On Wed, 22 Jan 2014, Torbjorn Granlund wrote:
Marc Glisse <marc.gli...@inria.fr> writes:By the way, do we have a policy about breaking binary compatibility? In this case, mixing old and new objects could result in crashes (almost certainly at -O0, seldom at -O3). It should be possible to prevent this issue by renaming __gmp_unary_expr and __gmp_binary_expr (say adding a '2' at the end). It wouldn't be 100%, but I think only questionable code would break. I don't think we need to support compiling foo.cc with one GMP release and bar.cc with another, and then link the objects to some libgmpxx.{so,a}.
Ok, but note that the issue doesn't stop at mixing .o directly, linking with various .so (unless they used visibility options to isolate themselves) or .a that weren't compiled with the same version of gmpxx would break, so in practice that's really a forced recompile of any library using gmpxx, same as a completely new ABI.
-- Marc Glisse _______________________________________________ gmp-devel mailing list gmp-devel@gmplib.org https://gmplib.org/mailman/listinfo/gmp-devel
