Shawn Walker wrote:

> On 04/14/10 04:44 PM, Danek Duvall wrote:
> >Shawn Walker wrote:
> >
> >>http://cr.opensolaris.org/~swalker/pkg-529-2/
> >
> >generic.py:
> >
> >   - line 863: fsobj_checkpath() doesn't appear to be called in any
> >     removal contexts, though if it were, the error on line 871 woudln't
> >     be appropriate.
> 
> I wasn't sure that we wanted/needed to do that.

So in what contexts would "not destination_fmri" be true?

> >   - line 866: is there any way we can run into the usual problem with
> >     os.path.commonprefix(), in that it doesn't actually act on path
> >     components, but characters?
> 
> I don't anticipate it, was there a specific scenario you envisioned?

Let's say that final_path is /usr/share/foo/thing.html.  And let's say that
/usr/share/foo is a symlink to ../shopping/foo.  Then parent_path is
/usr/share/foo, and real_parent_path is /usr/shopping/foo.  Because of the
behavior of commonprefix(), cmn_path is /usr/sh, which doesn't exist, so
the join on line 868 doesn't make much sense.

Danek
_______________________________________________
pkg-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/pkg-discuss

Reply via email to