[18:08] <nuclearbob> cyphermox or xnox: in the past few days ubiquity tests have been failing because the GTK_MODULES environment variable isn't populated. Is that something I should file a bug for, or do the tests need to be updated to work around that?
[18:18] <xnox> nuclearbob, i removed GTK_MODULES environemtn variable as overlays-scrollsbars has not been used that for many releases now
[18:18] <xnox> it's intentional change
[18:18] <nuclearbob> xnox: okay, thanks, I can update the tests to reflect that
[18:19] <xnox> nuclearbob, thank you!
[18:54] <nuclearbob> xnox: so it turns out the problem is in bin/ubiquity-wrapper, since it tried to append to GTK_MODULES if it's launched with --autopilot. Should I remove the two remaining GTK_MODULES references in the code, or have them create the variable if it doesn't exist?
[21:28] <nuclearbob> xnox: I've got a branch that removes the remaining references. If I make that in a running image, I can launch ubiquity via dbus again. I'm working on figuring out if there's a better way to test it, since dbus launches from /usr/bin/autopilot, so it's using what's rolled into the image
[21:28] <nuclearbob> xnox: it's here if you can take a look: https://code.launchpad.net/~canonical-platform-qa/ubiquity/no-gtk-modules/+merge/313473