• possible emsi protocol bug in binkd

    From Benny Pedersen@2:230/0 to All on Tuesday, January 30, 2018 13:47:54
    Hello All!

    just in case, if qico runs in xinetd with -a auto, then binkp clients stale, this is a bug in binkd

    ----- msged.txt begins -----
    Date: 2018-01-29 08:01:20
    From: Benny Pedersen of 2:230/0
    To: Sergey Poziturin
    Subj: hotdoged stopped ...
    Attr: Snt Loc Scn
    Conf: HOTDOGED

    Hello Sergey!

    18 Oct 2017 09:14, Sergey Poziturin wrote to Benny Pedersen:

    no more possible for me to use it, it try to dial me, but no files
    transfered :/
    How did this happen?

    found the qico bug:

    /usr/sbin/qico -a auto
    /usr/sbin/qico -a binkp

    only the last one works

    so binkp protocol clients does not always work with autodetect in qico, so i loose emsi on same port as binkp

    ----- qico begins -----
    service binkp
    {
    disable = no
    protocol = tcp
    flags = IPv4
    port = 24554
    socket_type = stream
    wait = no
    user = xpoint
    server = /usr/sbin/qico
    server_args = -a binkp
    }
    ----- qico ends -----

    works :=)

    wonder if thats same problem i have with aftershook ?
    Hmm.

    hope its not the problem

    note 2:230/0 is now supporting ipv6


    ----- msged.txt ends -----

    Regards Benny

    ... there can only be one way of life, and it works :)

    --- Msged/LNX 6.2.0 (Linux/4.9.76-gentoo-r1 (i686))
    * Origin: I will always keep a PC running CPM 3.0 (2:230/0)
  • From Michael Dukelsky@2:5020/1042 to Benny Pedersen on Tuesday, January 30, 2018 16:09:30
    Hello Benny,

    Tuesday January 30 2018, Benny Pedersen wrote to All:

    just in case, if qico runs in xinetd with -a auto, then binkp clients stale, this is a bug in binkd
    ^^^^^^^^^^^^

    It is rather in qico.


    Michael

    ... node (at) f1042 (dot) ru
    --- GoldED+/LNX 1.1.5-b20170303
    * Origin: Moscow, Russia (2:5020/1042)
  • From Michiel van der Vlist@2:280/5555 to Benny Pedersen on Tuesday, January 30, 2018 14:46:39
    Hello Benny,

    On Tuesday January 30 2018 13:47, you wrote to All:

    just in case, if qico runs in xinetd with -a auto, then binkp clients stale, this is a bug in binkd

    No, it is not a bug. You are trying to make it do it was never designed for.

    so binkp protocol clients does not always work with autodetect in
    qico, so i loose emsi on same port as binkp

    Of course it does not work. Binkp was never designed to share a port. Like most
    protocols, it was designed to have a port of its own.

    The solution is easy: forget about EMSI. Binkp was optimized for use over IP, EMSI was not. Who needs EMSI when you have binkp?

    Cheers, Michiel

    ...I am not a member of the more is better club.

    --- GoldED+/W32-MSVC 1.1.5-b20170303
    * Origin: http://www.vlist.eu (2:280/5555)
  • From Kees van Eeten@2:280/5003.4 to Michiel van der Vlist on Tuesday, January 30, 2018 15:40:28
    Hello Michiel!

    30 Jan 18 14:46, you wrote to Benny Pedersen:

    just in case, if qico runs in xinetd with -a auto, then binkp clients
    stale, this is a bug in binkd

    MvdV> No, it is not a bug. You are trying to make it do it was never designed
    MvdV> for.

    so binkp protocol clients does not always work with autodetect in
    qico, so i loose emsi on same port as binkp

    MvdV> Of course it does not work. Binkp was never designed to share a port.
    MvdV> Like most protocols, it was designed to have a port of its own.

    MvdV> The solution is easy: forget about EMSI. Binkp was optimized for use
    MvdV> over IP, EMSI was not. Who needs EMSI when you have binkp?

    I think you are confusing the matter. The way I would implement the auto
    detect, is to look at the fingerprint of waht comes in first. Does it look
    like a binkp or like an EMSI handshake and switch accordingly.

    qico should probably just needs an upgrade in its protocol recognistion.

    I think we have had more or less the same with an other mailer, where the
    calling system had to make shure that a lower version of the binkp protocol
    should be used.

    But the workaround is there, don't use auto but diffrent port instead. ;))

    Kees

    --- GoldED+/LNX 1.1.5
    * Origin: As for me, all I know is that, I know nothing. (2:280/5003.4)
  • From Benny Pedersen@2:230/0 to Michael Dukelsky on Tuesday, January 30, 2018 15:41:40
    Hello Michael!

