svn commit: trunk/uClibc: extra/scripts

Mike Frysinger vapier at gentoo.org
Tue Dec 12 03:59:32 UTC 2006


On Monday 11 December 2006 13:43, Erik Andersen wrote:
> On Mon Dec 11, 2006 at 11:16:28AM -0500, Mike Frysinger wrote:
> > On Monday 11 December 2006 01:41, Erik Andersen wrote:
> > > Ok, so now where does include/asm and include/linux/ come from?
> >
> > the top level makefile adds a -I for it ... was working for bfin before i
> > committed, but guess no longer eh
>
> I do not think adding a -I is going to be sufficient, as
> evidenced by the fact that uClibc does not compile.

it compiles if you use the right path ... that probably involves 
appending '/include' to the current value

ive update my build script to do just that

> Even if you 
> get uClibc to compile, how would a client application magically
> also get the correct -I paths added?

i dont think uClibc shouldnt really be concerned with this ... imo, the kernel 
headers should be handled by the user creating the toolchain, not uClibc 
trying to make a guess at what's correct (and what probably doesnt work 
anyways) ...

this is a logical decouple-ation (is that a word ?  well it is now ...) of the 
toolchain process ... especially considering the fact that we're moving in 
the direction of sanitized headers, the notion of pointing uClibc at the 
actual kernel source is an antiquated notion

if we want to kill off the "make headers" step, then this change is also a 
required growing pain
-mike
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 827 bytes
Desc: not available
Url : http://lists.busybox.net/pipermail/uclibc/attachments/20061211/5572434f/attachment-0002.pgp 


More information about the uClibc mailing list