Re: file backend

[Jonghyuk Choi <jongchoi@us.ibm.com>]

How about support indexing through file offsets ?
It seems not difficult to modify bdb/hdb index cache
to hold file offsets instead of IDs.
- Jong

>I've been thinking of making a simple 'file' read-only backend which
>will read a file of records into memory, build a cache with one entry
>for each record, and let searches go through all the entries and apply
>the search filter to each.  Before each search operation, the backend
>checks if the file has been updated, and if so rebuilds the cache.
>
>Configuration would be something like
>
>database         file
>suffix           ou=contacts,dc=uio,dc=no
>filename         /some/file/name
>record-separator "\n\n"
>field-separator  ":"
># list of attributes, one for each field in a record
>attributes       cn sn uid mail labeledURI
># which attribute to make the DN from
>dn               uid
># list of object classes for each entry
>objectclasses    top person
># add these object classes if the attributes exist in an entry
>objectclassForAttrs pilotPerson      mail uid
>objectclassForAttrs labeledURIObject labeledURI
>
>Maybe I'll add an option to make it accept files in LDIF format too,
>while I'm at it.

------------------------
Jong Hyuk Choi
IBM Thomas J. Watson Research Center - Enterprise Linux Group
P. O. Box 218, Yorktown Heights, NY 10598
email: jongchoi@us.ibm.com
(phone) 914-945-3979    (fax) 914-945-4425   TL: 862-3979