Le lundi 14 février 2011 à 14:28 +0100, Raphael Hertzog a écrit : > On Mon, 14 Feb 2011, Josselin Mouette wrote: > > There will be a default selected in gnome-session, that can be changed > > by user action. Without a session-wide default (any session manager can > > set one, of course) a random choice (possibly alphabetical) is used. > > How is that default communicated? An environment variable? A gconf key?
The defaults are set in /etc/gnome/defaults.list. This file is used in GNOME session since there is a XDG_DATA_DIRS indirectly pointing to it (set in a Xsession script). > > We are used to have much more terrible breakage in testing/unstable > > right after a release. If our concerns are now bugs wrt. setting the > > default browser, it must mean we are doing *great* :) > > Well, if we want to be serious about Constantly Usable Testing (and I wish > we do) I believe we should care about such things because they are far from > being details from an end-user perspective. Fully agreed. One of my concerns with CUT is that once such a change reaches testing without the other impacted packages, it can stay buggy for weeks, even months. > Filed #613381 to block glib2.0. Good. We can probably let it migrate once epiphany and iceweasel have been fixed; unless you want the configuration GUI, which means gnome-control-center 3.x - and if you really want to tie glib2.0 with it, I’ll let you deal with the RT fallback :) -- .''`. : :' : “You would need to ask a lawyer if you don't know `. `' that a handshake of course makes a valid contract.” `- -- J???rg Schilling -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/1297698733.8791.77.camel@meh