hp at redhat.com
Mon May 8 18:00:17 PDT 2006
Ralf Habacker wrote:
> I recognized that some core tools like dbus-monitor uses the glib
> mainloop. What does this means relating to the binding split from the
> core ?
> If glib will be a requirement for the core library, than on windows glib
> will be a requirement for dbus, which would not be good at all.
> If glib will not be a requirement, a dbus only mainloop has to be
> implemented for unix and windows.
I don't really consider dbus-monitor a "core tool", it's just a
debugging hack more or less. There's already a crappy main loop
abstraction used by dbus-daemon that dbus-monitor could use (or for that
matter dbus-monitor could just be blocking, I don't know why it isn't,
I'm surprised it isn't really ... don't have the source handy to look)
Anyhow, I would oppose public API offering a "dbus main loop" (for
obvious reasons) but using the existing dbus main loop for stuff that
ships with dbus is of course feasible.
More information about the dbus