> > I dont see the problem. You follow the link. At that point you worry about
> > creating an object. The race there is a userspace issue.
>
> There are kernel races. BTW, you do realize that in our case following
> link != reading the contents and doing lookup?
Now tell me why it shouldnt in this case ? We need to know the tail item of
the path and the directory the symlink pointed us at when we followed it. We
dont need to care if the symlink changes around that moment.
> > You might want to take that up with the relevant posix committee then
>
> What an elegant way of sending somebody to hell... ;-) Seriously, I think
> that at the very least we should get something along the lines of
> "portable programs can not rely on creat(2) turning dangling symlinks into
> normal ones" in there. Any clues on how one deals with POSIX committees?
> Aside of "take a ten-feet pole...", that is.
ENOCLUE. Since there are people here citing posix drafts hopefully they
know
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Thu Jun 15 2000 - 21:00:23 EST