Re: current control combination proposals

[Chris Ridd <chris.ridd@isode.com>]

On 17/5/04 4:11 pm, Kurt D. Zeilenga <Kurt@OpenLDAP.org> wrote:

> At 11:08 PM 5/16/2004, Chris Ridd wrote:
>> On 17/5/04 2:44 am, Ramsay, Ron <Ron.Ramsay@ca.com> wrote:
>>> constrainViolation is only appropriate for update operations.
>> 
>> So far.
> 
> Which is a good reason not to do that.
> 
> But Ron's point is quite valid.  ConstraintViolation has a well
> established meaning and expanding it to include other error
> conditions would cause problems.  A client would have no way
> to determine if the error indicated an update problem or
> indicated a control problem.  Likewise for most other existing
> result codes.

Well, I'm not going to defend this to death. The discussion is interesting
enough without thinking about the error code :-)

>> It would be much nicer of course if result codes were somehow extensible,
>> but that's another problem altogether :-)
> 
> The resultCode enumeration is extensible.  However, we should

Not flexibly - there's only a flat number space so different implementations
could easily collide their resultCode extensions.

> attempt to revise the specification to resolve whatever
> issues exist without resorting to adding a new code.

Agreed.

> It has been reported that existing implementations return
> protocolError here.  What's broke with that?

Well, for one thing no protocol's being broken. It seems a bit severe.
 
Cheers,

Chris