>PS, do we really need multiboot compliance ? >Eventually, you'll be able to use GRUB from within FreeMWare, >and until then we only want to run our own testkernels anyway. >It seems to me like a rather useless feature... unless you've >been trying to boot FIASCO already ;) Indeed, thats the next greater goal... ;) FIASCO ist _not_ a test kernel, but very small and sources are available. (and the author sits a few doors away from mine ;) >> ? patch8.cvsdiff >> ? guest/virtcode/keyboard.c >> ? guest/virtcode/minishell.c >> ? guest/virtcode/helper.c >> ? user/multiboot.h >> ? user/oskit_types.h > >These files were not included in the patch... I think cvs diff >isn't such a good idea. But should came around with the patch before. >Perhaps it would be useful to have *some* global variables... Yes, I noticed it when I changes to do came not to end. But then the most work was done, I didn't countermand it. >This is not the way you want to use it... you want to emulate >the whole REP INS in one go, something like: Yes, as I sayed I remembered that someone had already done such or a similar thing, hoped he would shuffle it in ;) >I'll clean up this weekend ... I'll fix this this weekend [and so on] Fine! >Okay, I forgot to ask about this one last time --- >I added VERSION when I cleaned up user.c (once again ;)), >but we never actually did discuss our versioning conventions. >We should think about this one... Perhaps we could take a proven example like the linux kernel. The patches were collected (in CVS repository) and from time to time takes someone the sources, rolls a tarball, makes a patch against the last version and puts it with some comments on a web page. As a first-try-demo, I made this on http://os.inf.tu-dresden.de/~jn4/freemware/. Comments? jens