    30 Jan 2018 16:09, Michael Dukelsky wrote to Benny Pedersen:

    just in case, if qico runs in xinetd with -a auto, then binkp clients
    stale, this is a bug in binkd
    ^^^^^^^^^^^^

    It is rather in qico.

    binkd timeout ?

    binkd client to binkd server no problem, but that does imho not mean its not any problems in binkd client code

    i am happy if its only a qico problem, so others can ignore it :/


    Regards Benny

    ... there can only be one way of life, and it works :)

    --- Msged/LNX 6.2.0 (Linux/4.9.76-gentoo-r1 (i686))
    * Origin: I will always keep a PC running CPM 3.0 (2:230/0)
  • From Benny Pedersen@2:230/0 to Michiel van der Vlist on Tuesday, January 30, 2018 15:50:52
    Hello Michiel!

    30 Jan 2018 14:46, Michiel van der Vlist wrote to Benny Pedersen:

    MvdV> No, it is not a bug. You are trying to make it do it was never
    MvdV> designed for.

    i will not take that debat here, you fail aswell here

    MvdV> Of course it does not work. Binkp was never designed to share a
    MvdV> port. Like most protocols, it was designed to have a port of its
    MvdV> own.

    correct binkp was never designed for other then binkp, true

    but qico have always supported multiple protocols on same port

    MvdV> The solution is easy: forget about EMSI. Binkp was optimized for use
    MvdV> over IP, EMSI was not. Who needs EMSI when you have binkp?

    with that rule i could easely take out AAAA in my fidonet hostname

    be carefull


    Regards Benny

    ... there can only be one way of life, and it works :)

    --- Msged/LNX 6.2.0 (Linux/4.9.76-gentoo-r1 (i686))
    * Origin: I will always keep a PC running CPM 3.0 (2:230/0)
  • From Michael Dukelsky@2:5020/1042 to Benny Pedersen on Tuesday, January 30, 2018 18:45:54
    Hello Benny,

    Tuesday January 30 2018, Benny Pedersen wrote to Michael Dukelsky:

    just in case, if qico runs in xinetd with -a auto, then binkp
    clients stale, this is a bug in binkd
    ^^^^^^^^^^^^
    It is rather in qico.

    binkd timeout ?

    binkd client to binkd server no problem, but that does imho not mean
    its not any problems in binkd client code

    i am happy if its only a qico problem, so others can ignore it :/

    If the same problem is with other binkp clients (for example Hotdoged, Aftershock) and qico server, then it is rather a problem in qico, isn't it?

    Michael

    ... node (at) f1042 (dot) ru
    --- GoldED+/LNX 1.1.5-b20170303
    * Origin: Moscow, Russia (2:5020/1042)
  • From Benny Pedersen@2:230/0 to Kees van Eeten on Tuesday, January 30, 2018 19:09:00
    Hello Kees!

    30 Jan 2018 15:40, Kees van Eeten wrote to Michiel van der Vlist:

    But the workaround is there, don't use auto but diffrent port
    instead. ;))

    and its still not proved its a qico bug, i just showed a working work around the problem

    i could try bforce if it compiles, or ifcico, dont know if it matter


    Regards Benny

    ... there can only be one way of life, and it works :)

    --- Msged/LNX 6.2.0 (Linux/4.9.76-gentoo-r1 (i686))
    * Origin: I will always keep a PC running CPM 3.0 (2:230/0)
  • From Benny Pedersen@2:230/0 to Michael Dukelsky on Tuesday, January 30, 2018 19:11:32
    Hello Michael!

    30 Jan 2018 18:45, Michael Dukelsky wrote to Benny Pedersen:

    If the same problem is with other binkp clients (for example Hotdoged, Aftershock) and qico server, then it is rather a problem in qico,
    isn't it?

    if binkd stale on that fail in qico, then its binkd bug aswell as in qico itself

    if it just timeout with that bug in qico, there is no bug in binkd


    Regards Benny

    ... there can only be one way of life, and it works :)

    --- Msged/LNX 6.2.0 (Linux/4.9.76-gentoo-r1 (i686))
    * Origin: I will always keep a PC running CPM 3.0 (2:230/0)
  • From Kees van Eeten@2:280/5003.4 to Benny Pedersen on Tuesday, January 30, 2018 19:42:10
    Hello Benny!

    30 Jan 18 19:09, you wrote to me:

    But the workaround is there, don't use auto but diffrent port
    instead. ;))

    and its still not proved its a qico bug, i just showed a working work around the problem

    Indeed.

    i could try bforce if it compiles, or ifcico, dont know if it matter

    I do not know about bforce, ifcico does not support binkp and mbcico from
    the mbse package cannot distinguish between binkp and emsi.
    A different paramater is given from inetd or xinetd and two portnumbers are
    used. Neither ifcico nor mbcico can listen on a port directly, the inet demon
    is always needed.

    Kees

    --- GoldED+/LNX 1.1.5
    * Origin: As for me, all I know is that, I know nothing. (2:280/5003.4)