possible breakage in kpasswd with 0.5?

[Wynn Wilkes <wynnw@sco.com>]

Hello,

(Sorry if this is a duplicate)

With the new release of Heimdal 0.5, there appears to be a problem in the 
process_reply method in lib/krb5/changepw.c

On line 178, there's a check to make sure the priv_data.data does not overflow 
the reply buffer- except the check is >=, when it looks like it should just 
be >

We've experienced the problem when running against Active Directory as the 
changepw server- I have not verified it against the heimdal or mit servers. 
Changing the test to > fixes the problems we've had.

Is this a real bug, or is there something else wrong?

Thanks,
Wynn