Re: [gpm] Control of generic backlight interface not working

[Richard Hughes <hughsient gmail com>]

On Mon, 2008-11-03 at 12:27 +0000, Rui Tiago Cação Matos wrote:
> I wonder what is the plan to support hardware that doesn't expose this
> feature through xrandr but can otherwise set the backlight brightness
> in software. I'm sure this change was done with the betterment of the
> free desktop architecture in mind but there is a line to draw
> somewhere about what constitutes an acceptable regression.

Most hardware is now exposing a kernel backlight device (or devices) and
also an XBACKLIGHT interface.

When we are in X, it makes sense to use XBACKLIGHT rather than poking at
the hardware, as we really want to avoid going behind the hardwares'
back, especially on Intel cards.

If you don't have XBACKLIGHT support, g-p-m will use the BIOS as before.
What it looks like here is that your xrandr interface is broken, and
your BIOS interface is not.

The logic is thus (src/gpm-brightness.c):

	brightness->priv->xrandr = gpm_brightness_xrandr_new ();
	if (gpm_brightness_xrandr_has_hw (brightness->priv->xrandr))
		brightness->priv->use_xrandr = TRUE;
	brightness->priv->hal = gpm_brightness_hal_new ();
	if (gpm_brightness_hal_has_hw (brightness->priv->hal))
		brightness->priv->use_hal = TRUE;
	/* we want to default to only use XRANDR if available, as some hardware can use either */
	if (brightness->priv->use_xrandr)
		brightness->priv->use_hal = FALSE;

If you make gpm_brightness_xrandr_has_hw() always return FALSE, do
things work for you?

Richard.