A discussion started on the nano devel mailing list about the option --nofollow. Ideally, this prohibits the editor from writing to a file referenced by a symlink.
At first it seems nobody was concerned about the problem, but I and one other gentleman pointed out that in a shared dir, root writing to a file that is changed, mid edit, to a symlink would pose a significant security problem.
Many editors warn if a file has changed, but in the age of race conditions, can you really tell how it changes between the time you hit save and the time it actually writes out to disk?
My question is, where should we go from here. Is this a real problem, or am I getting excited about nothing?

Ancient problem. There's plenty of examples resulting in arbitrarily overwriting files. As for how to move forward maybe start by reading comments here.

1 members found this post helpful.

Fixable since 3.6 - see the manpage for proc, look for protected_symlinks

1 members found this post helpful.

Thanks for the links, is there a general CVE or similar, for the linux kernel alerting people that this attack was possible?

It's always been possible ... but, that is why we invented symlinks!

To me, this sort of scenario becomes rather farfetched. It presupposes the existence of a rogue process that knows what you are doing and that contravenes that intention in a certain way and at a certain time. It also conveniently presupposes that the root user is the one whose activities are being interfered with, in this very-certain way. To me, this "vulnerability" is truly hypothetical.