On Mon, Jul 9, 2012 at 10:22 PM, Brian Dolbec <dol...@gentoo.org> wrote: > There have already been users on the forums with that very confusion of > what to do with the cryptic "[!icu?]". And there are currently many > forum threads involving the icu use flag, qt-webkit,...
Yup, this issue hit anybody who has qt-webkit and chromium installed. I wouldn't be surprised if that is half of the entire userbase. We ran into another confusing icu-related issue with qt-core a few weeks ago (bug 413541). I can understand that the qt maintainers want to get away from enabling icu for this reason, but chromium is a VERY popular package so it is really only disabled in the sense that it annoys a bazillion people who have to un-disable it and then still run into the problems. Better portage logic might help here, but I think we need to consider whether a non-optimal decision from a single package perspective is going to lead to a better overall experience for our userbase. Zac suggested adding icu to the profile, which would work, though really just adding it as the default for these two packages would really address the issue until portage can catch up. Those who REALLY don't want icu support in qt-webkit can always disable it manually now that the flag is there. If there is a fear that this default will lead to more bugs, those bugs will happen anyway, since anybody running chromium has to enable that flag. Rich
