Activation: b-a-s problems.

Michael Meeks michael@ximian.com
Mon, 05 Jan 2004 13:41:07 +0000


Hi Mikael,

On Mon, 2004-01-05 at 12:13, Mikael Hallendal wrote:
> > 	IMHO the most difficult issue with activation is getting the scope
> > right; it transpires the world is more ugly than most would imagine; eg.
> > multi-language desktops - you have to fork a new service for (the
> > hundreds) of different combinations of LC_ALL, LC_MESSAGES, LC_CTYPE,
> > LC_IJUSTMADETHISUP, LANG etc. since there is no good way to propagate
> > multi-language contexts / get gettext to play ball (AFAIK), worse you
> > have to wrestle with the daft-return value of setlocale, and/or somehow
> > parse the stuff yourself, and then work out what is unique.
> 
> Do you have any examples on when this will be used? Do users run with
> different locales within the same session or is it just for fixing the
> problem when changing the locale within the session?

	The examples lurk in bugzilla, and various cursings from users; it
turns out that large chunks of technical non-English speaking people
like to use odd things like LC_ALL=en_US.UTF-8 LC_CTYPE=de_DE.UTF-8 (or
something) to get English UI, with German something else, and better
they expect it to work per application.

> Since the bus is per session (which is per display, right?) I don't see
> this as a problem. Or am I missing something?

	No idea; how is the 'per session' stuff implemented ? what if you ssh
in from elsewhere ? or does it tunnel over the X connection ? if it
tunnels over X - on which machine is the service started ? local or
remote ?

> If you ask the bus for a service it checks if it is running (which would
> be on the correct display since it has no idea about services running on
> other displays). If it is not, it's started. 

	I thought there were multiple busses with different scopes: per-system,
per-session, per-user, but perhaps I was mis-informed ? b-a-s was always
per-machine/user so we saw these problems, in retrospect it would
perhaps have been better to stick with the under-engineered X property
on display rather than the horribly over-engineered b-a-s approach.

	Regards,

		Michael.

-- 
 michael@ximian.com  <><, Pseudo Engineer, itinerant idiot