As an advanced, long-time user of Sabayon with additional Gentoo experience, I've always mixed the use of Portage and Entropy on my systems— when I started using Sabayon, there was no Entropy. Despite all of the exclamatory warnings in the wiki about the dangers of this practice, I've actually found it to be surprisingly beneficial in helping me maintain a more stable, complete system. Here's the arrangement I've come to
- Entropy serves the majority of my package management needs, handling my system updates, and it's my go-to system when looking for new software
- I manage conflicts between packages which exist both in my Portage tree and the Entropy packageset using complementary entries in /etc/portage/package.* and /etc/entropy/packages/package.*
- I set up my Entropy system not to ignore whatever packages I've customized with Portage, but usually simply to not to downgrade newer versions of packages I've deliberately upgraded
Thinking about it now, it may well be possible to set something similar up through the careful use of keywords, masking, and perhaps even the employment of two Portage trees, clean and unclean (basically turning on and off all layman overlays manually instead of adding them in make.conf). Even so, Sabayon's dual-packager system has made this very natural to me, and fast as well. I'm really enjoying it, and I hope Sabayon keeps support for Portage indefinitely. I hope as well that as I continue to use Sabayon, improvements in Entropy, Portage, and my understanding of both will help me further refine this package management workflow I've stumbled into.
So there's my 2¢; a slightly different take on the usual advice regarding Sabayon and Portage. Has anyone else got similar experiences, disaster stories, or tips? I'd like to hear the devs way in on this, if possible.