directory-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Alex Karasulu <>
Subject Re: [jira] Updated: (DIRSNICKERS-102) The Length' s nuber of byte must not be 127
Date Sat, 12 Feb 2005 21:32:04 GMT
Vincent Tence wrote:

>> Those are two test cases related to the issue.
>> The Undefinite Length form is not allowed by RFC 2251 : chapter 5.1, 
>> limitation (1), so it could make sense to throw an exception in this 
>> particullary case. However, it may seems a little bit restrictive, so 
>> I suggest that a configuration flag should be use to allow or 
>> disallow this feature.
>> The point is : are we trying to build a generic ASN.1 compliant 
>> decoder, or a specific one. 
> I am gald you brough that point up Emmanuel, and I'd like to know what 
> the other guys think about this. What the best approach for us?
> 1. Build specific component and grow them to generic ones later on if 
> there's a need.
> 2. Try to build generic components from the start
> I would favor #1, so we concentrate our energies on our needs and we 
> build the best API for what we have to do. We can then evaluate 
> opportunities for extracting generic components.
> Opinions?

+1 I agree. 

The only subtle difference for me is that I would use interfaces and 
hierarchy to differentiate the codecs.  Like for example DER is a subset 
of BER and some would just say its constrained BER.   Also note that the 
LDAP BER is not full BER, it too is a subset.  So when we build a 
decoder for example it might be nice to have the same interfaces and 
swap out the underlying implementation when we want. 


View raw message