Wrong colors displayed by eog on calibrated screens

[Frodon]

Binary package hint: eog

On a calibrated screen which use an ICC profile launched through dispwin, Eye-of-Gnome display wrong colors. Indeed it seems Eye-of-Gnome don't use well the ICC profile used by X.

On contrary all other apps like gthumb, GIMP, Fspot, firefox, ... don't have this issue.

According to the following dscussion EoG needs to have some metadata in the image telling that it is sRGB, however even with images containing metadata telling that it is sRGB the colors displayed are wrong.
http://mail.gnome.org/archives/eog-list/2009-March/msg00010.html

Thanks to anyone willing to help with this. In such state EoG is just unusable for photographers as it doesn't display the right colors.

[Launchpad Janitor (janitor)]

Status changed to 'Confirmed' because the bug affects multiple users.

[Julien D (julien-dubreuille-ubuntu)]

I am experimenting such a bug, and have been able to characterize it a little bit (with test profiles, to be sure)
For what I can see, it happens on a dual monitor setup.
Eye-of-Gnome (3.10.2, on trusty) is always using the ICC profile from the primary monitor (the one with the launcher), even when the window is on the secondary monitor.

This, with a version newer than the one where bug #272584 has been fixed, seem to prove that the previous duplicate status was wrong.

[Julien D (julien-dubreuille-ubuntu)]

Hum... After a series of blind investigations, I executed darktable-cmstest (coming with darktable):

$ darktable-cmstest
eDP1 the X atom and colord returned the same profile
 X atom: _ICC_PROFILE (23300 bytes)
  description: YYY
 colord: "YYY.icc"
  description: YYY
HDMI1 the X atom and colord returned different profiles
 X atom: _ICC_PROFILE_1 (0 bytes)
  description: (none)
 colord: "XXX.icc"
  description: XXX
Better check your system setup
 - some monitors reported different profiles
You may experience inconsistent color rendition between color managed applications

--> It looks like X atom _ICC_PROFILE_1 is not set! (and eog is trying to use it)
==> This may not be an eog Bug!