Re: Problem with GUPnP root device / Question regarding GUPnP deinitialization

[Jens Georg <mail jensge org>]

On Di, 2011-06-28 at 11:28 +0200, Alexander Kirschner wrote:
> Hi all,
> 
>  
> 
> we are implementing a software tool using GUPnP 0.14.1-1. Our software
> contains a UPnP control point as well as a UPnP root device and is
> based on Qt 4.7.1. Right now we are facing two major issues...

Nice to hear that you opted to use GUPnP in Qt. :)

> At first we gave the internal webserver some time to host the files
> before accessing them (assuming GUPNP uses it's own webserver and does
> not access the XML-files directly from the file-system)... but this
> strategy did not solve the problem. Right now we are starting to debug
> glib because the output does not give us any hints. Up to now our
> debugging revealed that inside" gupnp_service_info_get_context()" the
> method "g_object_new()" creates somehow broken instances of
> GUPnPServiceInfo; sometimes with illegal values of GType, sometimes
> with memory allocation errors. Does anyone know how to deal with this
> problem?

Valgrind. Sounds like memory corruption. Also, which version of GSSDP do
you use with GUPnP?

> 2. Our second issue: In our application the GUPNP part needs often to
> be restarted (initialized / deinitialized). Because of this we have a
> question regarding deinitialization of GUPNP structures: The sample
> code provided on the web simply calls g_object_unref on the
> GUPnPContext; we added calls to g_signal_handler_disconnect for all
> signal-handlers we registered during the initialization phase. Is
> there anything more we should take care of?

If g_object_unref leads to destruction of the object (i.e. frees the
last reference to it), then there's no need to disconnect the signals.

Additionally a GUPnPContext can be set active/inactive. I'm not sure
what you mean by initializing/deinitializing.