"Peter N. Lundblad" <peter@famlundblad.se> writes:
> OK. We have a definition of a canonical path. We have
> svn_path_canonicalize (and, at least currently, svn_path_is_canonical()).
> It is reasonalbe to assume that, if you call svn_path_is_canonical on a
> path, you can pass that path to any API function that expects a canonical
> path and get the expected result. My point is that we *might* have APIs
> that rely on other properties of a canonical path than what
> svn_path_is_canonical checks currently. For example, svn_path_is_canonical
> doesn't check for multiple consecutive slashes. Some function might depend
> on each component being separated by exactly one slash. This might lead to
> other bugs or crashes that this fix pretends to prevent, but doesn't.
> Ofcourse, if svn_path_is_canonical gets fixed, this point isn't anymore.
Thanks, that's totally clear for me now!
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Sun Jan 16 02:38:10 2005