db-derby-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Kristian Waagan (JIRA)" <j...@apache.org>
Subject [jira] Commented: (DERBY-4702) Determine if the DRDA Layber B DSS extended total length field should carry a signed or unsigned integer
Date Thu, 17 Jun 2010 09:14:24 GMT

    [ https://issues.apache.org/jira/browse/DERBY-4702?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12879715#action_12879715

Kristian Waagan commented on DERBY-4702:

Comment copied from DERBY-1595 (:

 a) DRDA vol 3, page 300 (17/Jun/10 05:11 AM):
   "The length of the extended total length field must be a multiple of two bytes, and it
cannot be longer than necessary to express the length of the object's data."

 b) .../client/net/Request:
        // according to Jim and some tests perfomred on Lob data,
        // the extended length bytes are signed. Assume that
        // if this is the case for Lobs, it is the case for
        // all extended length scenarios.

  It sounds reasonable to expect that the code never sends negative lengths, but:
    - negative values may be reserved for a special purpose
    - the DRDA spec may have made a wrong turn in this area (i.e "wasting space" by using
signed instead of unsigned)
    - since the client only sends signed positive values that are small enough so that the
most significant bit isn't used, the difference between client and server doesn't cause problems
    - if the server is correct, our client is breaking the spec for a series of ranges of
LOB lengths (point a, the client uses 0x7FFFL, 0x7FFFFFFF and 0x7FFFFFFFFFFF). This doesn't
matter, as the server doesn't seem to enforce point a.
   (- btw, I haven't investigated if the extended length bytes are used when transferring
data from the server to the client)

  To move on with the reported issue, I have logged DERBY-4702 to track further discussion
of this topic.

> Determine if the DRDA Layber B DSS extended total length field should carry a signed
or unsigned integer
> --------------------------------------------------------------------------------------------------------
>                 Key: DERBY-4702
>                 URL: https://issues.apache.org/jira/browse/DERBY-4702
>             Project: Derby
>          Issue Type: Task
>          Components: Network Client, Network Server
>    Affects Versions:
>            Reporter: Kristian Waagan
>            Priority: Minor
> The client and the server disagrees on whether the extended total length field in the
DRDA protocol is signed or unsigned.
> A search in the DRDA specification (version 4) was fruitless.
> I don't think the current situation results in any practical problems, but it would be
nice to determine what the correct representation is and to make the client and the server
> This issue was brought up under DERBY-1595.

This message is automatically generated by JIRA.
You can reply to this email to add a comment to the issue online.

View raw message