SORT BY:

LIST ORDER
THREAD
AUTHOR
SUBJECT


SEARCH

IPS HOME


    [Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

    RE: iSCSI: numeric-range format question (pedantic)




    WADR I do not see an interoperability  issue - you just implement both anyway and  you admit for range both (it is not even costlier).

    Julo


    kevin_lemay@agilent.com

    06/02/03 02:11

    To
    Julian Satran/Haifa/IBM@IBMIL, kevin_lemay@agilent.com
    cc
    blandry@crossroads.com, ips@ece.cmu.edu, owner-ips@ece.cmu.edu
    Subject
    RE: iSCSI: numeric-range format question (pedantic)





    Julian,

    I guess now I question whether clearer wording is needed to specify that both are acceptable ( or not)

    I reasoned, that a range must be offered and a single value returned in reply. It is easier to check for the lack of the "~" to denote a single value.

    If 4096~4096 is allowed, then I must decode both values and ensure that they are indeed the same.

    This is not a big issue, but I would prefer that the spec pick one way or the other. I see no value in allowing both.

    Right now, we have an interoperabilty issue.

    Kevin
    -----Original Message-----
    From: Julian Satran [mailto:Julian_Satran@il.ibm.com]
    Sent: Wednesday, February 05, 2003 4:04 PM
    To: kevin_lemay@agilent.com
    Cc: blandry@crossroads.com; ips@ece.cmu.edu; owner-ips@ece.cmu.edu
    Subject: RE: iSCSI: numeric-range format question (pedantic)



    obviously 4096~4096 is also valid (but redundant). Julo



    kevin_lemay@agilent.com
    Sent by: owner-ips@ece.cmu.edu


    06/02/03 01:48


    To
    blandry@crossroads.com, ips@ece.cmu.edu

    cc

    Subject
    RE: iSCSI: numeric-range format question (pedantic)

                   




    Buck,

    I coded it this way....

    OFMarkInt=4096

    If you don't fine the "~" delimiter then it is a single value.

    I would interpret 4096~4096 as a range with only one acceptable value.

    Kevin

    -----Original Message-----
    From: Buck Landry [mailto:blandry@crossroads.com]
    Sent: Wednesday, February 05, 2003 3:35 PM
    To: ips@ece.cmu.edu
    Subject: iSCSI: numeric-range format question (pedantic)


    Is it draft 20's intent that any numeric-range may be represented by a single numerical-value, if that value is the only one acceptable?

    I ask because:
    -- the definition for numeric-range does not mention this (5.1)
    -- OFMarkInt and IFMarkInt are defined as <numeric-range> for offers (A.3.2)
    -- .. but the description for these keys says "In case it only wants a specific value, only a single value has to be specified".

    There are three conclusions I might make:
    a) When the draft says "only a single value has to be specified", it means only that (for instance) 4096~4096 might be specified.
    b) x is valid shorthand for x~x, but only for OFMarkInt and IFMarkInt
    c) x is valid shorthand for x~x, for any theoretical numeric-range

    The safest choice is obviously c).  If the intent was actually a), would somebody please speak up.

    Thanks, --buck





Home

Last updated: Thu Feb 06 10:19:12 2003
12294 messages in chronological order