On Fri, Jan 16, 2009 at 01:35:37PM +0100, Daniel Veillard wrote:
On Thu, Jan 15, 2009 at 05:18:40PM +0000, Richard W.M. Jones wrote:
> On Tue, Jan 13, 2009 at 05:43:08PM +0000, Daniel P. Berrange wrote:
> > With the domain events code, the callbacks triggered upon events get given
> > a virDomainPtr object instance. In many cases it'd be desirable to grab
> > this object and keep it in your app code. Unfortunately it is free'd the
> > moment the callback finishes executing.
> >
> > When allowing multiple threads to access a single virConnectPtr object
> > it is neccessary to ensure no thread releases it (virConnectCLose) while
> > another thread is still using it.
> >
> > The way to address both of these problems is to allow an application to
> > take an explicit reference on the object in question. So this patch
> > exposes methods to allow an app to increment the ref count on all our
> > public objects. To release the ref count, the existing virConectClose/
> > virDOmainFree, etc methods suffice
> >
> > include/libvirt/libvirt.h | 6 +
> > include/libvirt/libvirt.h.in | 6 +
> > src/libvirt.c | 183
+++++++++++++++++++++++++++++++++++++++++++
> > src/libvirt_public.syms | 12 ++
> > 4 files changed, 206 insertions(+), 1 deletion(-)
>
> Poor man's garbage collection ... +1.
Yeah it's unfortunate we have to add this at the API level, but
unfortunately there is no way around.
I wonder if we could resurrect some of the deprecated fields of
error structure now, seems to me that would be possible now.
Patch looks fine, +1
Comitted this patch
Daniel
--
|: Red Hat, Engineering, London -o-
http://people.redhat.com/berrange/ :|
|:
http://libvirt.org -o-
http://virt-manager.org -o-
http://ovirt.org :|
|:
http://autobuild.org -o-
http://search.cpan.org/~danberr/ :|
|: GnuPG: 7D3B9505 -o- F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 :|