[colord] dispwin sets the wrong X.org _ICC_PROFILE atom

[Tobias Ellinghaus]

Am Freitag, 25. November 2016, 23:51:52 CET schrieb Niccolò Belli:
> On venerdì 25 novembre 2016 16:43:04 CET, Tobias Ellinghaus wrote:
> > Am Freitag, 25. November 2016, 15:55:39 CET schrieb Niccolò Belli:
> >> As you can see, unless darktbale-cmstest is bugged, dispwin installed the
> >> profile to eDP1 (which is not even listed among the displays, because
> >> it's turned off in xrandr) instead of DP1-8 (which is display 1).
> > 
> > I would say it's a safe bet to assume darktable-cmstest to be buggy.
> 
> Is there any other way I can check the X atom to be sure?
> Do someone know how darktable-cmstest checks for the _ICC_PROFILE X atom?

Looking at this part as an example:

eDP1    the X atom and colord returned different profiles
        X atom: _ICC_PROFILE (2815280 bytes)
                description: U2515H #1 2016-09-25 18-25 D6500 2.2 M-S 
XYZLUT+MTX

The X atom mentioned should match the content shown, so _ICC_PROFILE has the 
profile U25...
The part that might be buggy is the mapping between _ICC_PROFILE and eDP1. All 
of these things lack either documentation or sample code how to use it in the 
real world without learning about several decades of X11 naming conventions 
and nomenclature. It also seems to be undefined in the case that xrandr is 
used.

The code of darktable-cmstest is here:

https://github.com/darktable-org/darktable/blob/master/src/cmstest/main.c

I fixed a potential bug in there recently, maybe try compiling it standalone 
(see comment in the top of the file for how) and check if it shows a different 
result.

> Niccolò Belli

Tobias
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 833 bytes
Desc: This is a digitally signed message part.
URL: <https://lists.freedesktop.org/archives/colord/attachments/20161126/b5949f3b/attachment.sig>