Re: UTF-8 full support in LDIF / LDIF v2

Steven Legg wrote:
> 
> I would be more amenable to a here-document scheme if there were no
> special treatment for line separators. That is, if the UTF-8 character
> sequence between the end of the introducer and the beginning of the
> terminator is the literal directory attribute value without any 
> modification.

That would be easy to implement, just pick one line separators instead of 
allowing several. But, is this the right way of doing things ?

When you deal with multi-line entries, isn't the LDAP client at the end of 
the chain modifying your line separator into a different one, one 
appropriate for its platform ?

If so, how much more difficult would it be to let every element in the chain 
use whichever line separator they want, and translate as needed ?


> The downside of a here-document scheme is the need to scan the value
> before hand to make sure it doesn't contain the terminator sequence,
> which becomes awkward for streaming applications, so I'd be inclined to
> keep the current scheme as an alternative.

It should be easy to pick a sufficiently complicated word that you know will 
never appear in any of your entries (e.g.: X**ENDOFTEXTDONOTUSE**X), and you 
can still scan for it from your entries and raise an exception in the 
unlikely case that it does appear (no need to scan ahead, still fool proof). 
This is a slight inconvenient, but the advantages of not having to escape 
characters, and of making the file much easier to read make it, in my 
opinion, worth it.

-- 
Yves.
http://www.sollers.ca/