Update sodium compat#146
Conversation
|
|
||
| while (!success) { | ||
| if (cancellationToken.isCancelled()) { | ||
| // In theory, this would never be reached - SkyNotTheLimit |
There was a problem hiding this comment.
spamming F3 + A is an edge case and race condition nightmare
There was a problem hiding this comment.
This was more of a comment of where we happen to inject. Since the target method returns early if the cancellationToken was cancelled, then that statement should never be triggered, and I don't think it ever has, because the handler was not marked as cancellable even thought we call CallbackInfoReturnable#setReturnValue
There was a problem hiding this comment.
since it has to wait for the buffer queue to free space if there isn't enough, it is possible that it gets cancelled while waiting
I just have never been able to fill the index buffer queue
There was a problem hiding this comment.
Ah, I see what you mean now
Would you like me to remove the comment?
There was a problem hiding this comment.
Also, I'm fairly sure I was correct in marking the injector as cancellable, unless you want me to revert that too.
There was a problem hiding this comment.
Also, I'm fairly sure I was correct in marking the injector as cancellable, unless you want me to revert that too.
The cancellable is fine, I missed it when working on the fixes
|
The sodium changes seem to affect the transparency resorting significantly |
|
"Asynchronous Graph Culling" worries me, the baked BE resorting relies on having the graph at the right order, and I don't know what would happen if it has a graph with more sections then the current one |
|
Ok, from what I can see it should be fine. Just have to make sure that the task is cancelled and the index buffer dropped if the async culling removes a section that has been queued for baked BE resorting. |
|
huh, weird, the terraformersmc maven is failing to provide the jars |
|
Oh, so it's not just me? |
|
not even github is being able to get it |
2 participants
Update sodium and modmenu deps
Long2ReferenceMap<RenderSection> sectionByPositionbecameSectionStorage renderSectionsin CaffeineMC/sodium@7d86fa6SectionCollectorwas removed in CaffeineMC/sodium@a327f63 and seems to have been functionally replaced byVisibleChunkCollectorTesting: uncommented runtime sodium dependency and it works:tm: so I think it's fine (probably)