secondlife / jira-archive

3 stars 0 forks source link

[BUG-234487] ACME Physics bug - Wile-e-prim does not realise there is nothing underneath it. #11404

Open sl-service-account opened 1 year ago

sl-service-account commented 1 year ago

What just happened?

https://i.gyazo.com/042afc25e86f2f216e94c60dd1851590.jpg The following video shows the issue:

https://i.gyazo.com/76bceff37288d4236c4801512360179e.mp4

Setup: A small physical object, is placed at rest on a hollowed cube. The hollow size has been previously set to be small enough that the physics engine "convexifies" the cube, sealing the hole. small object: Size <0.03266, 0.03266, 0.03266> set to physical. larger object: Size <1.49838, 1.49838, 1.49838> Hollow: 4.996 With the small object placed over the hole, it will not fall because of the convex physics. This is entirely expected.

We now increase the hollow size. to 9.9996, the "convexification" on the server is no longer applied, and the correct physics shape is now active; however, the physical prim that was at rest remains entirely unaware of the fact until, when given a suitable event to remind it that gravity is a thing, it will plummet.

What were you doing when it happened?

testing FS

What were you expecting to happen instead?

Any object at rest should be awakened when a change to the object it rest upon occurs.

Other information

https://gyazo.com/8039b5b709b7ff8504d349d68ecdba6e

Original Jira Fields | Field | Value | | ------------- | ------------- | | Issue | BUG-234487 | | Summary | ACME Physics bug - Wile-e-prim does not realise there is nothing underneath it. | | Type | Bug | | Priority | Unset | | Status | Accepted | | Resolution | Triaged | | Reporter | Beq Janus (beq.janus) | | Created at | 2023-10-03T22:10:05Z | | Updated at | 2023-10-26T17:40:44Z | ``` { 'Build Id': 'unset', 'Business Unit': ['Platform'], 'Date of First Response': '2023-10-06T19:00:53.944-0500', "Is there anything you'd like to add?": 'https://gyazo.com/8039b5b709b7ff8504d349d68ecdba6e', 'ReOpened Count': 0.0, 'Severity': 'Unset', 'System': 'SL Simulator', 'Target Viewer Version': 'viewer-development', 'What just happened?': '!https://i.gyazo.com/042afc25e86f2f216e94c60dd1851590.jpg!\r\nThe following video shows the issue:\r\n\r\nhttps://i.gyazo.com/76bceff37288d4236c4801512360179e.mp4\r\n\r\nSetup: A small physical object, is placed at rest on a hollowed cube. The hollow size has been previously set to be small enough that the physics engine "convexifies" the cube, sealing the hole. \r\nsmall object: Size <0.03266, 0.03266, 0.03266> set to physical.\r\nlarger object: Size <1.49838, 1.49838, 1.49838> Hollow: 4.996\r\nWith the small object placed over the hole, it will not fall because of the convex physics. This is entirely expected. \r\n\r\nWe now increase the hollow size. to 9.9996, the "convexification" on the server is no longer applied, and the correct physics shape is now active; however, the physical prim that was at rest remains entirely unaware of the fact until, when given a suitable event to remind it that gravity is a thing, it will plummet. \r\n\r\n\r\n\r\n\r\n', 'What were you doing when it happened?': 'testing FS', 'What were you expecting to happen instead?': 'Any object at rest should be awakened when a change to the object it rest upon occurs.', 'Where': 'You are at 27.1, 81.9, 3,205.5 in New Babbage located at simhost-06d1c22f942641068.agni\r\nSLURL: http://maps.secondlife.com/secondlife/New%20Babbage/27/82/3206\r\n(global coordinates 161,307.0, 257,362.0, 3,205.5)\r\nSecond Life Server 2023-09-07.6113592855', } ```
sl-service-account commented 1 year ago

Gwyneth Llewelyn commented at 2023-10-07T00:00:54Z

Kudos for giving this issue a memorable name (and an image!). I will never forget this for the next decade or so, when the Linden devs, once they finally stop laughing at it, come up with a solution to fix it... :)

sl-service-account commented 1 year ago

Maestro Linden commented at 2023-10-24T15:09:34Z

I can reproduce this. The issue seems to be that when an object is at rest (dynamic but not moving) and a nearby object updates its collision shape by toggling from convex hull to prim physics, the resting object is not reactivated.

This differs from the behavior I see in this flow:

  1. Smaller cube is physical and resting on a larger cube, both using prim physics
  2. Larger box's "hollow" parameter is increased to a large value (say 95.0), such that the smaller cube sits over the hollow area
  3. The smaller cube falls through the hollow cavity