Tracker item #2046262, was opened at 2008-08-11 12:45 Message generated for change (Comment added) made by tgurr You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=989708&aid=2046262&group_id=204462
Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: None Group: None Status: Open Resolution: None Priority: 5 Private: No Submitted By: Timo Gurr (tgurr) Assigned to: Nobody/Anonymous (nobody) Summary: open-vm-tools-2008.08.08-109361 fails to configure Initial Comment: When performing ./configure it bails out with following error message: checking for GCC flag -Wno-pointer-sign... yes checking for GCC flag -Wno-unused-value... yes checking for GCC flag -fno-strict-aliasing... yes checking for GCC flag -Wno-unknown-pragmas... yes checking for GCC flag -Wno-uninitialized... yes configure: error: conditional "am__fastdepCXX" was never defined. Usually this means the macro was only invoked conditionally. Using: gcc-4.3.1 automake-1.10.1 Attached patch fixes the issue for me. ---------------------------------------------------------------------- >Comment By: Timo Gurr (tgurr) Date: 2008-08-13 12:01 Message: Logged In: YES user_id=1687320 Originator: YES Worked perfectly, just had to create the missing m4 directory. Thanks for looking into this, greatly appreciated! ---------------------------------------------------------------------- Comment By: Adar Dembo (adembo) Date: 2008-08-13 09:03 Message: Logged In: YES user_id=1867590 Originator: NO File Added: configure.ac.patch ---------------------------------------------------------------------- Comment By: Adar Dembo (adembo) Date: 2008-08-13 09:02 Message: Logged In: YES user_id=1867590 Originator: NO Alright, I've spent a while looking at this and I think I have it figured out. The reason this problem manifests itself only when you run autoreconf is because you probably have libtool-2.2 installed in your system. Internally we build the configure script using libtool-1.5.22, and when I changed our build to use libtool-2.2, I also saw the same errors. So I've gone ahead and done three things: - I've upgraded our internal compile test infrastructure to use libtool-2.2 instead of libtool-1.5.22, because it fixes a few other issues and is more cutting edge. This necessitated a few other changes to configure.ac (which you'll see in the patch I'm uploading). - I've also made the AC_PROG_CXX check unconditional as you did in your patch. I looked at how autoconf implements the AC_PROG_CXX macro and saw that unlike AC_PROG_CC, it does _not_ throw an error if the compiler isn't found. So it should be safe to have everyone use AC_PROG_CXX, even if they aren't using libicu. - To make sure that we actually do have a C++ compiler, I added a call to AC_CHECK_TOOL on CXX in the case where we're using libicu. Here's the patch; you'll need to use autoreconf because I'm too lazy to generate diffs on configure and all the Makefile.in files. Let me know if it solves the problem. File Added: Makefile.am.patch ---------------------------------------------------------------------- Comment By: Timo Gurr (tgurr) Date: 2008-08-12 11:34 Message: Logged In: YES user_id=1687320 Originator: YES Tarball, because the file was too big to attach in plain/text. File Added: configure.fail.tar.bz2 ---------------------------------------------------------------------- Comment By: Timo Gurr (tgurr) Date: 2008-08-12 11:29 Message: Logged In: YES user_id=1687320 Originator: YES File Added: build.log ---------------------------------------------------------------------- Comment By: Timo Gurr (tgurr) Date: 2008-08-12 11:28 Message: Logged In: YES user_id=1687320 Originator: YES Attached is my config.log. After reading your comment I also found out that if I don't force the rebuild of the of the autotools scripts it works like in your case, I'm also attaching my full build.log and configure.fail (the autoconf 2.62 regenerated configure script). File Added: config.log ---------------------------------------------------------------------- Comment By: Adar Dembo (adembo) Date: 2008-08-11 21:38 Message: Logged In: YES user_id=1867590 Originator: NO Hmm, I don't think that patch is correct. If you look further into configure.ac you'll see that AC_PROG_CXX is already invoked in the case where icu is enabled. If icu is disabled (which it is in your case) we shouldn't need the C++ compiler/linker, and so we shouldn't require it for all open-vm-tools builds. That's why it's surprising to see this failure. When I run your ./configure line locally (after removing my /usr/bin/g++ symlink) I don't see this problem. Could you share your config.log file? Is anyone else experiencing this bug? ---------------------------------------------------------------------- Comment By: Timo Gurr (tgurr) Date: 2008-08-11 13:01 Message: Logged In: YES user_id=1687320 Originator: YES Forgot the most important, following ./configure parameters were given: ./configure --prefix=/usr --host=x86_64-pc-linux-gnu --mandir=/usr/share/man --infodir=/usr/share/info --datadir=/usr/share --sysconfdir=/etc --localstatedir=/var/lib --without-kernel-modules --without-icu --without-x --disable-multimon --libdir=/usr/lib64 --build=x86_64-pc-linux-gnu ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=989708&aid=2046262&group_id=204462 ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/ _______________________________________________ open-vm-tools-devel mailing list open-vm-tools-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/open-vm-tools-devel
