I suspect that the answer is 'badly'. V7's chroot() seems to be more than a little bit of a fast hack, one that was good enough for some things but not at all comprehensive or problem free.
(There are PDP-11 emulators and V7 disk images available through tuhs.org, so intrepid people who want to find out for themselves can actually try this out on a live V7 system.)
That is indeed the modern behaviour of chroot. It is not clear that the "original implementation" shown in the article shares that property, though.
EDIT: That's not to say it doesn't prevent that behaviour. If the usual code to prevent you ..ing above the root checks against u.u_rdir instead of /, it might well work as modern implementations do.
I think that chroot predates anything in kernel that would care about canonical paths. OS X manpage (first google hit for "getcwd history") claims that "The getwd() function appeared in 4.0BSD.", which is to this day essentially the only syscall that returns path name.
Very early versions of Unix didn't enforce a hierarchical file system---it was, in fact, a directed graph (not only could you have multiple hardlinks to files, but multiple hardlinks to directories!). (information from "Unix Implementation", Bell System Technical Journal, July-August 1978). So things back then were a lot less ... structured.
12 comments
[ 3.0 ms ] story [ 47.3 ms ] thread(There are PDP-11 emulators and V7 disk images available through tuhs.org, so intrepid people who want to find out for themselves can actually try this out on a live V7 system.)
EDIT: That's not to say it doesn't prevent that behaviour. If the usual code to prevent you ..ing above the root checks against u.u_rdir instead of /, it might well work as modern implementations do.