Releasing the "windo_filter" texture, then using it again causes it to be replaced with a sprite displayed on screen

[drojf]

Disclaimer: I haven't tested this on the stock game, so it could be some other of my changes which causes this to happen. Also, it's not possible? for this to happen on the base game, but it is more of a "annoying for developers" thing.

Currently, before we switch window filters (the background for the textbox), we call ReleaseTextures(); just before it:

higurashi-assembly/Assets.Scripts.UI/MainUIController.cs

Lines 477 to 480 in 2393805

	GameSystem.Instance.MainUIController.bgLayer.ReleaseTextures();
	GameSystem.Instance.MainUIController.bgLayer2.ReleaseTextures();
	GameSystem.Instance.MainUIController.bgLayer.DrawLayer("windo_filter", 0, 0, 0, null, gameSystem.MessageWindowOpacity, /*isBustshot:*/ false, 0, 0f, /*isBlocking:*/ false);
	GameSystem.Instance.MainUIController.bgLayer2.DrawLayer("windo_filter", 0, 0, 0, null, gameSystem.MessageWindowOpacity, /*isBustshot:*/ false, 0, 0f, /*isBlocking:*/ false);

When testing my own code, I found that if you call this twice with "windo_filter" as the image, it can result in this happening.

It only happens with "windo_filter" in particular, because there is a hardcoded cache for that particular image in:

higurashi-assembly/Assets.Scripts.Core.AssetManagement/AssetManager.cs

Lines 242 to 245 in 2393805

	if (textureName == "windo_filter" && windowTexture != null)
	{
	return windowTexture;
	}

However, if you remove the ReleaseTextures();, then the problem goes away.

I'm guessing that if you call releaseTextures on windo_filter, it doesn't invalidate (it doesn't set windo_filter to null), and when the game tries to use that texture, it ends up using some random texture instead.

I could either fix this by disabling the caching (currently you're not caching the ADV/NVLADV textures), or adding an extra condition check which disables the cache. Edit: If I just use a third type and switch ADV->NVL->OG->ADV... then the problem goes away so I guess it's not really an issue for now.

I don't think this is related to #45 even though it looks similar, because the sprite appears slightly transparent (it draws it with the window opacity).