Re: get_preferred_height_vfunc for Gtk::CellRenderer seems to have no effect on 3.14 to the mailing list gtkmm-list

[Gaute Hope <eg gaute vetsj com>]

Excerpts from Krzesimir Nowak's message of October 28, 2014 13:40:
> 2014-10-28 13:22 GMT+01:00 Gaute Hope <eg gaute vetsj com>:
> > It turns out this was only a problem on my mac book pro with a retina
> > screen (hidpi). When I overloaded the function:
> >
> >   get_request_mode_vfunc and return CONSTANT_SIZE or HEIGHT_FOR_WIDTH
> >
> > and also wrap:
> >
> >   get_preferred_width_vfunc (...)
> >
> > things work out. The return value of the latter functions seems to be
> > irrelevant. But if I do not wrap it (and set the values to a positive
> > int) I get an assertion failure upon runtime.
> >
> > Is this expected behaviour? If so, I could not find any documentation.
>
> gtkmm is just a wrapper and by itself usually does not add any
> policies, but instead it follows ones imposed by wrapped C library. In
> this case it is GTK+. It is always good to consult the wrapped library
> docs as a fallback. It has a section about geometry management[1].
> Should be helpful in your case.
>
> [1] https://developer.gnome.org/gtk3/stable/GtkWidget.html#geometry-management

Thanks for the pointer, that is useful. But it does not explain why this
is an issue with a hidpi screen. It is probably a GTK+ issue though, and
not specific to gtkmm.

Regards, Gaute