Re: GTK_MODULE (G)Option breakage & init ordering ...
- From: michael meeks <michael meeks novell com>
- To: James Henstridge <james jamesh id au>
- Cc: Mark McLoughlin <markmc redhat com>, Owen Taylor <otaylor redhat com>, accessibility mailing list <gnome-accessibility-devel gnome org>, Bill Haneman <bill haneman sun com>, bonobo <gnome-components-list gnome org>, Gtk Hackers <gtk-devel-list gnome org>
- Subject: Re: GTK_MODULE (G)Option breakage & init ordering ...
- Date: Wed, 08 Feb 2006 10:10:12 +0000
On Wed, 2006-02-08 at 18:01 +0800, James Henstridge wrote:
> > :-) Either way - there is a workaround now; and as Federico says prolly
> >going via the environment is a more robust & cleaner solution.
> The only issue with using the environment is that it is inherited by
> child processes. This might cause a program to think it is being
> invoked as a factory when it isn't.
Sure - of course, there's a gotcha for the unwary ;-) assuming though
that people init bonobo early; we can clobber the environment variable
after capturing it though.
It would be nasty surprise OTOH to have IOR:1010... arriving
unexpectedly in random streams of early-spawned children ;-)
michael meeks novell com <><, Pseudo Engineer, itinerant idiot
] [Thread Prev