[turba] First shot at a new ldap schema.

Adam Tauno Williams adam at morrison-ind.com
Sun Mar 16 15:48:32 PST 2003


>>>>One might want to consider create a subtree for each application, so
>>>>that "team" can create schema without concern about OID collision.
>>>> Like
>>>>1.3.6.1.4.1.13040.2.1.* Horde global attributes
>>>>1.3.6.1.4.1.13040.2.2.* Horde global objectclasses
>>>>1.3.6.1.4.1.13040.3.1.* IMP attributes
>>>>1.3.6.1.4.1.13040.3.2.* IMP objectclasses
>>>>1.3.6.1.4.1.13040.4.1.* Turba attributes
>>>>1.3.6.1.4.1.13040.4.2.* Turba objectclasses
>>>>
>>>>It is a nice way to share an OID space across projects/divisions
>>>That sounds like a good idea to me. If someone can organize the latest
>>>revisions into a schema file, I'll look at actually getting something
>>>into CVS...
>>My attempt is attached.
>>I included the new Turba attributes, as well as an auxillary objectclass
>>to allow them to be added to existing objects.
>Awesome, Adam.  Thanks.  After seeing this I'm going to have to redo
>min :-)  
>I'm in the process of implementing the new schema[s] while upgrading to 
>openldap 2.1.12.  Upon adding the rfc2377.schema, errors are 
>generated on the matching and substring rules that I was able to quiet
>by changing from 
>EQUALITY caseIgnoreMatch
>UBSTR caseIgnoreMatch
>to
>EQUALITY caseIgnoreIA5Match
>SUBSTR caseIgnoreIA5SubstringsMatch
>That solves the immediate problem and it seems to work.  I'm pretty
>sure I shouldn't do that but what would be the correct solution?  Has
>anyone else runing 2.1.xx seen this problem with the rfc2377 schema?
>How did you solve it?

What you did was correct.  An attribute of type  1.3.6.1.4.1.1466.115.121.1.26
should have an EQUALITY/SUBSTRING of
caseIgnoreIA5Match/caseIgnoreIA5SubstringsMatch (assuming the search should be
case insensitive).  caseIgnoreMatch/caseIgnoreSubstring match are by attributes
of type 1.3.6.1.4.1.1466.115.121.1.15 (see the "name" attribute in core.schema
as an example).

This is a bug in the rfc2739.schema and fixing it is perfectly legal, since the
text from the RFC is just flat out wrong. I'd just forgotten about that issue
and sent the cannonical (aka "the busted") version.

Adam Tauno Williams
Network & Systems Administrator
Morrison Industries
Grand Rapids, Mi. USA


More information about the turba mailing list