On Sat, 2006-05-20 at 10:22 -0400, Steve M. Robbins wrote: > Otherwise, it would help me if you could point out what part of gmp > you believe has changed. In this bug report, the errors have to do > with missing methods of obby::RSA::Key: [snip] > Since you know the obby source better than I, maybe you could let me > know what parts of gmp are implicated. For example, what is the type > of obby::RSA::Key::set_n()'s argument after the rebuild?
Well, no problem: The version currently in the archives provides set_n with the following signature (provides by objdump -T -C): obby::RSA::Key::set_n(__gmp_expr<__gmpz_value, __gmpz_value> const&) The recompiled version has instead this one: obby::RSA::Key::set_n(__gmp_expr<__mpz_struct [1], __mpz_struct [1]> const&) So the ABI changed and gmp's development packages need to be renamed, as we just use "set_n(const mpz_class& n)" in the C++ source. We did not rely on internals. Do you agree with my observation? If so, should I file a RC bug against gmp for tracking-purposes? Thanks for your help, Philipp Kern -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
