Subject: Re: [ruby-ffi] New release of FFI gem |
From: Jon |
Date: 12/3/10 9:12 AM |
To: ruby-ffi@googlegroups.com |
<wmeissner@gmail.com> wrote:On Fri, Dec 3, 2010 at 7:57 AM, Wayne Meissner*think* I fixed up the win32 build. I couldn't test it, since I'm having issues getting rake-compiler to build a cross-ruby on my machine ... and I have no win32 vm to run it in anyway. But, if anyone wants to try building out of master for win32 and letting me know if it all works, that would be nice.btw, Ihttps://gist.github.com/726989 The problem is that is looking for strdup and not _strdup, perhaps the problem is libffi?I was able to compile ffi with Ruby 1.9.2-p0 and GCC 4.5.1, but now getting this on the specs:
whoa..you guys are quick! I've only had time to do the following with master, but seems that ffi_c.so builds fine with both 1.9.2p80 and 1.9.3dev (2010-12-03 trunk 30060) on my Win7 Ultimate 32-bit. No testing yet. :( C:\Users\Jon\Documents\RubyDev\ffi-trunk\ext\ffi_c>ruby -rdevkit extconf.rb Temporarily enhancing PATH to include DevKit... checking for rb_thread_blocking_region()... yes checking for ruby_thread_has_gvl_p()... yes checking for ruby_native_thread_p()... yes checking for rb_thread_call_with_gvl()... yes creating extconf.h creating Makefile C:\Users\Jon\Documents\RubyDev\ffi-trunk\ext\ffi_c>make C:/ruby192/bin/ruby -e "puts 'EXPORTS', 'Init_ffi_c'" > ffi_c-i386-mingw32.def Configuring libffi ....SNIP.... gcc -shared -s -o ffi_c.so AbstractMemory.o ArrayType.o Buffer.o Call.o ClosurePool.o DataConverter.o DynamicLibrary.o ffi.o Function.o FunctionInfo.o LastError.o MappedT ype.o MemoryPointer.o MethodHandle.o Platform.o Pointer.o Struct.o StructByReference. o StructByValue.o StructLayout.o Type.o Types.o Variadic.o -L. -LC:/ruby192/lib -L. -Wl,--enable-auto-image-base,--enable-auto-import ffi_c-i386-mingw32.def /c/Users/Jo n/Documents/RubyDev/ffi-trunk/ext/ffi_c/libffi/.libs/libffi_convenience.a -lmsvcrt-ru by191 -lshell32 -lws2_32 Jon --- http://jonforums.github.com/