Why sometimes the signal/method call can't be sent successfully

[Edison_lin-林俊翰]

Hi,
I add a dbus_bus_request_name(dbconn, THE_NAME_TO_REQUEST, DBUS_NAME_FLAG_REPLACE_EXISTING, &dberr) in the sender code
, and it seems that the sending request can be received by the dbus-daemon and passed to receiver every time. Why is that ? 
By the way, I first use the same name as the receiver(running on the background) has requested, and both are set with the flag DBUS_NAME_FLAG_REPLACE_EXISTING only.
It still works. Why is that ? Shouldn't any confliction occur if I didn't apply DBUS_NAME_FLAG_ALLOW_REPLACEMENT ?

BR,
Edison Lin
  ----- Original Message ----- 
  From: Edison_lin-林俊翰 
  To: Naveen Verma 
  Cc: dbus at lists.freedesktop.org 
  Sent: Wednesday, May 27, 2009 3:29 PM
  Subject: Re: Why sometimes the signal/method call can't be sent successfully


  Hi,
  I find another case that dbus-daemon doesn't notice the sending request at all.

  BR,
  Edison Lin
    ----- Original Message ----- 
    From: Edison_lin-林俊翰 
    To: Naveen Verma 
    Cc: dbus at lists.freedesktop.org 
    Sent: Wednesday, May 27, 2009 11:35 AM
    Subject: Re: Why sometimes the signal/method call can't be sent successfully


    Hi,
    The attached files are the logs after open verbose mode. After comparing these two logs I found one place different : After removing the message from outgoing queue, for failed case the dispatch status always shows complete, as for the successful case there's always data remains. You could find this after the verbose message "_dbus_connection_send_preallocated_and_unlock middle"

    For your reference,
    BR,
    Edison

    ----- Original Message ----- 
      From: Edison_lin-林俊翰 
      To: Naveen Verma 
      Cc: dbus at lists.freedesktop.org 
      Sent: Tuesday, May 26, 2009 8:23 PM
      Subject: Re: Why sometimes the signal/method call can't be sent successfully


      Hi,
      I think it's probably not my sender's fault, because sometimes the sending request seems to disappear
      through observing dbus-monitor even if I use the following command:
      dbus-send --session --type=signal --dest=com.wiley.test / com.wiley.test.TestSignal string::"rock"
      Any idea is appreciated !

      BR,
      Edison Lin
        ----- Original Message ----- 
        From: Edison_lin-林俊翰 
        To: Naveen Verma 
        Cc: dbus at lists.freedesktop.org 
        Sent: Tuesday, May 26, 2009 6:33 PM
        Subject: Re: Why sometimes the signal/method call can't be sent successfully


        Hi Naveen,
        I'm looking at session bus, and I'm using low level C API to do this. The attached is the source code of the sender, do I miss something ?

        Appreciate,
        Edison Lin
          ----- Original Message ----- 
          From: Naveen Verma 
          To: Edison_Lin-林俊翰 
          Cc: dbus at lists.freedesktop.org 
          Sent: Tuesday, May 26, 2009 6:02 PM
          Subject: Re: Why sometimes the signal/method call can't be sent successfully






            Hi,
            I used the dbus-monitor to monitor the status of message bus, and I found that sometimes after sending a signal/method call and flush the connection, nothing happened(I meant the information about the sending action) in the description produced from dbus-monitor. After several times of retries then it shows the sending message, at the same time the receiver got the request. I'm wondering the reason causing this ? Is it a usual case so that I have to establish a mechanism to ensure the delivering of request ?

          -This is strange, I am also using dbu-monitor quit a lot but never face this issue. Are you looking at system bus or session bus? 


            Appreciate,
            Edison Lin

          -Br
          Naveen
           


            _______________________________________________
            dbus mailing list
            dbus at lists.freedesktop.org
            http://lists.freedesktop.org/mailman/listinfo/dbus



-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.freedesktop.org/archives/dbus/attachments/20090527/16ecc869/attachment.htm