Re: iSCSI: Case-sensitivity in iSCSI names

[Mark Bakke <mbakke@cisco.com>]

Robert Snively wrote:
> 
> I did want to point out one characteristic of this.
> 
>         Whatever mechanism you use to assign the names SHALL
>         cross check each name you assign for compliance with
>         NAMEPREP normalization.

True, however most of use will likely just assign names using
digits and lower-case ASCII.  Since those are already normalized,
anything that generates based on this is OK.  Even if I allow
a user to enter part of the name, I can still just do tolower()
and use isalnum() to check that no other characters have been
added.

The place where one would have to run NAMEPREP is wherever
names are generated using characters outside ASCII.  At that
point, your procedure below must be used.

> This gets especially sticky as
> the folded cases and the multi-character cases are
> resolved.  Basically, I would expect that the name would
> get run through NAMEPREP's latest revision.  If it came
> out unchanged, you have selected a valid name.  If it
> came out changed, you have to either use a different name
> or use the normalized changed name.  In either of the
> mismatch cases, you must be careful to avoid duplications.
> 
> Bob Snively                        e-mail:    rsnively@brocade.com
> Brocade Communications Systems     phone:  408 487 8135
> 1745 Technology Drive
> San Jose, CA 95110

-- 
Mark A. Bakke
Cisco Systems
mbakke@cisco.com
763.398.1054