Apparently, though unproven, at 02:43 on Thursday 18 November 2010, Walter Dnes did opine thusly:
> On Wed, Nov 17, 2010 at 02:00:48AM +0200, Alan McKinnon wrote > > > If you then mentioned that their defaults broke Dale's setup, they'd > > likely answer "Who's Dale?" followed shortly by "None of us have > > hardware like Dale to test. Sorry 'bout that. Set USE=-hal" > > Of course the USE flag advice is given *AFTER* the new flag breaks > your system. That's why I use "-*" at the beginning of my USE in > /etc/make.conf. I never found out whether hal would break my system<G>. > If Dale had used "-*" his X would not have broken, even if some other > ebuild pulled it onto the machine as a hard-coded dependancy. Looks like the *actual* problem is non-application of OYFEAL[1],not what the devs do. Dale should have seen a new package being installed - an "N" inside "[ ]", should have seen new flags highlighted in colour, and should have decided. If he decided to go without hal, nothing would have changed for him. He decided to go with hal, and he got the breakage he did. Either way, seeing the USE flag changes tells him nothing about the impending breakage. He can only know that by *doing it*, or reading about others that did it. Let's look at this sanely and realise that there's nothing magic about hal and what it did. It has bugs. Big deal. So did jpeg and look at the carnage that one caused. How would your method of handling USE have assisted in preventing that breakage? Please note that the breakage in jpeg is much *much* more common than changes to default USE. -- alan dot mckinnon at gmail dot com
