[Cluster-devel] Upcoming change to the build system.

[Cluster-devel] Upcoming change to the build system.

[Fabio M. Di Nitto]

Hi guys,

I just committed a "big" change to the build system in a private branch 
(fabbione_build).

http://git.fedorahosted.org/git/cluster.git?p=cluster.git;a=commitdiff;h=e112eca49345e13aa2344c50dacfcac142ee2d06

Please let me know if you have any objections to it or I'd like to commit 
it monday morning my time.

Thanks
Fabio

--
I'm going to make him an offer he can't refuse.

[Cluster-devel] Upcoming change to the build system.

[Christine Caulfield]

Fabio M. Di Nitto wrote:
> 
> Hi guys,
> 
> I just committed a "big" change to the build system in a private branch 
> (fabbione_build).
> 
> http://git.fedorahosted.org/git/cluster.git?p=cluster.git;a=commitdiff;h=e112eca49345e13aa2344c50dacfcac142ee2d06 
> 
> 
> Please let me know if you have any objections to it or I'd like to 
> commit it monday morning my time.


I'm all in favour of fixing a build system that doesn't remake targets 
on "make install" as I routinely build over NFS and it drives me insane 
when it fails.

Now all I need is for someone to fix the rest of the world ... ;-)
-- 

Chrissie

[Cluster-devel] Upcoming change to the build system.

[Joel Becker]

On Thu, Jul 31, 2008 at 09:49:16AM +0200, Fabio M. Di Nitto wrote:
> I just committed a "big" change to the build system in a private branch  
> (fabbione_build).
>
> http://git.fedorahosted.org/git/cluster.git?p=cluster.git;a=commitdiff;h=e112eca49345e13aa2344c50dacfcac142ee2d06
>
> Please let me know if you have any objections to it or I'd like to commit 
> it monday morning my time.

	Unfortunately, it doesn't work still.  install still depends on
all, so I get:

debian2:/nfs/cluster-working# make install
[ -n "" ] || make -C config all
make[1]: Entering directory `/build/jlbec/cman/cluster-working/config'
set -e && \
	for i in libs plugins tools; do \
		make -C $i all; \
	done
make[2]: Entering directory `/build/jlbec/cman/cluster-working/config/libs'
set -e && \
	for i in libccsconfdb; do \
		make -C $i all; \
	done
make[3]: Entering directory `/build/jlbec/cman/cluster-working/config/libs/libccsconfdb'
make[3]: *** No rule to make target `/usr/include/libxml2/libxml/parser.h', needed by `libccs.o'.  Stop.
make[3]: Leaving directory `/build/jlbec/cman/cluster-working/config/libs/libccsconfdb'
make[2]: *** [all] Error 2
make[2]: Leaving directory `/build/jlbec/cman/cluster-working/config/libs'
make[1]: *** [all] Error 2
make[1]: Leaving directory `/build/jlbec/cman/cluster-working/config'
make: *** [config] Error 2

Joel

-- 

"And yet I find,
 And yet I find repeating in my head.
 If I can't be my own, 
 I'd feel better dead."

Joel Becker
Principal Software Developer
Oracle
E-mail: joel.becker at oracle.com
Phone: (650) 506-8127

[Cluster-devel] Upcoming change to the build system.

[Fabio M. Di Nitto]

On Thu, 2008-07-31 at 15:03 -0700, Joel Becker wrote:
> On Thu, Jul 31, 2008 at 09:49:16AM +0200, Fabio M. Di Nitto wrote:
> > I just committed a "big" change to the build system in a private branch  
> > (fabbione_build).
> >
> > http://git.fedorahosted.org/git/cluster.git?p=cluster.git;a=commitdiff;h=e112eca49345e13aa2344c50dacfcac142ee2d06
> >
> > Please let me know if you have any objections to it or I'd like to commit 
> > it monday morning my time.
> 
> 	Unfortunately, it doesn't work still.  install still depends on
> all, so I get:

Ok. I think it's possible to remote the install: all in favour of
install: but the bug you see here is not in our build system (at least I
believe so).

> 
> debian2:/nfs/cluster-working# make install
> [ -n "" ] || make -C config all
> make[1]: Entering directory `/build/jlbec/cman/cluster-working/config'
> set -e && \
> 	for i in libs plugins tools; do \
> 		make -C $i all; \
> 	done
> make[2]: Entering directory `/build/jlbec/cman/cluster-working/config/libs'
> set -e && \
> 	for i in libccsconfdb; do \
> 		make -C $i all; \
> 	done
> make[3]: Entering directory `/build/jlbec/cman/cluster-working/config/libs/libccsconfdb'
> make[3]: *** No rule to make target `/usr/include/libxml2/libxml/parser.h', needed by `libccs.o'.  Stop.
> make[3]: Leaving directory `/build/jlbec/cman/cluster-working/config/libs/libccsconfdb'
> make[2]: *** [all] Error 2
> make[2]: Leaving directory `/build/jlbec/cman/cluster-working/config/libs'
> make[1]: *** [all] Error 2
> make[1]: Leaving directory `/build/jlbec/cman/cluster-working/config'
> make: *** [config] Error 2

The issue seems to be that .d files (generated with gcc -MMD option)
still include libxml/ headers in the dependency chain. This looks wrong
to me.

If I generate the .d file using -MD I can see all system includes
(including openais headers from /usr/include/openais), but with -MMD
openais goes away together with all the other system includes. xml2
stuff is still there.

I wonder if gcc is unhappy because xml2-config pulls
in /usr/include/libxml2 and marks that as non system standard path.

Fabio

[Cluster-devel] Upcoming change to the build system.

[Fabio M. Di Nitto]

On Thu, 2008-07-31 at 15:03 -0700, Joel Becker wrote:
> On Thu, Jul 31, 2008 at 09:49:16AM +0200, Fabio M. Di Nitto wrote:
> > I just committed a "big" change to the build system in a private branch  
> > (fabbione_build).
> >
> > http://git.fedorahosted.org/git/cluster.git?p=cluster.git;a=commitdiff;h=e112eca49345e13aa2344c50dacfcac142ee2d06
> >
> > Please let me know if you have any objections to it or I'd like to commit 
> > it monday morning my time.
> 
> 	Unfortunately, it doesn't work still.  install still depends on
> all, so I get:

I remove the install: all depends from my private branch and it works
for me. It also allow me to remove libxml2-dev from the target machine
and i don't hit your problem anylonger.

Fabio