On Fri, 19 Mar 2010, Carlos R. Mafra wrote: > Right now linux.c will be compiled in other OSes too, right?
yes. this is exactly how it has always been, it just hasn't been called linux.c. the code fails gracefully when it cant do its thing. > Do you have a bigger plot in mind, where this will be used strictly > on linux systems? Well, the cleanup is worth doing regardless, so yes. the same functionality is achieved via sysctl(3) or kvm*(3) (need to figure out which one do i need; some of the kvm calls are privileged which are out of the question) on the bsds and solaris. other platforms people will either step up or continue working as they did before (ie being a stub, basically). > I am just curious whether you have more coming on in this direction :-) yes. there's lots to be done, most of which i have no idea about (platform detection in autoconf, for example), but it's a start. > I mean, have you seen more OS-centric stuff out there which will be > worth moving to osdep/ too? i don't know (afk now), but even if this is the only one, it at least makes it apparent for future adapters that there are os-dependent parts. -- [-] mkdir /nonexistent -- To unsubscribe, send mail to [email protected].
