Jeremie Courreges-Anglas <j...@wxcvbn.org> writes: > Jeremie Courreges-Anglas <j...@wxcvbn.org> writes: > >> Hi, >> >> Bruno Haible kindly notified us of the release of libffcall-1.13. >> >> http://marc.info/?l=openbsd-ports&m=149830931016250&w=2 >> >> This new release will hopefully let us drop the not-so-nice patches we >> have. >> >> With the following diff, ''make test'' works fine on amd64 and arm. >> ''make test'' passes for lang/clisp and lang/gforth on amd64, lang/clisp >> can now be unlocked on arm. >> >> Test reports are welcome, especially on those architectures: >> >> alpha hppa i386 mips64 mips64el powerpc sparc64 >> >> ok/comments? > > > Updated diff that copes with the renaming of cl_cv_c_struct_return_small > to ffcall_cv_c_struct_return_small. Someone with a sparc64 at hand > could delete the sparc64-specific hack in the Makefile and check that > everything still works fine.
The list of tested architectures is now: - amd64: ok - arm: ok - sparc64: ok, the vacall test segfaults, like in -current - powerpc: the vacall test dies, like in -current. The callback test now errors out in a strange way. lang/gforth is happy with 1.13. - mips64el: ok - mips64: ugly sigbus/segfaults/double free in ar, ranlib, perl on GENERIC-IP32 -> I shamelessly ran away NB: vacall isn't installed, so failures don't matter much. -- jca | PGP : 0x1524E7EE / 5135 92C1 AD36 5293 2BDF DDCC 0DFA 74AE 1524 E7EE
