Re: Blank Attribute - further question

[Daniel Henninger <daniel@ncsu.edu>]

> > In reference to the previous question from August:
> >
> >
> > > At 07:22 AM 8/11/2005, Ivan.Garcia@leroymerlin.es wrote:
> > >
> > > > adding new entry "uid=30000001,ou=people,ou=es-  
> > > > lm,o=corp.happyparty.com"
> > > >
> > > > ldap_add: Invalid syntax (21)
> > > >        additional info: mail: value #0 invalid per syntax
> > > >
> > > > this appears when the mail attribute is blank, but when the mail
> > attribute
> >
> > > > is writted that´s all OK.
> > >
> > > If you example the technical specification for the LDAP syntax
> > > of the 'mail' attribute you will find that it doesn't allow
> > > empty values.  Hence, slapd(8)'s response.
> > >
> > > Kurt
> >
> > I am having trouble figuring out what it is about the mail attribute
> > that indicates "can not be blank".
>
> See Section 9.3.3 (RFC 822 Mailbox) of RFC 1274 (also known in
> as the 'mail' attribute in LDAP).  Note the SIZE restriction.
> slapd(8) is coded to enforce this restriction.
>
> Also note that a valid RFC 882 Mailbox [RFC 2822] cannot be
> zero-length.  If the object has no mailbox associated with it,
> the 'mail' attribute should not be present.  Providing a
> 'mail' attribute (or any other attribute for that matter)
> with garbage values will lead to interoperability problems.

Right, but that doesn't look the same as what I'm seeing in the  
OpenLDAP schema files.

     rfc822Mailbox ATTRIBUTE
         WITH ATTRIBUTE-SYNTAX
             caseIgnoreIA5StringSyntax
             (SIZE (1 .. ub-rfc822-mailbox))
     ::= {pilotAttributeType 3}

Just to make sure we are on the same page, I assure you that I am not  
planning on trying to set the actual 'mail' attribute to 0 length.  I  
simply used that as an example.  I'm actually creating an entirely  
new custom attribute but have been unable to figure out how to make  
it permit a 0 length string.  The SIZE restriction in the OpenLDAP  
schema file looks like 256 max to me.  Am I reading that wrong?

attributetype ( 0.9.2342.19200300.100.1.3
        NAME ( 'mail' 'rfc822Mailbox' )
        DESC 'RFC1274: RFC822 Mailbox'
    EQUALITY caseIgnoreIA5Match
    SUBSTR caseIgnoreIA5SubstringsMatch
    SYNTAX 1.3.6.1.4.1.1466.115.121.1.26{256} )

{256} is the SIZE, is it not?  And it looks like a maximum (as that  
is obviously not the minimum).  But lets say I have something like this:

attributetype ( 1.3.6.1.4.1.234.1.15
        NAME 'ncsuPreferredMiddleName'
        DESC 'Preferred middle name for display'
        EQUALITY caseIgnoreMatch
        SUBSTR caseIgnoreSubstringsMatch
        SYNTAX 1.3.6.1.4.1.1466.115.121.1.15 SINGLE-VALUE )

Now, this can be blank if you want your middle name to not be listed  
at all.  We have a ncsuMiddleName as well, which is your legal middle  
name.  The theoretical check would go like this:

if (ncsuPreferredMiddleName) { use that}
else { use ncsuMiddleName }

Well, the user might want to specify an empty preferred middle name,  
thereby saying "I don't want you to list the middle name".  Whereas  
if I simply did not have the ncsuPreferredMiddleName there, one might  
assume "ok, they don't have a preference, stick with legal".

So how would I indicate that it's permissible to have a 0 length  
string for ncsuPreferredMiddleName?  =/

Is it possible to do something like this?

attributetype ( 1.3.6.1.4.1.234.1.15
        NAME 'ncsuPreferredMiddleName'
        DESC 'Preferred middle name for display'
        EQUALITY caseIgnoreMatch
        SUBSTR caseIgnoreSubstringsMatch
        SYNTAX 1.3.6.1.4.1.1466.115.121.1.15{0..} SINGLE-VALUE )
or..
       SYNTAX 1.3.6.1.4.1.1466.115.121.1.15{0..256} SINGLE-VALUE )

Daniel

--
Daniel Henninger <daniel@ncsu.edu>
Systems Programmer
Information Technology Division