[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Is there a reason why GNUnet forcefully sends a message of type 6?
From: |
Schanzenbach, Martin |
Subject: |
Re: Is there a reason why GNUnet forcefully sends a message of type 6? |
Date: |
Sat, 23 May 2020 09:58:18 +0200 |
Hi,
if your service is not actually a service (that serves clients) maybe you should
try making it a program instead? Look at src/rest/gnunet-rest-server.c
OTOH if you do have clients then you _must_ to call client_continue which will
send messages of type 6 as this is part of the GNUnet client/server IPC logic.
This does not really have to do anything with the other GNUnet services but
with how services signal to clients (=programs using the API of the service)
that they may "continue" with sending.
BR
> On 22. May 2020, at 23:52, Alessio Vanni <address@hidden> wrote:
>
> Hello,
>
> I wrote a small service which is meant to be part of a bigger project.
> While I was checking stdout to see if everything worked correctly, I was
> notified that there was no handle for a message of type 6 and size 4
> (and subsequently that the handle in the service didn't call
> `GNUNET_SERVICE_client_continue' before the timeout.)
>
> Due to the fact that this specific service is not meant to talk with
> other GNUnet services (it simply uses GNUnet's utilites to talk to
> services that actually use GNUnet and to operate on the received data),
> I enumerated the types of message starting from 0, instead of checking
> which numbers were "free" (i.e. not used by GNUnet itself). I only have
> 3 message types so far, so receiving a message of type 6 is something
> that can't happen.
>
> Looking at the source code, it appears that GNUnet forcefully adds a
> handler for this message type upon starting the service in
> `GNUNET_SERVICE_start', but for some reason when starting my service
> this handle was not added, yet the message was sent to the service.
>
> I can't find where the message is sent in the code, so my research stops
> here, but the question remains: why is GNUnet asking to send an AGPL
> URL, even though the application has nothing to do with the GNUnet core
> itself and might even be licensed differently?
>
> In particular, this behaviour is not documented at all as far as I know,
> so if by disgrace I had more than 6 message types I would be here trying
> to debug a piece of code that would otherwise have no problems.
>
> Thanks,
> A.V.
>
signature.asc
Description: Message signed with OpenPGP
- Is there a reason why GNUnet forcefully sends a message of type 6?, Alessio Vanni, 2020/05/22
- Re: Is there a reason why GNUnet forcefully sends a message of type 6?, Christian Grothoff, 2020/05/22
- Re: Is there a reason why GNUnet forcefully sends a message of type 6?, Alessio Vanni, 2020/05/22
- Re: Is there a reason why GNUnet forcefully sends a message of type 6?, Christian Grothoff, 2020/05/23
- Re: Is there a reason why GNUnet forcefully sends a message of type 6?, Alessio Vanni, 2020/05/23
- Re: Is there a reason why GNUnet forcefully sends a message of type 6?, Christian Grothoff, 2020/05/23
- Re: Is there a reason why GNUnet forcefully sends a message of type 6?, Alessio Vanni, 2020/05/25
- Re: Is there a reason why GNUnet forcefully sends a message of type 6?, Christian Grothoff, 2020/05/25
- Re: Is there a reason why GNUnet forcefully sends a message of type 6?, Alessio Vanni, 2020/05/27
- Re: Is there a reason why GNUnet forcefully sends a message of type 6?, Christian Grothoff, 2020/05/27
Re: Is there a reason why GNUnet forcefully sends a message of type 6?,
Schanzenbach, Martin <=