cylent wrote:please explain this statement a bit more. i keep hearing about how this will happen but i have never ran into problems.
Think about it this way: Application A can be compiled in 3 different ways. It depends on Dependency A which can also be compiled in 3 different ways. When entropy compiles Application A in a certain manner, it also compiles Dependency A in a certain manner.
If you change how Application A is compiled, it might not like how Dependency A is compiled (although that's fairly unlikely). If you change how Dependency A is compiled, Application A has a good chance of crashing.
One option is just to compile both of them again, right? Well, I didn't mention that Application B,C,I, and Z also use Dependency A. As you can see, it's already getting kind of complicated pretty quickly.
Portage is great for figuring how to get applications and their dependencies to compile and work with each other. But by compiling things the way you want to, you break expected compatibility with binary packages that are compiled with USE flags. If you are only compiling applications at the top most level (they are not dependencies for other applications), then mixing portage and entropy is safe. Worst case scenario, the application you compile won't work, but your system should be fine. But if you compile some low level libraries in a way that doesn't work for the entropy provided binary applications that use them...your system will be quite unhappy.