Re: [SR-Users] auth_challenge() and async

To be clear, the two t_newtran() calls in the main request_route were naturally a typo. :-) Only the first one was intended.

On Dec 12, 2022, at 3:45 PM, Alex Balashov abalashov@evaristesys.com wrote:

Hi,

Is there any good solution for the scenario of an auth_challenge() with two separate transaction suspensions?

To clarify (simplified):

request_route {
    ...

    t_newtran();

    if(!is_present_hf("Authorization") && !is_present_hf("Proxy-Authorization")) {
       auth_challenge("$fd", "1");
       exit;
    }

    # TM suspend/continue
    t_newtran();

    async credentials query("RESUME_AUTH");
}

route[RESUME_AUTH] {
    # Credentials received into PVs.

    if(!pv_auth_check("...")) {
        auth_challenge("$fd", "1");
        exit;
    }

    # Create transaction shell if not exists already from auth query.
    if(!t_lookup_request())
       t_newtran();

    # TM suspend/continue
    async route query("RESUME_ROUTING");
}

route[RESUME_ROUTING] {
    # Unmarshal etc.

    t_relay(); # etc.
}

What happens here is that the first auth_challenge() results in a retransmission of its 407 challenge without absorbing the negative ACK. This is despite the ostensibly stateful behaviour of auth_challenge() without "force_stateless_reply" enabled.

In this case, it eclipses the second, unrelated 407 challenge from the subsequent routing query (407 challenge with +1 CSeq):

<Screenshot 2022-12-12 at 3.44.15 PM.jpeg> -- Alex

-- Alex Balashov | Principal | Evariste Systems LLC

Tel: +1-706-510-6800 / +1-800-250-5920 (toll-free) Web: http://www.evaristesys.com/, http://www.csrpswitch.com/