BTW it is also more effective to use RTLD_LAZY than RTLD_NOW.
In what way is RTLD_LAZY more effective than RTLD_NOW? Using RTLD_NOW can make future latencies lower (or at least more deterministric) by performing most dynamic symbol resolutions immediately in one big batch. This trades a higher initial cost (and possibly more total work, if not all the resolutions will be used) for lower latency in the future. -- -- fedora-devel-list mailing list fedora-devel-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/fedora-devel-list
