DisplaySize and X and DPI

Keith Packard keithp at keithp.com
Tue Sep 25 16:37:41 PDT 2007

On Tue, 2007-09-25 at 17:47 -0400, mcr at xdsinc.net wrote:

>   I think that the situation is twofold, 
>     a) X/video "hardware"/etc. doesn't actually know what the
>        DisplaySize *is*, and should say "0" or unknown.
>        I don't think this happens very often, unfortunately.

In the core protocol, it cannot report zero size as applications often
divide by this size to compute DPI themselves.

However, in RandR 1.2, it *does* report 0,0 if it doesn't know the size
of the output. The size reported in the core protocol can now be changed
on-the-fly by clients, while the size reported in the RandR output is as
detected by the hardware (or 0,0 if that is unknown).

>     b) X thinks it knows, and reports a value, but has no way to know
>        that it is wrong.

Either X gets data from EDID and sets an appropriate value, or X picks
'96dpi' for the core protocol.

>   I haven't looked deeply into randr 1.2, but I see that the palette of
> resolutions is gone, which is great.  I didn't see if the display X/Ymm
> was settable seperately from the X/Y resolution. Is it?

Yes. You can set the reported screen physical geometry to whatever value
you'd like. The xrandr utility lets you specify either dpi or mm.

keith.packard at intel.com
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part
URL: <http://lists.x.org/archives/xorg/attachments/20070925/3b50cf93/attachment.pgp>

More information about the xorg mailing list