-
Notifications
You must be signed in to change notification settings - Fork 3.5k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Changing terrain provider moves billboard to incorrect location #8006
Comments
One detail that may be relevant. If there weren't different pixel offsets for each billboard in this sandcastle, the billboards would render on top of other billboards in the set. So they end up rendering on top of each other when this issue happens. |
So many billboard tickets out there, but I think this is the most relevant to what I'm seeing and I wanted to add my observations. So here it goes... Stationary billboards do not clamp correctly when entities are removed just before terrain is reloaded and entities are added while terrain is still loading. Note that in the example code(see link below), all entities are removed, but even if one or more entities are removed and new ones are added you can see the issue. Observations
It looks like the QuadtreePrimitive customData list keeps growing and some stale height callbacks are being made with old coordinates. This seems to be in agreement with @lilleyse comment in #8650 For my case I can delay loading of entities (by timer or wait until all tiles are loaded), but it really isn't optimal. Note that we have been seeing this issue since at least 1.57, even possibly as far back as 1.51, but have not had time to debug. |
Note that the delay in loading entities does not always help and also I noticed that the customData length continues to grow each terrain reload so that is a bit worrisome. |
…ltiple callbacks being registered for the same billboard in invalidateAllTiles, resulting in leaked/stale callbacks and bad clamped locations.
This came up on the forum. Here's a Sandcastle example. To reproduce:
This may be just another manifestation of #4776 but I couldn't dig deep enough to figure out what the underlying issue was.
The text was updated successfully, but these errors were encountered: