Re: C++ Variant type for DBus object path

[Andrejs Hanins <andrejs hanins ubnt com>]

On 2015.04.11. 10:46, Kjell Ahlstedt wrote:
> Den 2015-04-10 15:09, Andrejs Hanins skrev:
> > Hi,
> >
> > On 2015.04.09. 14:01, Andrejs Hanins wrote:
> > > Hi Kjell,
> > >
> > > On 2015.04.08. 18:08, Kjell Ahlstedt wrote:
> > > > A reply, at long last! I have filed bug https://bugzilla.gnome.org/show_bug.cgi?id=747508 with a patch 
> > > > that I think fixes your problem.
> > > >
> > > > It would be fine if you could test the patch with your quite complicated type of Variant. Can you apply 
> > > > and test a patch without great trouble? 
> > > Thanks for the patch, I'll try it later and will let you know about the results.
> > >
> > > > If not, I can send you updated versions of the glibmm/variant.h and variant.cc files.
> > > No need, thanks, patch is enough for me.
> > I have applied the patch to hg and ccg files, but it does not cause real .h and .cc files to be 
> > regenerated by gmmproc. I even tried to delete the source files, but still no luck. There seems to be no 
> > automatic make rule for regeneration (bug?). Unfortunately, I failed to understand how to regenerate the 
> > source files in reasonable amount of time.
> >
> > Can you please help me how to regenerate the files under glib/glibmm or send regenerated variant.h/.cc 
> > files as you initially offered?
> You must run ./configure with the --enable-maintainer-mode option.
So easy, shame to me :) Ok, I can confirm that your patch definitely solves my case. I can now parse 
a{oa{sa{sv}}} signature by using Glib::ustring as a template parameter of Glib::Variant for the 'o'. Thank 
you Kjell! I suppose, this patch will be included in next Glibmm version 2.45 ?
I have also updated the Bugzilla ticket.
>