Conference rocks::dec_edi

3110.0. "3780 P_MCDD script/profile question" by CSC32::R_GOLLEHON () Fri Apr 25 1997 21:04

    Hello,
    
    I have a customer who is getting duplicate data sent to MCI using the
    3780 gateway on 2.1C.  After looking at the scripts, I have a question.
    
    It would appear that when we re-run the connection the VAN realizes the 
    problem from the last run and resends it's data, but doesn't expect us
    to resend the data from the last connection (the result is we sent
    duplicate data).  
    
    IE:
    
    .
    .
    .
    write(sf,"TEXT DECEDI$P_MCDD_LOGIN.DAT")
    write(sf,"BRANCH NOT NRMEOF TO 201")
    write(sf,"##")
    write(sf,"RECEIVE DECEDI$P_MCDD_MAILBOX.DAT "+local$tft)
    write(sf,"BRANCH NOT NRMEOF TO 301")
    write(sf,"##")
    write(sf,"TEXT DECEDI$P_MCDD_XMIT.DAT "+local$tfb)
    write(sf,"BRANCH NOT NRMEOF TO 203")
    write(sf,"##")
    write(sf,"RECEIVE DECEDI$P_MCDD_STAT.DAT "+local$tft)
    write(sf,"BRANCH NOT NRMEOF TO 303")  ! we die here, end up exiting w/303
    write(sf,"##")
    write(sf,"999 LOG *** 3780Plus Job Completed Successfully ***")
    .
    .
    .
    write(sf,"303 LOG %%% 3780Plus Receive Stat Failed %%%")
    write(sf,"VOICE")
    write(sf,"QUIT 303")
    .
    .
    .
    if (stat .EQ. 303)
            log("*** P_MCDD303: 3780Plus Receive Stat Failed ***")
    !       Abnormal session termination; trash all work done
    !       For profile B140 session must terminate correctly for work to
    be accepte
    d
    !       VAN will resend data already received
            send_complete(FALSE)
            spawn("DELETE/NOCONFIRM
    DECEDI$3780:DECEDI$P_MCDD_NEW_MAIL.DAT;")
            exit FAIL
    endif
    
    
    Should the VAN be trashing the data we sent if we are using the proper
    profile and the session terminates abnormally?
    
                                                                           
    Thanks,
    
    -Robert

    Answered my own question...the answer is that MCI should trash the
    whole session...both data sent and received...however, in my customer's
    case MCI doesn't see a problem with the connection where as we
    do...namely, it appears we are getting an EOT before the STAT file is
    received which MCI says isn't terribly abnormal...I have a problem
    believing that though.
    
    Regards,
    
    -Robert