Re: DPI in intel modesetting driver

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Le jeudi 04 janvier 2007 à 12:05 -0500, Adam Jackson a écrit :
> On Thu, 2007-01-04 at 17:25 +0100, Nicolas Mailhot wrote:
> 
> > - OLPC has to dynamically adjust GNOME dpi when screen mode switches (but
> > the scaling preferences of the user didn't change!)
> 
> This doesn't actually happen in current OLPC builds.  Whacking the
> grayscale button doesn't change font sizes.

I doesn't because IIRC in the background a process receives a dbus event
and changes the value of the gconf dpi key to compensate for pixel
density change. Because since this key aggregates hardware factor and
user scaling preferences it is not resilient to hardware changes.

> > The "scaling factor" is the excuse GNOME people have found to avoid fixing
> > this. "it's not a real value" => "we don't have to compute it correctly"
> 
> But considering it a scaling factor is _correct_.

Current GNOME dpi is hardware density × scaling factor. The two terms
need separation to be hardware-changes-agnostic

> If it means other
> things besides scale factor, then that needs deconflation within Gnome
> itself.
> 
> I'm aware that probably requires fixing Gnome, but there's certainly no
> fixing this within core X.

The problem is certainly not xorg-size but 100% GNOME-size

> Mangling DisplaySize in the config file is not a fix.

I'm not advocating putting a false DisplaySize value in the config file
(that would only continue the current mess). I'm advocating having the
GNOME xorg setup app let users put the real DisplaySize there if xorg
misdetected it, and replace "GNOME DPI" with an actual scaling factor
(in %) applied over the dynamically computed xorg dpi. (Two terms for
two different things). With maybe a floor value to avoid feeding
freetype scaling we know won't permit correct font rendering.

Not the current mess where one false value (wrong xorg computation) is
replaced by another false value (random user-provided "dpi")

Regards,

-- 
Nicolas Mailhot

Attachment: signature.asc
Description: Ceci est une partie de message =?ISO-8859-1?Q?num=E9riquement?= =?ISO-8859-1?Q?_sign=E9e?=

-- 
fedora-devel-list mailing list
fedora-devel-list@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/fedora-devel-list

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Index of Archives]     [Fedora Announce]     [Fedora Kernel]     [Fedora Testing]     [Fedora Formulas]     [Fedora PHP Devel]     [Kernel Development]     [Fedora Legacy]     [Fedora Maintainers]     [Fedora Desktop]     [PAM]     [Red Hat Development]     [Gimp]     [Yosemite News]
  Powered by Linux