I recall waay back on Feb 01 when Nathan Hawkins wrote: > Personally, I'm not interested in creating another BSD. (Not in the sense > of Free/Open/Net BSD.) I'd just like to integrate parts of FreeBSD (kernel > and closely related programs) into Debian. So it could simply be the > FreeBSD source code patched to fit into the Debian filesystem and package > structure. Rather more likely, there'll be numerous patches... :)
That's pretty much what it looks like it would be, from what I've seen anyway.. > I think enthusiasm would be premature. I think enthusiasm is what it might end up taking to get people moving on the project =D > There's nothing getting done, so far > as I know. (I'm waiting on some hardware that I should have had back in > early December. I don't know much about what anyone else is doing, because > this list has mostly been discussing politics. :-( ) See earlier message, I did some playing around with dpkg and apt on FreeBSD 4.0, and they seem portable enough. It takes a few important changes (and the usage of gmake) but it works out ok. The only thing I haven't figured out yet is how to make mmap() work in apt. It appears to be giving problems. Once those things are worked out, some more tools can be ported using the same method and people ought to be able to start playing with compiling Debian packages, and creating Debian packages from the FreeBSD bin distribution. I know there are many more details, but I'm taking the gung-ho see how far we can get strategy at the moment. > It's obvious that only > a minority are interested in this from either side, but I believe there's > great potential to make both systems better. It seems to me that one prudent way of advancing is to try to make the minimum number of changes to either system independent of the system itself (few Debian patches as possible, few BSD patches as possible) and then let people roll their own as they like. The base distro might include the basic install (like the Debian pre-dselect install), plus a few packages for things like gcc and dpkg devel tools, and then you'd go from there. This way, you don't waste disk space on the Debian server with Yet Another Recompile (I've actually seen this argument on devel) but it is still possible to setup a system. One cool thing to do would be to incorporate patches from Ports where available. I don't know the feasability of this because of licensing, but I'm assuming it's ok since FreeBSD does much the same thing with the pre-compiled packages for download. Anyway, there are patches in ports to make a given program compile on FreeBSD. There are also patches with a Debian source package. Might be cool to have a util that looks for both and tries to make a FreeBSD-compatable Debian package. Ok, that's pretty far fetched but I'm brainstorming again (out of politics for me!) -- And on the eighth day, we bulldozed it.
