cvs commit: src/contrib/openbsm - Imported sources
Robert Watson
rwatson at FreeBSD.org
Thu Aug 31 06:14:32 UTC 2006
On Thu, 31 Aug 2006, Robert Watson wrote:
> On Mon, 28 Aug 2006, Warner Losh wrote:
>
>> Import on vendor branch two files that have been tweaked to unbreak
>> the build. The openbsm folks are free to fix it in any other way they
>> see fit once they resurface.
>>
>> Basically, make everything always be const char **, even though const
>> char ** usually should be 'const char * const *' in most cases. This
>> makes the three different definitions consistant and allows world to
>> build.
>
> The audit_record.h file in src/contrib/openbsm/bsm is not used in building
> FreeBSD; instead, src/sys/bsm is used. If audit_record.h is being used from
> the contrib tree, there is a problem with the build parts. Looking at the
> BSD makefiles associated with libbsm, I'm not sure I see how that can
> happen.
>
> Notice that world builds on all platforms but amd64, and that the build
> failure is not an amd64-specific warning/error in the compile: it has to do
> with matching prototypes, not with pointer/int size, format strings, etc.
> I'm not convinced OpenBSM is at fault here, although clearly it's triggered
> by the OpenBSM import.
Just as a follow-up to this thread for the more general benefit, as portions
went offline: it looks like the problem is that the compat32 support on the
amd64 platform is using different include paths or include installs than is
used for non-compat32 builds, resulting in libbsm under compat32 using the
wrong include file (in contrib). Since the contrib one differs from the base
system one, due to some changes not yet merged to the FreeBSD kernel in the
base system, it conflicts with src/sys/bsm, which contains the include files
intended to be built as part of the base system. I'm a little surprised that
your change allows the base to continue to build, given that your "fix" didn't
change src/sys/bsm, only the contrib tree.
So it sounds like we need to fix the compat32 build environment before your
build fixes can be reverted.
Robert N M Watson
Computer Laboratory
University of Cambridge
More information about the cvs-src
mailing list