Mantis Bugtracker

Viewing Issue Simple Details Jump to Notes ] View Advanced ] Issue History ] Print ]
ID Category Severity Reproducibility Date Submitted Last Update
0000265 [Croquet] Jasmine minor always 10-06-04 03:18 12-30-04 19:25
Reporter jlombardi View Status public  
Assigned To Croqueteer
Priority normal Resolution fixed  
Status resolved  
Summary 0000265: Pointer ray not tracking closely
Description When selecting the mirror halo, the pointer does not track the yellow selection cube closely enough as one rotates or moves the mirror.
Additional Information
Attached Files

- Relationships

- Notes
(0000475 - 177 - 187 - 187 - 187 - 187 - 187)
10-19-04 10:00

Please check to see if this is still a problem (DAS fixed a number of things here). If not, please describe what "not tracking closely enough" means (how close is close enough?)
(0000483 - 620 - 630 - 630 - 630 - 630 - 630)
10-19-04 19:15

I just did an update and the problem still exists. By "not tracking closely enough" I mean that when you grab the top of the mirror and drag the mirror back and forth, the pointer lags behind so that the red arrowhead appears to be moving in a very jerky and imprecise way. It would be far better if the refresh on the arrow was the same as the refresh on the movement of the window. This way users could percieve other user's manipulations more accurately. Doing so will be important to maintaining a quality of experience for users engaged in collaborative interaction involving manipulation of objects in a scene.
(0000484 - 351 - 357 - 357 - 357 - 357 - 357)
10-20-04 06:59

I see. The problem is clearly in TAvatar>>pointerTransform:. Something in the computation is wrong but it converges after a few cycles. I can't quite figure out where the problem is since usually the changes are too small to see anything interesting. I'm assigning this to DAS - he wrote the code originally so I'm sure he can figure out what's wrong.
(0000994 - 692 - 692 - 692 - 692 - 692 - 692)
12-30-04 19:25

This was due to the laser transform using the previous pointer transform values. Note that there is still something of an issue here, as the laser pick position is often recalculated by the target object - for example the TWindow - which will sometimes generate what appears to be a slightly incorrect value during a pointer move event. This is because the TWindow (for example) will generate another pick event on pseudo-objects like a cylinder in the case of rotation or a plane in the case of translation. Also note that by definition, we compute the object the real pointer is looking at during render time and display the resulting laser position and length during the following render.

- Issue History
Date Modified Username Field Change
10-06-04 03:18 jlombardi New Issue
10-19-04 10:00 andreas Note Added: 0000475
10-19-04 10:00 andreas Status new => feedback
10-19-04 19:15 jlombardi Note Added: 0000483
10-20-04 06:59 andreas Note Added: 0000484
10-20-04 06:59 andreas Assigned To  => Croqueteer
10-20-04 06:59 andreas Status feedback => assigned
12-30-04 19:25 Croqueteer Status assigned => resolved
12-30-04 19:25 Croqueteer Resolution open => fixed
12-30-04 19:25 Croqueteer Note Added: 0000994

Mantis 1.0.8[^]
Copyright © 2000 - 2007 Mantis Group
50 total queries executed.
37 unique queries executed.
Powered by Mantis Bugtracker