[Bug sanitizer/61314] Building GCC 4.9.0 breaks in libbacktrace on Ubuntu Lucid Lynx
gk at torproject dot org
gcc-bugzilla@gcc.gnu.org
Mon May 26 20:01:00 GMT 2014
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61314
--- Comment #2 from Georg Koppen <gk at torproject dot org> ---
Created attachment 32856
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=32856&action=edit
config.log
Attached is the config.log. Here comes some additional information that might
help to understand this bug. The build failure does not happen when running
|make| but |make install| which is surprising. It might be due to our
deterministic build setup which uses Gitian and libfaketime. We had and still
have a lot of trouble compiling GCC when using libfaketime and what we are
currently doing is compiling GCC with -j1 we seems to avoid the problems
involved in faking the timestamps. Until now. The lines immediately before the
pieces I posted above are:
make[2]: Leaving directory `/home/ubuntu/build/gcc/gcc'
make[2]: warning: Clock skew detected. Your build may be incomplete.
make[2]: Entering directory `/home/ubuntu/build/gcc/intl'
make[2]: Warning: File `Makefile' has modification time 0.23 s in the future
make[2]: Nothing to be done for `install'.
make[2]: warning: Clock skew detected. Your build may be incomplete.
make[2]: Leaving directory `/home/ubuntu/build/gcc/intl'
make[2]: Entering directory `/home/ubuntu/build/gcc/libbacktrace'
make[2]: Warning: File `dwarf.lo' has modification time 0.035 s in the future
Here is what seems to happen: the build system seems to realize that there
might have been issues in the build and is recompiling dwarf.c during the
make-install-step but not using the compiler the configure was for but the one
Lucid ships. That this is a non-issue for compiling on Precise might be due to
me not compiling 4.9.0 in the Gitian/libfaketime environment. Does this make
some sense? If so, do you see a workaround for us here? Pointing to the
compiler the configure setup was for after |make| ran? Or...?
More information about the Gcc-bugs
mailing list