Hello, With the basic configure support and null-gc fix applied, I've rebased the series and provide alternative patches for the configure libm and console-unix FIONREAD issues.
Concerning a potential serial-null.c file, this would require to move some enum definitions to a shared header, say serial.h. The Makefile currently distinguishes between Windows and Unix hosts and has serial.c for Unix only; I have therefore opted to provide a separate implementation inside existing serial.c as intermediate step. Feedback on whether the return values are okay as no-ops would be appreciated. The conversion of the signal macros to functions is not yet done. Is anything wrong with the mono-config patch? Regards, Andreas Andreas Färber (10): configure: Don't hardcode libm dependency config: Add support for Haiku mini: Add support for BeOS-style signal handlers mini: Fix unguarded uses of siginfo_t Handle lack of SA_SIGINFO io-layer: Define makedev for Haiku console-unix: Compile fix for Haiku serial: Stub implementation for Haiku eglib: Workaround hardcoded library dependencies for Haiku configure: Fix linking against eglib configure.in | 11 +++++-- eglib/src/Makefile.am | 2 +- eglib/test/Makefile.am | 2 +- mono/io-layer/processes.c | 3 ++ mono/metadata/ChangeLog | 9 +++++- mono/metadata/console-unix.c | 39 ++++++++++++++++------ mono/metadata/mono-config.c | 2 + mono/mini/ChangeLog | 21 ++++++++++++ mono/mini/mini-posix.c | 15 +++++++- mono/mini/mini-x86.c | 8 +++++ mono/mini/mini-x86.h | 18 ++++++++++ mono/mini/mini.h | 6 +++ support/serial.c | 72 ++++++++++++++++++++++++++++++++++++++++++ 13 files changed, 189 insertions(+), 19 deletions(-) _______________________________________________ Mono-devel-list mailing list [email protected] http://lists.ximian.com/mailman/listinfo/mono-devel-list
