I just finished reviewing our SMTP code, and we are using the adjusted value for the comparison. The error message uses the raw value because rejection messages are often displayed to a user in some way, either by being included in a DSN bounce message or being shown by a client like Thunderbird.
This behavior is consistent with other protocols and webmail where we use the adjusted limit when checking the size of the message, but use the raw limit in messages that might be displayed to a user.
Andrew Barker
Software Developer
SmarterTools Inc.
www.smartertools.com