On Sat, 24 Mar 2012 02:30:49 -0300 Lucas De Marchi
<lucas.demar...@profusion.mobi> said:

> On Fri, Mar 23, 2012 at 9:59 PM, Carsten Haitzler <ras...@rasterman.com>
> wrote:
> > On Fri, 23 Mar 2012 15:40:10 -0300 Lucas De Marchi
> > <lucas.demar...@profusion.mobi> said:
> >
> >> On Fri, Mar 23, 2012 at 2:23 AM, Enlightenment SVN
> >> <no-re...@enlightenment.org> wrote:
> >> > Log:
> >> > add -cleanvt
> >> >
> >> >
> >> >
> >> > Author:       raster
> >> > Date:         2012-03-22 22:23:44 -0700 (Thu, 22 Mar 2012)
> >> > New Revision: 69575
> >> > Trac:         http://trac.enlightenment.org/e/changeset/69575
> >> >
> >> > Modified:
> >> >  trunk/exquisite/ChangeLog trunk/exquisite/src/bin/main.c
> >> >
> >> > Modified: trunk/exquisite/ChangeLog
> >> > ===================================================================
> >> > --- trunk/exquisite/ChangeLog   2012-03-23 03:46:36 UTC (rev 69574)
> >> > +++ trunk/exquisite/ChangeLog   2012-03-23 05:23:44 UTC (rev 69575)
> >> > @@ -0,0 +1,10 @@
> >> > +2012-02-10  Carsten Haitzler (The Rasterman)
> >> > +
> >> > +        1.0.0 release
> >> > +
> >> > +2012-03-23  Carsten Haitzler (The Rasterman)
> >> > +
> >> > +        * add -cleanvt to enable clean "fade out" animation before
> >> > giving
> >> > +        up the VT.
> >>
> >> IMO it's much better to tell X to run on the same VT (i.e. the first
> >> one), start X with "background none" and use the same background. Then
> >> you have a flicker free boot and user has the feeling that boot was
> >> much faster then it really was.
> >
> > x will not be able to own that vt because exquisite already owns it. also
> 
> That's why you still ask it to "exit gracefully"... that means: keep
> it in gfx mode. X with no background, in the same VT, will just use
> the last image left by the splash (given your gfx driver supports
> that).
> 
> > remember that the driver will most likely reconfigure the gfx chip -
> > possibly
> 
> I'm talking about drivers with KMS (intel here). It sure doesn't work
> with proprietary nvidia driver since it doesn't support kernel mode
> setting.
> 
> > with a different resolution (this definitely happens on  nvidia drivers in
> > many cases as the kernels vesa support often can't detect the correct
> > resolution).
> 
> If you have a driver with KMS, don't worry about that.
> 
> > i'm actually using fifos mostly to cut down the overhead of writing status
> > updates as it needs much simpler tools. :)
> 
> It's almost the same work to support abstract sockets.... but it's ok,
> it was already written.
> 
> >
> >> "2. You need a /tmp directory for the UNIX socket that is used to talk to
> >> exquisite. The exquisite-write utility sends messages to exquisite via a
> >> UNIX domain socket created in /tmp. This needs to be writable for
> >> exquisite to start."
> >>
> >> This implies that you need to have /tmp mounted or a writable / early
> >> on boot sequence. If you use abstract sockets you can start exquisite
> >> earlier. With dietsplash I use this and run the run the splash even
> >> before handing off to the init process.
> >
> > exqusitie needs a lot of infra - eina, ecore, evas, edje, an edje file,
> > embryo etc. etc. to get up and running. these in turn need modules.
> > using /tmp isn't a tall order :) also i think i wrote this before abstract
> > socket support was in
> 
> If you want your splash to appear asap, you better start it early.
> Supposing we are using KMS, that means:

sure - but it needs pretty much a large part of the fs up - so /tmp.. as a
ramdisk.. is minimal extra. :) adding abstract socket support is a few lines of
code because ecore-con already supports it - but it just never was written to
begin with.

> 1. bootloader set up gfx and put an img as background
> 2. kernel comes up, see the gfx is in the right setup and doesn't touch it
> 3. splash comes up, start animations and give control to init
> 4. We are about to execute X: ask bootsplash to exit nicely
> 5. X comes up and use what the bootsplash left
> 
> The hardest part is (1). The others can be done for drivers that have
> KMS, then you have a flicker-free boot.

if you want to do that - it's moot. you dont need the -cleanvt option. by
default if someone wants the vt exquisite instantly exits. you have choices.
exquisite cant go assuming you have a full kms friendly setup. it supports it
by behavior already,

> > ecore_con or very early after it was added. i did it then never touched it
> > again as it worked.
> 
> Ok. We could remove that limitation: using abstract sockets is veeery
> simple, no need to support that in ecore_con (example of a client
> using it:
> http://git.profusion.mobi/cgit.cgi/dietsplash.git/tree/src/dietsplashctl.c)

ecore_con already supports abstract sockets - has for a while, i just wrote
this early on - either before ecore-con did, or just after it was added.

-- 
------------- Codito, ergo sum - "I code, therefore I am" --------------
The Rasterman (Carsten Haitzler)    ras...@rasterman.com


------------------------------------------------------------------------------
This SF email is sponsosred by:
Try Windows Azure free for 90 days Click Here 
http://p.sf.net/sfu/sfd2d-msazure
_______________________________________________
enlightenment-devel mailing list
enlightenment-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/enlightenment-devel

Reply via email to