On Tue, 2004-12-07 at 01:41, Peter N. Lundblad wrote:
> On Mon, 6 Dec 2004, Philip Martin wrote:
>
> > VK Sameer <sameer@collab.net> writes:
> >
> > > +svn_error_t*
> > > +svn_path_is_valid_in_svn (const char *path,
> > > + apr_size_t len,
> > > + apr_pool_t *pool)
> > > +{
> > > + int i;
> > > + for (i = 0; i < len; i++)
> > > + {
> > > + if (!svn_ctype_isutf8(path[i]))
> > > + return svn_error_createf(SVN_ERR_FS_PATH_SYNTAX, NULL,
> > > + "Invalid UTF8 character in '%s'",
> > > + svn_path_local_style (path, pool));
> >
> > UTF-8 is a multibyte encoding so this doesn't ensure that the path is
> > valid UTF-8. We already have code to check for valid UTF-8, see
> > utf_validate.c, and it's already in use. The above is either
> > redundant if combined with the existing code, or insufficient if it
> > replaces the existing code.
> >
> > I think you should either drop attempts to test UTF-8 or build on the
> > existing UTF-8 validation code.
> >
> It is redundant, since we already do this check when converting from
> native to UTF8, even if native is UTF8. Validating input on the server
> side when expecting valid UTF8 wouldn't hurt, though.
OK, will look into splitting function into client-side and server-side
checks, without and with utf8 checks, respectively.
Thanks
Sameer
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Tue Dec 7 02:23:58 2004