and btw it is tricky to reproduce! you need process large amount of data to be able to see the issue. Thanks!
Glad to hear you figured out why it’s happening.
Yeah, I’m using 2.0.6 version for now. But of course a bunch of other people have already uploaded split sequences by now. I deleted and have/will reupload mine.
And yeah, it does require a large amount of files. I can get it to reproduce almost 100% of the time with like 25k images, but that’s insane to debug. And since you seem to be saying it’s because of parallel processing, then image size (hard drive reading) matters. Yikes!
I am unaware of DU having source code/repository or alpha/beta channel, but I can run a dry test or two on my images if you want to make a test version.