ACk, honza can you please merge it? Thanks again for the work! Fabio On 4/3/2013 11:45 AM, Andrei Belov wrote: > > On Apr 3, 2013, at 13:12 , Fabio M. Di Nitto <fdinitto@xxxxxxxxxx> wrote: > >> Hi Andrei, >> >> On 4/3/2013 10:55 AM, Andrei Belov wrote: >> >>>> We spent a huge amount of time moving to "detect features" vs >>>> "hardcoding" because it´s a lot more flexible and lasts longer. this >>>> approach is moving backwards. >>> >>> Thanks for the explanation - I see why this makes sense now. >>> >>> Please find the improved patch attached. I've just tested it on CentOS 6.0, >>> SunOS 5.11 and Mac OS X 10.8.3, and it seems to work as expected. >>> >>> GCC versions are below: >>> >>> CentOS: gcc version 4.4.4 20100726 (Red Hat 4.4.4-13) (GCC) >>> SunOS: gcc version 4.7.2 (GCC) >>> Mac OS X: gcc version 4.2.1 (Based on Apple Inc. build 5658) (LLVM build 2336.11.00) >> >> Much better patch, only a few minor bits: >> >> +if test "$ap_cv_cc_as_needed" = "yes"; then >> + SEC_LDFLAGS="$SEC_LDFLAGS -Wl,--as-needed" >> +fi >> >> please don´t overload SEC_LDFLAGS. SEC stands for SECURITY (see other >> configure.ac bits related). >> At that stage of the check you can simply add it to LD_FLAGS if it´s >> supported. > > Oh, my fault with SEC - I thought it means SECondary. :) > > >> For the --version-script: >> >> Please rename EXTRA_LDFLAGS to something more specific. Maybe >> VERSCRIPT_LDFLAGS or something like that. this is merely to avoid >> conflict with those distribution automatic build system that override >> default LD_FLAGS via EXTRA_LDFLAGS and such, and to keep the variable >> isolated. > > Changed to VERSCRIPT_LDFLAGS (VERSION_SCRIPT_LDFLAGS or > LIB_VERSION_SCRIPT_LDFLAGS looks too long). > > >> Nice work btw! Thanks! >> >> Fabio > > Third variant is attached. > > _______________________________________________ discuss mailing list discuss@xxxxxxxxxxxx http://lists.corosync.org/mailman/listinfo/discuss
