[libvirt] [libvirt-glib] Correct namespace prefix for GVirConfig symbols

[Christophe Fergeau]

On Tue, Jan 03, 2012 at 08:47:02PM +0200, Zeeshan Ali (Khattak) wrote:
> 
>   So FWIW I found the patch in Boxes that causes a problem with current API:
> 
> https://bugzilla.gnome.org/attachment.cgi?id=204140&action=edit
> 
>   Compiling that patch should lead you to this:
> 
> vm-configurator.c: In function ‘boxes_vm_configurator_set_video_config’:
> vm-configurator.c:713:62: error: ‘GVIR_TYPE_DOMAIN_VIDEO_MODEL’
> undeclared (first use in this function)
> vm-configurator.c:713:62: note: each undeclared identifier is reported
> only once for each function it appears in
> vm-configurator.c: In function ‘boxes_vm_configurator_set_sound_config’:
> vm-configurator.c:758:62: error: ‘GVIR_TYPE_DOMAIN_SOUND_MODEL’
> undeclared (first use in this function)
> vm-configurator.c: In function ‘boxes_vm_configurator_set_tablet_config’:
> vm-configurator.c:804:62: error: ‘GVIR_TYPE_DOMAIN_INPUT_BUS’
> undeclared (first use in this function)

Ah great, thanks! At first I thought your patch was just about making
things cleaner, so I was a bit surprised that you mentioned issues with
vala just now. And having a clear problem to think about really helps to
focus rather than " I promise you, I've seen some problem once, but I
forgot all the details"

The issue comes from typeof(GVirConfig.DomainVideoModel). Vala wants to
change this to GVIR_TYPE_CONFIG_DOMAIN_VIDEO_MODEL but it doesn't really
have a way of getting this right (I couldn't find an obvious way of getting
this name from the .gir) so it ends up with GVIR_TYPE_DOMAIN_VIDEO_MODEL
which doesn't exist. However, the .gir has:
    <enumeration name="DomainVideoModel"
                 glib:type-name="GVirConfigDomainVideoModel"
                 glib:get-type="gvir_config_domain_video_model_get_type"
                 c:type="GVirConfigDomainVideoModel">
My feeling is that there would be less magic involved on the Vala side if
it were to use the value of the glib:get-type attribute instead of trying
to build a GVIR_TYPE_DOMAIN_VIDEO_MODEL symbol, but it may have very good
reasons for doing things the way it does.

Then I'm not saying because of what is above, your patch is not needed,
it's just me trying to understand what kind of breakage we have because of
the unusual namespacing.

Christophe

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 198 bytes
Desc: not available
URL: <http://listman.redhat.com/archives/libvir-list/attachments/20120104/4cbeba21/attachment-0001.sig>