On 2 February 2010 10:08, Jeff Rose <ros...@gmail.com> wrote: > So what would that look like? Is it sufficient to throw every > canonical directory path in a set and test for membership before > traversing down any directory?
Yes, I think so, as long as you can fit all of them in memory. The other thing you could do is make sure not to follow symlinks. > Or is that even sufficient because you > could arrive at the same location through different paths... I wonder No, they will all have the same canonical path. That's what canonical means :) > if you can get something more unique, like an inode ID or something > along those lines? The canonical path should be just as unique as the inode number in theory. I don't know how it deals with things like bind mounts etc. in practice, but I don't think that should cause problems, because you can't create loops with bind mounts (AFAIK). Of course the above is Unix-centric. On Windows you might not have inode numbers (VFAT. Not sure if ntfs has an inode number equivalent.) Also, I think I've heard that in Windows you can create hard links to directories, so I don't know how they avoid loops. > On Feb 1, 11:21 pm, ataggart <alex.tagg...@gmail.com> wrote: >> On Feb 1, 2:14 pm, Jim Van Donsel <j...@jimandbarb.net> wrote: >> >> > Fair enough. >> >> > But I guess the moral is that a file-seq should not be used unless you >> > have a well-controlled filesystem. >> >> Or more generally, all sequences may be infinite unless you know the >> backing data has no cycles, or you have some terminating/filtering >> conditions. -- Michael Wood <esiot...@gmail.com> -- You received this message because you are subscribed to the Google Groups "Clojure" group. To post to this group, send email to clojure@googlegroups.com Note that posts from new members are moderated - please be patient with your first post. To unsubscribe from this group, send email to clojure+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/clojure?hl=en
