Re: [PATCH] some performance improvements

[Rodney Dawes <dobey novell com>]

Generally looks ok.

It looks like all of the code in gnome_settings_xrdb_init () is being
removed. Is there any particular reason to keep the function around, but
empty?

When does the window created in gnome_settings_xrdb_load () get
destroyed? Are we not leaking it? There are no calls to any unref or
destroy methods in the current code, so I presume we are, no?

Then again, I would also wonder if there is any particular reason to
keep xrdb merging around at all. How many applications on a normal
desktop actually need this stuff?

As for the 2.12 stuff, perhaps we should get some testing done with it
in HEAD first? Would be nice to really test not doing xrdg -merge at all
as well.

-- dobey


On Fri, 2005-10-14 at 15:41 +0200, Rodrigo Moya wrote:
> updated patch attached, which fixes the use of the GConfClient, found by
> Lorenzo.
> 
> On Fri, 2005-10-14 at 12:16 +0200, Rodrigo Moya wrote:
> > Hi
> > 
> > Attached patch includes:
> > 
> > * part of my performance improvements, essentially the lazy startup of
> > the screensaver and the typing break (1s gain as of Lorenzo's analysis
> > [1]), and the async wait of children started in
> > gnome_settings_daemon_spawn_with_input.
> > * patch from Erwann for http://bugzilla.gnome.org/show_bug.chi?id=314774
> > 
> > I think this should be also committed to the 2.12 branch, ok?
> > 
> > [1] http://www.gnome.org/~lcolitti/gnome-startup/analysis/