GIcon regression?
[ RUN ] TestVolumeImp.GetIconName
/<<BUILDDIR>>/unity-7.5.0+18.10.20180905/tests/test_volume_imp.cpp:71: Failure
Expected: volume_->GetIconName()
Which is: ". GThemedIcon gnome-dev-cdrom gnome-dev-cdrom-symbolic"
To be equal to: icon_name
Which is: "gnome-dev-cdrom"
[ FAILED ] TestVolumeImp.GetIconName (0 ms)
All that test does is create GIcon
with g_icon_new_for_string ("gnome-dev-cdrom", NULL)
and then compares output from g_icon_to_string (icon)
...
From g_icon_to_string
focumentation:
If icon is a GThemedIcon with exactly one name, the encoding is simply the name (such as network-server).
To me this looks like !72 (merged) has introduced regression. Should I update unity tests? Or this must be fixed in glib to restore documented behavior?