Red Hat Bugzilla – Bug 187687
GCC fails to rebuild itself for FC5 update kernel
Last modified: 2007-11-30 17:11:29 EST
Description of problem: The command "rpmbuild --rebuild gcc-4.1.0-4.src.rpm" fails during the install stage. Version-Release number of selected component (if applicable): gcc-4.1.0-4 How reproducible: Always Steps to Reproduce: 1. Execute "rpmbuild --rebuild gcc-4.1.0-4.src.rpm". Actual results: Configure stage aborts with the following output: "configure: error: The following requested languages could not be built: ada Recognised languages are: c,ada,c++,fortran,java,objc,obj-c++,treelang" Expected results: The package should have compiled cleanly. Additional info: This worked up to FC5 final. I got aware of the issue when I wanted to update the package to gcc-4.2.0-20060401. Here, compilation aborts but as pointed out above also already for the current offical rawhide SRPM. Package "gcc-gnat-4.1.0-4" *is* installed.
Sorry, in "Description of problem:", I obviously meant " .. during the configure stage."
Works for me just fine.
Your system is really, really an FC6/rawhide one, I mean, you do not compile "gcc-4.1.0-4.src.rpm" (FC6/rawhide package) on FC5 or other?
Yes, it was a rawhide system, with *4.1.0-4* installed.
Well, some progress on the GCC front: I reverted my system to FC5 from scratch. All volumes formatted, no additional or customized dot files, one plain FC5 test user account. My default shell is "csh" but "bash" makes no difference. My local ".rpmmacros" file reads: %_usrsrc /home/test/build %_tmppath /home/test/build/var/tmp and the corresponding directory tree has been created in test's home directory in order to avoid building the package as "root". Result after installing FC5 w/o any updates: "checking for gnatbind... gnatbind checking whether compiler driver understands Ada... yes" After fully updating the system as of 04/04/06 and booting the system again, the Ada driver stops working (!): "checking for gnatbind... gnatbind checking whether compiler driver understands Ada... no" In the first case, the system boots "kernel-2.6.15-1.2054_FC5", in the second case "kernel-2.6.16-1.2080_FC5". After rebooting the updated system again and booting "kernel-2.6.15-1.2054_FC5", case #1 occurs, thus no error message. This is no delayed April's fool day hoax! Rebooting the system several times confirms that for the original FC5 kernel, the configure script passes successfully when checking Ada support, and that for the update kernel the script fails. Any clues?
The SRPM that failed to compile now of course is "gcc-4.1.0-3.src.rpm".
*** This bug has been marked as a duplicate of 187853 ***
Isn't DUPLICATE marking handled according to the chronological order of the involved bug reports? Just wondering .. . This somehow always suggests that you hadn't checked the existing bug reports prior to filing your own (duplicate) one.