This is the mail archive of the
java@gcc.gnu.org
mailing list for the Java project.
Re: patch to enable inet_addr() on Win32 [ok for mainline?]
- From: Mark Mitchell <mark at codesourcery dot com>
- To: Adam Megacz <gcj at lists dot megacz dot com>
- Cc: Bryce McKinlay <bryce at waitaki dot otago dot ac dot nz>, "java at gcc dot gnu dot org" <java at gcc dot gnu dot org>, "oliva at redhat dot com" <oliva at redhat dot com>
- Date: Tue, 28 May 2002 00:22:00 -0700
- Subject: Re: patch to enable inet_addr() on Win32 [ok for mainline?]
--On Tuesday, May 28, 2002 12:05:04 AM -0700 Adam Megacz
<gcj@lists.megacz.com> wrote:
>
> Mark Mitchell <mark@codesourcery.com> writes:
>> >> Okay, I just took a look at this. Turns out that gcj's configury uses
>> >> AC_CHECK_FUNCS, which doesn't work right when crosscompiling.
>
>> But why not? AC_CHECK_FUNCS is certainly *supposed* to work when
>> cross-compiling.
>
> The problem is two-fold:
>
> 1) gcj's configury still assumes in many places that if you're
> crosscompiling, you must be targeting newlib. This is wrong, but
> easy to fix.
>
> 2) Even when I fixed (1), configure was still invoking the platform's
> native 'gcc' to test if compiles succeeded.
These problems are important, and need to be fixed. We can't do that
by faking out autoconf for every target; that defeats the point of
autoconf.
It sounds to me that, rather than applying your change, you and the
rest of the GCJ team should work to fix the configury. There's no
reason that every other part of the compiler should work when
cross-compiling, but that GCCJ will not.
Alexandre, can you help with this? You understand better than most of
us how to get TARGET_*_TO_PASS set up correctly, methinks.
--
Mark Mitchell mark@codesourcery.com
CodeSourcery, LLC http://www.codesourcery.com