well, I got around most of the compilation issues by going with an older ohphone snapshot. Now I have another infuriating problem, though. On x86-64 (and ONLY x86-64), the build of ohphone fails: make P_SHAREDLIB=1 opt make[1]: Entering directory `/home/mandrake/rpm/BUILD/ohphone' g++ -D_REENTRANT -Wall -O2 -g -pipe -Wp,-D_FORTIFY_SOURCE=2 -fexceptions -DP_64BIT -I/usr/share/pwlib/include -DPTRACING -I/usr/include/openh323 -O2 -g -pipe -Wp,-D_FORTIFY_SOURCE=2 -fexceptions -O2 -g -pipe -Wp,-D_FORTIFY_SOURCE=2 -fexceptions -DLDAP_DEPRECATED -c main.cxx -o obj_linux_x86_64_r/main.o g++ -D_REENTRANT -Wall -O2 -g -pipe -Wp,-D_FORTIFY_SOURCE=2 -fexceptions -DP_64BIT -I/usr/share/pwlib/include -DPTRACING -I/usr/include/openh323 -O2 -g -pipe -Wp,-D_FORTIFY_SOURCE=2 -fexceptions -O2 -g -pipe -Wp,-D_FORTIFY_SOURCE=2 -fexceptions -DLDAP_DEPRECATED -c vidlinux.cxx -o obj_linux_x86_64_r/vidlinux.o g++ -D_REENTRANT -Wall -O2 -g -pipe -Wp,-D_FORTIFY_SOURCE=2 -fexceptions -DP_64BIT -I/usr/share/pwlib/include -DPTRACING -I/usr/include/openh323 -O2 -g -pipe -Wp,-D_FORTIFY_SOURCE=2 -fexceptions -O2 -g -pipe -Wp,-D_FORTIFY_SOURCE=2 -fexceptions -DLDAP_DEPRECATED -c xlibvid.cxx -o obj_linux_x86_64_r/xlibvid.o make -C /usr/src opt make[2]: Entering directory `/usr/src' make[2]: *** No rule to make target `opt'. Stop. make[2]: Leaving directory `/usr/src' Full log at http://kenobi.mandriva.com/queue/failure/cooker/main/release/20070601071609.adamw.kenobi.26266/log/ohphone-1.4.5-0.20050322.6mdv2008.0/build.0.20070601071836.log I believe this is caused by this section in openh323u.mak (which is included by the ohphone Makefile): ifndef LIBRARY_MAKEFILE ifdef DEBUG $(OH323_LIBDIR)/$(OH323_FILE): $(MAKE) -C $(OH323_SRCDIR) debug else $(OH323_LIBDIR)/$(OH323_FILE): $(MAKE) -C $(OH323_SRCDIR) opt endif # DEBUG endif # LIBRARY_MAKEFILE Since we've defined openh323dir (and hence oh323_srcdir) to be /usr (so that it correctly finds /usr/share/openh323 , /usr/include/openh323 ...), it's doing a 'make -C opt' in /usr/src , which is obviously silly. As far as I can work out that code's expecting OH323_SRCDIR to be a directory of clean openh323 source code. I'm sure that's wrong for ohphone compilation and it should never be hitting this bit of code, but it obviously is. So what's wrong here? Why does it happen, and why does it only happen on x86-64? My i586 build goes through just fine. Is LIBRARY_MAKEFILE set for an i586 build and not for an x86-64 build for some reason, or what? (to recap - building an ohphone snapshot dated 20050322 against openh323 1.18.0 and pwlib 1.10.7. But AFAICT nothing related to this has actually changed in current CVS). -- adamw ------------------------------------------------------------------------- This SF.net email is sponsored by DB2 Express Download DB2 Express C - the FREE version of DB2 express and take control of your XML. No limits. Just data. Click to get it now. http://sourceforge.net/powerbar/db2/ _______________________________________________ Openh323-devel mailing list Openh323-devel@xxxxxxxxxxxxxxxxxxxxx https://lists.sourceforge.net/lists/listinfo/openh323-devel