[Date Prev][Date Next] [Chronological] [Thread] [Top]

Re: Empty IA5String


Kurt,

Kurt D. Zeilenga wrote: 
At 11:41 AM 11/10/2004, Hallvard B Furuseth wrote:

Kurt D. Zeilenga writes:

I assume what you mean here is that you are increasing
believing that LDAPprep should be redesigned to accept
as input an zero-length string and output a zero-length
string for this input.

This would require a major redesign of the algorithm. 

Why?  Just put the zero-length shortcut you gave at the top of
stringprep algorithm instead of in the matching rule(s) which uses it:



Even if we presume that the proper LDAPprep("") output is "",
specifying this in the I-D would require significant revision
of the text.


I don't see it. Which of the steps in LDAPprep would be broken if the
output of the previous step (or the original input) happened to be a
zero-length string ?

> And, if we presume that LDAPprep("") output for 
some character string matching rules is not "", the I-D would
require significant redesign.


LDAPprep already has internal steps that replace an empty string with " ",
so why isn't this cause for a significant redesign ?

Regards,
Steven


In short, LDAPprep was designed for X.520-based matching rules.
It wasn't designed to handle empty inputs and redesigning it
to handle such inputs is, IMO, non-trivial.

However, before we consider any change to LDAPprep, we first
must reach consensus the semantics of empty values in IA5
Strings.  I note as well that I intend to argue that IA5
Strings, as specified in RFC 1274, were intended to be
non-empty and expected in many to be non-empty and that by
allowing non-empty values in IA5 string will lead to far
more problems than it will resolve.

Kurt