On Nov 10, 2014 2:59 PM, "Antti Kantee" <[email protected]> wrote: > > On 10/11/14 13:44, Martin Lucina wrote: > > [email protected] said: > >> I can relate to that problem. Though, did you just try using it with > >> rumprun-xen? > > > > No, I assumed that if gcc -m32 can't create a host executable then it no > > work. I'll give it another go. > > I think buildrump.sh incorrectly tests for this (and that's possibly why > the baremetal build needs -H with travis). > > I think I'll just add KERNONLY around that probe. > > ... so many inconsistencies because buildrump.sh initially targeted > POSIX platforms ... > > > So, I'll make you a pull request for what we have now and continue the work > > I want to do while communicating with the various projects about > > upstreaming the minios changes. >
Actually I think buildrump with -k should not test whether the compiler can build executables that work at all. I added a fudge for OS "none" that should go away if that is tested. Justin
------------------------------------------------------------------------------
_______________________________________________ rumpkernel-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/rumpkernel-users
