Re: utility functions for handling Distinguished Names

[Graham B. <gb...@po...>]

On Tue, Apr 23, 2002 at 06:59:55AM -0700, Kurt D. Zeilenga wrote:
> At 04:32 AM 2002-04-23, Graham Barr wrote:
> >OK, I made a couple of tweaks and it passes the testsuite.
> >
> >One change is that
> >
> >  CN=#0102
> >
> >Now decodes as   { CN => \"\01\02" }
> 
> That's incorrect.  The # format is hex encoded BER.  The
> first octet (01) is the (wrong) tag, the second is the data
> length (02), and the data is missing.  Hence, it cannot be
> converted to the LDAP-specific encoding of the value.

I know this is a bad BER example. What I was trying to show is that the value
is decoded into bytes, but the structure holds a reference so that it knows
it has to encode it back as #\01\02

Graham.

> >That is the value is a reference. Ans canonical_dn regenerates it correctly.
> >
> >Graham.
> >
> >On Tue, Apr 23, 2002 at 11:19:40AM +0100, Graham Barr wrote:
> >> On Tue, Apr 23, 2002 at 11:21:57AM +0200, Norbert Klasen wrote:
> >> > --On Freitag, 19. April 2002 08:43 +0100 Graham Barr <gb...@po...>
> >> > wrote:
> >> > > Whats wrong with the subs in Net::LDAP::Utils ?
> >> > 
> >> > ldap_explode_dn does not unescape escaped characters, e.g. the DN 'CN=\#John
> >> > Smith\ ' becomes
> >> > $VAR1 = {
> >> >           'CN' => '\\#John Smith\\20'
> >> >         };
> >> > 
> >> > instad of
> >> > 
> >> > $VAR1 = {
> >> >           'CN' => '#John Smith '
> >> >         };
> >> 
> >> Hm, I guess canonical_dn needs to be split from usinf explode_dn,
> >> or explode_dn needs to be told not to expand. Otherwise you cannot tell the difference
> >> between
> >> 
> >>   CN=#a0
> >> 
> >> and
> >> 
> >>   CN=\#a0
> >> 
> >> Do all testcases pass with your functions in place of the existing ones ?
> >> 
> >> Graham.
> >> 
>