https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=250702
--- Comment #8 from Dimitry Andric <d...@freebsd.org> --- The simple case, e.g. '_ZZ5func1vENK3$_0clEv' has regressed with https://sourceforge.net/p/elftoolchain/code/3531/ ("libelftc: revamped demangler support for function return type"), aka https://github.com/elftoolchain/elftoolchain/commit/68c2224b7 . The https://sourceforge.net/p/elftoolchain/tickets/539/ has another example however, '_ZZN9libunwind17LocalAddressSpace18findUnwindSectionsEjRNS_18UnwindInfoSectionsEENUlP12dl_phdr_infojPvE_8__invokeES4_jS5_', which GNU c++filt unpacks to 'libunwind::LocalAddressSpace::findUnwindSections(unsigned int, libunwind::UnwindInfoSections&)::{lambda(dl_phdr_info*, unsigned int, void*)#1}::__invoke(dl_phdr_info*, unsigned int, void*)'. With the elftc demangler, before upstream r3531, it simply returns "not a valid mangled name", while after upstream r3531, it asserts "(ddata->output.size > 0)". What seems to happen is that r3531 added a cpp_demangle_data::cur_output member, and this is being used for any cpp_demangle_push_str() call. When cpp_demangle_read_sname() reads the '5func1' part of the original string, it calls cpp_demangle_push_str to push the 'func1' name that it decodes, but at that point cur_output is *not* pointing to ddata->output, which the assert expects. -- You are receiving this mail because: You are the assignee for the bug. _______________________________________________ freebsd-toolchain@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-toolchain To unsubscribe, send any mail to "freebsd-toolchain-unsubscr...@freebsd.org"
