When a file under CONFIG_PROTECT doesn't exist yet, Portage simply installs it. It would be nice to have an option (say --protectforreal) that makes emerge create a .cfg_0000_* file also when the destination file does not exist yet in the CONFIG_PROTECT area. When first installing a system, one wouldn't want to use this option, but once a system is fully installed and stable, I wish to know about every change and addition in /etc. In a slightly different form this was discussed earlier: http://article.gmane.org/gmane.linux.gentoo.devel/42320 http://article.gmane.org/gmane.linux.gentoo.devel/42343 http://article.gmane.org/gmane.linux.gentoo.devel/42359
I don't think such a feature is useful or desired by the majority of users.
FYI, here's a reference to the discussion on gentoo-user that was the trigger for this request: http://thread.gmane.org/gmane.linux.gentoo.user/182695/focus=182701
(In reply to comment #1) > I don't think such a feature is useful or desired by the majority of users. If majority is a criterion, then it's time to get rid of options like -B, -d, and -F. Surely most users have no need or desire for those.
(In reply to comment #3) > If majority is a criterion, then it's time to get rid of options like -B, -d, > and -F. Surely most users have no need or desire for those. Yes, but those features are already implemented and have been for a very long time. Why would the portage devs spend the time to implement what you're asking when it's not useful to them and very few users would ever use it? You're always welcome to whip up a patch(es) for this and attach it to this bug.
(In reply to comment #4) > You're always welcome to whip up a patch(es) for this and attach it to this > bug. I've tried. But it is hell in there. (In portage.py around line 7708.) But even if I'd succeed, what are the chances of getting the patch accepted if the feature is not considered useful?
(In reply to comment #5) > (In reply to comment #4) > > You're always welcome to whip up a patch(es) for this and attach it to this > > bug. > > I've tried. But it is hell in there. (In portage.py around line 7708.) Yeah, the vdb merge code isn't as nice as it could be. > But even if I'd succeed, what are the chances of getting the patch accepted if > the feature is not considered useful? Don't let yourself get discouraged by individual opinions. I see how such a feature could be considered useful (though I don't think I'd use it myself). One suggestion for this would be to make this configurable on a per directory base (e.g. by extending the CONFIG_PROTECT syntax somehow or using a new variable eventually) so it could be enabled for things like /etc/foo.d/ only instead of all of /etc.
I don't think anyone in the portage team wants to work on this, but we'll look at and consider provided patches.
