be more than happy to demonstrate. If anyone has any other information on this issue, I would be very interested in hearing more.

Tested On: Second Life 1.18.6 (77968) Jan 29 2008 20:06:25 (Second Life Release)
Tested At: 255343.2, 255097.7, 701.2 in Sandbox Wanderton MONO located at sim3000.aditi.lindenlab.com (8.2.33.226:13005)
Mono 1.19.0.78407


Jennifer Brennon added a comment - 22/Feb/08 04:31 PM
Please see above Message....

Jennifer Brennon added a comment - 23/Feb/08 03:14 PM
Forgive my post from earlier. I was in error. The sim I tested on was not the updated server. Upon moving to a updated server, I have verified that the bug is indeed gone. Many thanks and as always my appreciation for the hard work from Linden Labs to resolve these issues.