It's easy to laugh at them but I've had corruptions from Linux's existing NTFS implementation too.
Isn't Linux's existing NTFS implementation read-only? I'd be very surprised you would get corruptions from that..
FUSE one is not just read-only. It had capacity to do write on NTFS partitiona for decade at least.
But, it is FUSE.
Am I correct in saying that there shouldn’t be any reason that being FUSE would in and of itself cause corruption?
Performance is a different question, since it’s not in kernel.
Off the top of my head... ordering violations. FUSE has a lot less control about the exact order in which things get written to disk, meaning that you are more likely to wind up with the wrong parts having committed down first if there's a crash.
That’s FUSE which is in userspace. The kernel driver is read-only. I believe it never progressed past that due to licensing issues.
I know what FUSE driver is, but my educated guess would be that almost no one ever used in-tree kernel-space driver which you referring to. So when people talking about issues with old driver that's almost certainly NTFS-3g
I've had issues with their Windows HFS+ driver too, shit wiped out a hard drive that I later learned wasn't fully backed up
+1 had the same issue with their HFS+ driver, and it did not take that long to encounter it.
I guess it's not that surprising really[0], but it is quite amusing in combination with the company name.
[0] If you tie your company's worth to a proprietary technology.
I mean the thread is they're open sourcing their work (if by commercial you mean proprietary).