Open podiki opened 6 years ago
Additional information: the profile applied with xiccd
and with e.g. dispwin -L
(from displaycal
) have the same number entries, as extracted via xprop -root | grep ICC
. However, they are not identical, and differ in many places. Due to their length I am not including the ICC profiles here, but could make them available.
@agalakhov have you had the chance to investigate this at all?
Sorry, I totally forgot about that. Could it be that there is something that overwrites the X atom? For example, the most recent version of Xfce handles ICC profiles on its own.
No, I don't think so, as I'm in a bare WM without anything else that should be touching colord/ICC stuff. Just now I tried where the X atom had a profile, but nothing with colord (xiccd wasn't running). Running xiccd loaded a profile for colord, but again differs in size and contents from doing dispwin -L
. Is there any debug output I could get from xiccd? Maybe it doesn't like something about the profile.
There seems to be an issue with the ICC profile in the X atom, where it is a different size than the actual file (e.g. compared to same profile in colord). I noticed this by running
darktable-cmstest
(from the Darktable photo editor), where even though the output will show the same description of the profile in the X atom and in colord, the size does not match the actual file, and thus this program reports a possible issue with the profile. One can look at the ICC size as reported here compared to the actual file and see that the X atom is smaller. This does not happen when using something like Gnome color manager.(Note: I'm not sure if this has a practical effect in actual color management, without knowing what has changed from the actual ICC profile file.)