On 7/19/2019 4:28 AM, Corinna Vinschen wrote: > I see what you're doing here, but it's totally non-obvious from the > commit message why this fixes the problem and doesn't introduce weird > side-effects.
Thanks. I was pretty careless with this patch. There's a new patch series on the way that (I hope) does it right. > An editorial note: While looking into your patch it occured to me that > it would be about time to go over all the is***device() methods and > clean up the mess. E.g., is_fs_device() is used by is_lnk_special() > only, is_auto_device() doesn't have much meaning, I've removed is_fs_device() and is_auto_device() > some funcs have > underscores, some don't. The convention seems to be that is<something> uses underscores if and only if "something" is a single word. The only exception I saw is isctty_capable. I didn't bother changing this, but I could if you want me to. Ken