Re: Chroot and ".."

[Alfred M\. Szmidt]

Sorry for continuing this, but I might just as well try setting some
things a bit straighter.  And I got bored and started reading some old
manuals...

   HISTORY OF "..":

Most of what you wrote about why .. was introduced seems quite wrong.
Early versions UNIX didn't have fsck, this is correct (I think fsck as
such was introduced with BSD).  But they all had file-system
consitency checkers (dcheck, icheck and check, guess what they do),
including version 1.  dcheck and icheck (or check on V1, icheck was
introduced a bit later) look like they were simply doing a two-pass
run over the file-system, each pass having a bound of log(n).

The early versions of UNIX didn't have chroot(), which was introduced
somewhere after V7, and where Korn et al raised the concerns about
`..' and suggesting changes to namei().

But anyway, using .. for path resolution is idiotic, and should have
been implemented the way Korn et all suggested long long ago.