Re: libeggdbus v. libdbus-glib

[Colin Walters <walters verbum org>]

On Fri, Sep 18, 2009 at 5:08 PM, David Zeuthen <david fubar dk> wrote:
> On Fri, 2009-09-18 at 15:04 +0000, Colin Walters wrote:
>> On Thu, Sep 17, 2009 at 1:58 PM, David Zeuthen <david fubar dk> wrote:
>> >
>> > Yes, EggDBus only uses the mainloop integration from dbus-glib. Yes,
>> > there's a reason for why this is so. Yes, the explanation is in the
>> > archives.
>>
>> Can we stick just the mainloop integration in Gio?  I'd be happy to
>> change dbus-glib to use that; how exactly it should work is an open
>> question; we could have #define GIO_USE_UNSTABLE_API or require people
>> to dlsym().  Other ideas welcome.
>
> I'd rather do it the same way we have gio-unix-2.0. E.g.
>
>  - provide a gio-libdbus1-2.0.pc file
>  - put headers in /usr/include/dbus-1.0/gio-libdbus1-2.0/gio/*

You know actually the other way around with this is to have extension
points in libdbus, so gio (or for that matter, dbus-glib) could drop a
.so somewhere with the mainloop code, and then you'd just have
dbus_bus_get_with_mainloop (DBUS_BUS_SESSION, "glib", &error) or
something.

> I also think we should be a bit more ambitious than just providing
> mainloop integration. Things like

True, but the mainloop is by far the biggest problem.  Well that and
some sort of arbitration for the message filter.

> I guess I should just finish the gdbus-standalone stuff and then we can
> either put it into libgio (for 2.30) or I can release it under another
> prefix and get on with my life...

Hmmm not sure what gdbus-standalone entails.