Re: [ldapext] new version of LDIF

[Hallvard B Furuseth <h.b.furuseth@usit.uio.no>]

Ludovic Poitou writes:
> I personally would like to see full support for UTF-8 characters in  
> LDIF,

yes...

>  and possibly other encodings.

That should be optional, so LDAP implementations are not burdened with a
requrement to support charset encoding if they do not actually need it.

If a file or entry has 'charset: ...' this should not only affect
'attr: val', not 'attr:: val' and 'attr:< file'.  The two latter may
contain non-text data, e.g. a jpeg picture.

So an app which writes latin-1 may still need to convert to utf-8, in
case it needs to write a value which must be base64-encoded.  E.g. a
value which contains a newline.

A fix would be to have two base64 attribute syntaxes: one which is
charset-converted after base64-decoding, and one which is not.


A looser grammar would be useful too.  Allow an LDIF file
with no entries.  Extra CRLFs at the end of the file.


I have various notes about wishes for LDIF spread around in my
mailbox...  Probably I should have undertaken to update LDIF format
myself, but I never seem to get around it.  I'll see what I can find.

-- 
Hallvard
_______________________________________________
Ldapext mailing list
Ldapext@ietf.org
https://www.ietf.org/mailman/listinfo/ldapext