The asset configuration window is one of the slowest interfaces in Studio. Depending on the size of the model you are publishing, it can be even slower.
This video perfectly captures the issue I’m encountering. The model I published here has 33,000 instances.
In comparison, here is the same process trying to save a single part to Roblox:
Our games have a system loader which fetches the most up-to-date models from Insert Service on start. It is crucial to our development this window perform well and consistently.
Hey Usering, we’re going to be working on a rough change to not render anything if it’s over a certain threshold of instances. Our hypothesis is that after a certain point, it’s significantly more likely that the model isn’t an individual entity (like a building with a detailed interior), and is instead something like what you’re doing of entire landscapes/cities, so we should be able to exit out early. Thanks for your report!
Any updates on this? It is a serious pain having to wait a few minutes each time I need to push this model. The race condition of showing the wrong group/user from what has been selected is also a major pain point still.
I don’t think my fix got in by code cutoff, but from what I can gather we can feasibly update just the Toolbox and be fine. I’ll kickstart that process on Monday.
Looks like the fix for the large models rolled out. Hard to tell when there is still other performance issues prevalent.
In this video, there is significant load time rendering all models. When it does finally load, the wrong models are loaded into view; in the video it loads my personal models after selecting the “Drive World” group.
Again, we use this window religiously for our system loaders across many teams. The time it takes to publish a model should be 5 seconds, not 30.
It hasn’t been turned on yet, you shouldn’t have noticed anything (though I think some other flags related to the publish window have been flipped that may have done something), will make a reminder for myself for Monday.
It’s now Monday, so an update: there’s some stuff that needs to be done from people other than me before we can flip this on that I expected to take less time, but am regularly monitoring when it’s done and leaving a bunch of reminders for myself
Hey Usering, I flipped this this afternoon. The limit is 10,000 descendants. Does this help you? If not, I can work with you to find a value that does work so that we can gauge that as a new setting.
The window runs much faster now with our larger model, the fix there is much appreciated. However I constantly run into the issue of the wrong user/group being loaded; this happens if you select a different owner while results are loading for your own user models. If that race condition could be patched my model publishing experience would be much improved.