Nate wrote
Abelius wroteI've only had a "Some images couldn't be loaded" just once, probably because of some kind of memory leak.
If this happens you can check your spine.log for why it failed. Spine should never leak memory.
Hi Nate,
My project is growing with dozens (if not hundreds) of those images exported at canvas size, and this issue is getting more frequent.
So, I need to exit Spine and enter again, and it works well for a while, but I'm afraid this will get worse and I can't even work for five minutes straight.
Let me paste you what I've got from the log last time...:
ERROR: Unable to read image file: C:/_HHS/SPINE/Side/./images/Female/Foot_F_Sock_White_Shaded.png
gS: Unable to allocate memory for pixmap: 1144x2602, rgba8888
at com.badlogic.gdx.graphics.g2d.Gdx2DPixmap.<init>(SourceFile:124)
at aj.<init>(SourceFile:118)
at Pg.D(SourceFile:1078)
at lPt6.D(SourceFile:3070)
at it.call(SourceFile:64)
at java.base/java.util.concurrent.FutureTask.run(Unknown Source)
at java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(Unknown Source)
at java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(Unknown Source)
at java.base/java.lang.Thread.run(Unknown Source)
The thing is, the way I manage attachments could contribute to this problem...
This is what I usually have in each slot. Several meshes pointing at canvas size images. I always leave a "catalog" of meshes outside the skin placeholder, so I can easily make a copy and put it inside with each skin I make.
But I suspect the reason I run out of memory is that Spine must use a lot when I mouse over each of them on the Hierarchy. I could easily pass through 8-10 in just one second, and maybe that's why Spine "overloads".
Also, once this happens and I start getting "error loading images" messages, it doesn't matter if I disable all skins. Not a single image gets loaded after that.
As I mentioned, I have 11 GB of VRAM on my video card, but something tells me this doesn't have anything to do with my VRAM. :rolleyes:
I vaguely remember you mentioned Spine will be able to address memory on a 64-bit address space in the next version (?). Could that fix my issues? Or could I do other things in the meantime?
I guess reducing the number of attachments on the slots (by placing them alone in their skins) would make things better... but I'd really like to keep those "catalogs" of sorts, so I don't need to remember what I have.
Thanks for your time.
Update: I've just made a little additional test.
I've opened Spine fresh and started to mouseover up and down those meshes on the Hierarchy, until the errors started to show up. Much sooner than usual.
I've made a little video illustrating it...:
I can also see on the Task Manager (which wasn't captured, lol) that the used RAM goes up fast, but I have 32 GB and it goes from 14-something to almost 17 GB of use, when I do this.