DragonFly kernel List (threaded) for 2005-04
On Fri, Apr 22, 2005 at 10:39:05AM +0200, Joerg Sonnenberger wrote:
> (1) We have to support it first. I have more important work at hand,
> the tool chain update is deferred for that reason.
So if anyone else wants to take a stab at it, why not let them give it a try. I
am really unsure if I will have time for any computer things the coming weeks,
but if I do I could probably give a helping hand on this part.
> (2) It needs fixes for certain parts of the tree, not everything is going
> to compile with it.
The way switching of gcc work means that we don't need to fix everything before
gcc 4 is imported. When gcc 4 then is in the tree everyone can do a test
buildworld and see where it breaks and fix it.
> (3) It needs TESTING. Things might break, things might not break. You
> don't know.
The best way to test it is to have it in the tree. There is not testing like
having different kind of people having different needs beating up it.
We can have gcc 4 in the tree without supporting it as a stable piece of code
yet. As I said above, the only way to get it to a point where it is stable is to
test it. And our way of having several versions of gcc is great for this kind of
testing. You can even just do a buildworld with gcc 4 and just put it into a