A Sunday 24 February 2008 13:18:58, Danny Backx wrote: > Hi John, > > Very nice work you did ! >
Indeed! > On Sun, 2008-02-24 at 09:23 +0000, john wrote: > > I'm now trying to tie up a few loose ends, and one thing I couldn't sort > > out was an error when trying to build the socket extensions for Ruby - > > one the functions that they use is h_errno; the code compiles cleanly, > > but when it comes to linking the final binary it falls over with an > > undefined reference to "__h_errno_location". I can't find this symbol > > anywhere. > > Well, when you say "sockets" and "Windows" then you need to think > winsock. Combine this with the fact that "errno" is a posix thing which > Windows API doesn't support, and here we are. > > That said, I'm sure we can get through this. Usually when people use > sockets/winsock, the better thing to do is to replace all use of errno > by the winsock equivalent WSAGetLastError(). > He's using cegcc, not mingw32ce, so it's kind of a mismatch to be needing that. cegcc privides an errno and a strerror already, and wraps the socket handling so you'd not think winsock. Well, it's far from perfect, so don't expect it to work without hickups. Since Ruby has been ported to build with MSFT's tools, and that code is upstream, a better solution IMHO, would be to use mingw32ce. I think a lot of code in that wince dir would not be needed with mingw32ce, since we provide it in mingwex. Well, if it's needed, it's there. -- Pedro Alves ------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ _______________________________________________ Cegcc-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/cegcc-devel
