Re: sysklogd -17.1 NMU build broken in mips/mipsel

[Chris Stromsoe <cbs@cts.ucla.edu>]

On Thu, 23 Feb 2006, Henrique de Moraes Holschuh wrote:
> On Thu, 23 Feb 2006, Bastian Blank wrote:
> > On Thu, Feb 23, 2006 at 03:00:19PM -0300, Henrique de Moraes Holschuh 
> > wrote:
> > > So far so good, but why is it allowed for a kernel header to include a 
> > > file that does not exist?
> >
> > mips is not managed in the tree of linus. So it is likely that it 
> > regulary breaks.
>
> Then it is not "only a sysklogd bug", but rather two bugs, one in 
> sysklogd, one in the kernel headers.


This is old, but was never really resolved.

In module.h I commented out

#include <asm/atomic.h>

and rebuilt sysklogd without any errors.  Is there any platform that needs 
asm/atomic.h for some reason?  If not, a bug to remove the #include should 
help.

Current sysklogd in stable (1.4.1-17) has an error in the rules file that 
caused it to be built without large file support, which means I can't keep 
weekly (or in some cases daily) log files on busy machines and have to do 
frequent rotations to avoid losing data at 2gb.

sysklogd 1.4.1-17.1 has the patched rules file and builds correctly with 
lfs.  What will it take to get a corrected -17.2 without the #include into 
stable?



-Chris