Re: new ffi tests fail

[Sam S. <sd...@gn...>]

> * Bruno Haible <oe...@py...t> [2004-08-27 22:47:27 +0200]:
>
> There is absolutely no need to have a /lib/libc.so file. You need to
> understand the roles of these files:
>
>   - *.so.X.Y files are used by the dynamic linker (i.e. at program startup
>     and at dlopen() time). The version number here is mandatory because
>     it must be avoided to link dynamically against a shared library with
>     an incompatible ABI, compared to the one that was intended.
>   - *.so files are used by the linker (ld). These can be symlinks to
>     *.so.X.Y files, or linker scripts (such as /usr/lib/libc.so).

some functions are ABI independent.

>> # cd /lib; ln -s libc.so.6 libc.so
>
> This will not work on GNU/Hurd, where /usr == / and /lib/libc.so is a
> linker script.

/lib == /usr/lib on cygwin too.
if hurd does not have /lib/libc.so, it's not a unix.

> Btw, why do you attempt to use MSVCRT.DLL on Cygwin, in ffi.tst? Why
> use the Microsoft broken malloc when you have a GNU malloc in
> cygwin.dll?

OK, I will change that.

BTW, I don't think it's GNU malloc there.
since they can license cygwin1.dll commercially, I don't think there is
any FSF code there.

-- 
Sam Steingold (http://www.podval.org/~sds) running w2k
<http://www.camera.org> <http://www.iris.org.il> <http://www.memri.org/>
<http://www.mideasttruth.com/> <http://www.honestreporting.com>
Good programmers treat Microsoft products as damage and route around it.