[Mingw-users] [Fwd: Re: [xscripts] build of win32api fails]

[Oliver B. <oli...@ae...>]

Apparently my mail original reply didn't reach the mailing list...


-------- Original Message --------
Subject: Re: [Mingw-users] [xscripts] build of win32api fails
Date: Fri, 12 Jun 2009 10:47:46 +0200
From: Oliver Bock <oli...@ae...>
To: MinGW Users List <min...@li...>
References: <4A3...@ae...>
<200...@us...>
<4A3...@gm...>

On 12.06.2009 00:03, JonY wrote:
> On 6/12/2009 04:02, Keith Marshall wrote:
>> On Thursday 11 June 2009 18:17:05 Oliver Bock wrote:
>>> When I use a current CVS checkout of the "xscripts" cross-compile
>>> framework, it fails when it tries to build the win32api:
>> FWIW, a current CVS checkout should be no different from the current
>> 1.0 release.
>>
>>> -----snip-----
>>> ./x86-mingw32-build.sh: stage 1: build w32api ...
>>> checking build system type... i686-pc-linux-gnu
>>> checking host system type... i586-pc-mingw32
>>> checking target system type... i586-pc-mingw32
>>> checking for a BSD-compatible install... /usr/bin/install -c
>>> checking for i586-pc-mingw32-gcc... /usr/bin/gcc-4.2
>>> checking for i586-pc-mingw32-ar... i586-pc-mingw32-ar
>>> checking for i586-pc-mingw32-as... i586-pc-mingw32-as
>>> checking for i586-pc-mingw32-ranlib... i586-pc-mingw32-ranlib
>>> checking for i586-pc-mingw32-ld... i586-pc-mingw32-ld
>>> checking for i586-pc-mingw32-dlltool... i586-pc-mingw32-dlltool
>>> checking for i586-pc-mingw32-windres... i586-pc-mingw32-windres
>>> -----snip-----
>>>
>>> As you can see the problem seems to be the configure script which
>>> fails to detect/find the cross-compile GCC (i586-pc-mingw32-gcc)
>> Yes, I can see that.
>>
>>> which was built and installed successfully in the previous step of
>>> the script.
>> Strange indeed, if it was installed correctly.  The binutils clearly
>> have been been correctly installed, and are detected; are you *sure*
>> that the compiler also has?
>>
>>> Instead the output shows that the host GCC
>>> (/usr/bin/gcc-4.2) is found/used.
>> Not for me!  I just ran it again, from my own CVS sandbox; this is
>> what I see:
>>
>> x86-mingw32-build.sh: stage 1: build w32api ...
>> checking build system type... i686-pc-linux-gnu
>> checking host system type... i586-pc-mingw32
>> checking target system type... i586-pc-mingw32
>> checking for a BSD-compatible install... /usr/bin/install -c
>> checking for i586-pc-mingw32-gcc... i586-pc-mingw32-gcc
>> checking for i586-pc-mingw32-ar... i586-pc-mingw32-ar
>> checking for i586-pc-mingw32-as... i586-pc-mingw32-as
>> checking for i586-pc-mingw32-ranlib... i586-pc-mingw32-ranlib
>> ...
>>
>> So, why should it behave differently for you?
>>
>> For reference, my x86-mingw32-build.sh.conf says:
>>
>>       assume GCC_VERSION                   3.4.5-20060117-2
>>       assume BINUTILS_VERSION              2.19.1
>>       assume RUNTIME_VERSION               3.15.2
>>       assume W32API_VERSION                3.13
>>
> 
> Hi,
> Do you have CC defined in your environment variables?

Duh! A hidden leftover of some previous experiment! Should have used a
fresh shell :-$

Thanks,
Oliver