On 25/05/14 16:49, Justin Cormack wrote: >> Dunno, it sounds like a lot more work for us to provide some -m/-n >> grammar instead of just allowing to specify compiler flags. It's also >> more work for the users, since they can't quite be sure if -m/-n params >> map to the compiler flags they know (and I do assume that anyone who >> wants to compile for earmv7hf knows what they're doing ;) > > True. How about rather than -F CFLAGS=... accepting all the -m options > like -mpcu=... -mabi= -marm -mtune= and passing them to the > appropriate CFLAGS/AFLAGS/LDFLAGS (most can just go to all three).
If there's something that should go to all of C/A/CPP/LDFLAGS, maybe we can just add a shortcut -F ALLFLAGS which does that. Once again, I'm hesitant to add too much "smartness" into the script, because "smartness" seems to invariably do the wrong thing. >> I'm not sure if there's a good answer for what an "ambiguous" compiler >> should produce with buildrump.sh. Maybe the best thing to do is to >> remove magic like EXTRA_CFLAGS=-march=armv6k, throw an error, and ask >> that the user is more specific. > > Thats an odd situation, is that the NetBSD compiler on pi or the Linux > one? It is not clear that buildrump can work out all the unsupported > situations. NetBSD uses wrong assembly instructions: http://gnats.netbsd.org/47401 ------------------------------------------------------------------------------ The best possible search technologies are now affordable for all companies. Download your FREE open source Enterprise Search Engine today! Our experts will assist you in its installation for $59/mo, no commitment. Test it for FREE on our Cloud platform anytime! http://pubads.g.doubleclick.net/gampad/clk?id=145328191&iu=/4140/ostg.clktrk _______________________________________________ rumpkernel-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/rumpkernel-users
