On Mon, 2009-08-17 at 09:40 +0100, Peter Robinson wrote:
>> I'm having a few issues compiling the syncevolution 0.9
release on
>> either F-11 or rawhide.
>>
>> I'm not sure if its a missing dep that isn't picked up by ./configure
>> or something else. The last beta compiled OK so I suspect its
>> something quite minor that I've missed.
>
> I'm not aware of build changes in the SyncEvolution and libsynthesis
> source since the last beta. Perhaps comparing against a build of that on
> your system will give a clue?
>
> You are compiling both in exactly the same way, I presume? The paths
> below (syncevolution-0.9/src/build-synthesis/src) indicate that you
> compile from source .tar.gz. Did you rebuild libtool and friends?
Nope. Just a plain ./configure and then a make :)
So did you patch libtool or not? The output below shows DIE_RPATH_DIE,
which indicates that it was patched.
>> I get the following undefined
>> reference errors below.
>
> The root problem of those is that libsynthesis.so.0 is not found:
>
>> /bin/sh ../libtool --tag=CXX --mode=link g++
>> -I/home/perobinson/rpmbuild/BUILD/syncevolution-0.9/src/build-synthesis/src
>> -O2 -g -pipe -Wall -Wp,-D_FORTIFY_SOURCE=2 -fexceptions
>> -fstack-protector --param=ssp-buffer-size=4 -m64 -mtune=generic
>> -uSyncEvolution_Module_Version -Wl,--export-dynamic -o syncevolution
>> syncevolution-syncevolution.o syncevolution-test.o
>> core/libsyncevolution.la -lglib-2.0
>> mkdir .libs
>> DIE_RPATH_DIE="/usr/lib64/syncevolution:$DIE_RPATH_DIE" g++
>> -I/home/perobinson/rpmbuild/BUILD/syncevolution-0.9/src/build-synthesis/src
>> -O2 -g -pipe -Wall -Wp,-D_FORTIFY_SOURCE=2 -fexceptions
>> -fstack-protector --param=ssp-buffer-size=4 -m64 -mtune=generic
>> -uSyncEvolution_Module_Version -Wl,--export-dynamic -o
>> .libs/syncevolution syncevolution-syncevolution.o syncevolution-test.o
>> core/.libs/libsyncevolution.so -lglib-2.0
>> /usr/bin/ld: warning: libsynthesis.so.0, needed by
>> core/.libs/libsyncevolution.so, not found (try using -rpath or
>> -rpath-link)
>
> I suspect that wiping out -rpath inside core/.libs/libsyncevolution.so
> too aggressively is causing this. But this is just a guess.
Yes, that looks like it. I get the following below when compiling it
as before (it worked fine in beta3).
Can you double-check that you are really compiling beta3 and 0.9 final
in exactly the same way?
I checked, ltmain.sh and aclocal.m4 (the main input for libtool) are the
same in both .tar.gz archives, so I really don't see why they should
behave differently.
The standard fix for this in
Fedora is to run the following 2 sed commands which generally fixes
the problem. Apparently in this case it doesn't :-)
sed -i 's|^hardcode_libdir_flag_spec=.*|hardcode_libdir_flag_spec=""|g'
libtool
sed -i 's|^runpath_var=LD_RUN_PATH|runpath_var=DIE_RPATH_DIE|g' libtool
Are you doing this to avoid the rpath warning? Wouldn't it be easier to
just ignore the warning? rpaths aren't nice for system packages, but not
getting the software to compile is the bigger problem ;-)
FWIW, there SyncEvolution is getting packaged for F-11:
http://bugzilla.moblin.org/show_bug.cgi?id=5061
--
Best Regards, Patrick Ohly
The content of this message is my personal opinion only and although
I am an employee of Intel, the statements I make here in no way
represent Intel's position on the issue, nor am I authorized to speak
on behalf of Intel on this matter.