Damaged files on USB drive?
patric aristide (434) 418 posts |
Is there a way RISC OS or FAT32FS might corrupt files on a flash drive? For testing purposes I played some mpeg files from a FAT32 formatted pen drive. KinoAmp seemed happy with them but surprisingly my DVD recorder later refused to read those two files. They were in fact missing from the list. Thought I accidentally deleted or copied them and plugged it into my XP machine to investigate. Nope, they’re still there but after a few seconds windows explorer crashed and refused to correctly identify the drive at all. Interesting. Haven’t tried it under Linux (though I’d assume that the DVD harddisk recorder uses some kind of Linux) but it seems as if accessing the drive on my BB did create problems. No harm done concerning the data as such but still. RISC OS reads everything okay, correctly identifying the 8GB drive as “USB FLASH DISK” (Vendor ID 090C; Product ID 1000). |
Dave Higton (281) 668 posts |
I’ve recently had cause to suspect that Fat32Fs doesn’t (always?) write out all of its cached content on the first dismount operation. Now I routinely dismount twice if I’ve been using Fat32Fs. This would only affect you if you used Fat32Fs to write one or more files to the drive. I have never suspected it of writing when it should only have been reading. Do make sure you’re using the most recent version of Fat32Fs; Jeff Doggett has updated it several times. I’m sorry to be pointing the finger of suspicion at Fat32Fs. The problem is that I don’t know how to diagnose for certain whether my suspicion is right or wrong. If I’m wrong, I’ll be glad to admit it and then keep quiet! |
Ronald May (387) 407 posts |
The doesn’t (always?) bit is going to make it hard. I wonder if having two drives mounted would help draw a conclusion. If after dismounting a drive, you do something with the other drive, it would force the cache to change, or is there a cache for each drive? If you are getting results with dismounting twice maybe this method (or others) could help expand on the ‘feature’ |