[Buildroot] {buildroot 0004384]: GCC 4.3.1 fails with cannot compute suffix of object files
Maxim Grigoriev
maxim at tensilica.com
Wed Feb 4 04:02:29 UTC 2009
Hello SOlO and everybody,
>> I would be very glad, too hear something regarding this issue.
>>
>> S0l0
The best solution I was able to find is switching to static
linkage of GMP/MPFR for GCC 4.3.x compilers.
It worked out pretty nicely for me.
Here is the patch :
diff -u package/gmp/gmp.mk.orig package/gmp/gmp.mk
--- package/gmp/gmp.mk.orig 2009-01-21 15:43:41.788558000 -0800
+++ package/gmp/gmp.mk 2009-02-03 18:07:44.643402000 -0800
@@ -83,7 +83,7 @@
--prefix="$(GMP_HOST_DIR)" \
--build=$(GNU_HOST_NAME) \
--host=$(GNU_HOST_NAME) \
- --enable-shared \
+ --disable-shared \
--enable-static \
$(DISABLE_NLS) \
)
diff -u package/mpfr/mpfr.mk.orig package/mpfr/mpfr.mk
--- package/mpfr/mpfr.mk.orig 2009-01-21 15:43:42.259089000 -0800
+++ package/mpfr/mpfr.mk 2009-02-03 17:58:09.369092000 -0800
@@ -97,7 +97,7 @@
--prefix="$(MPFR_HOST_DIR)" \
--build=$(GNU_HOST_NAME) \
--host=$(GNU_HOST_NAME) \
- --enable-shared \
+ --disable-shared \
--enable-static \
--with-gmp=$(GMP_HOST_DIR) \
$(DISABLE_NLS) \
-- Maxim
Maxim Grigoriev wrote:
> I think setting LD_LIBRARY_PATH as an extra step is unacceptable,
> even though, it works nice.
>
> I'm working on this. I will definitely share my results,
> when I have a reasonable solution.
>
> -- Maxim
>
>
> S0L0 at gmx-topmail.de wrote:
>> I have the same problem. Build fails and I can not find the
>> patchfile. :(
>>
>> I would be very glad, too hear something regarding this issue.
>>
>> S0l0
>> -------- Original-Nachricht --------
>>
>>> Datum: Sun, 01 Feb 2009 11:05:53 -0800
>>> Von: Maxim Grigoriev <maxim at tensilica.com>
>>> An: Maxim Grigoriev <maxim at tensilica.com>
>>> CC: buildroot at uclibc.org
>>> Betreff: Re: [Buildroot] {buildroot 0004384]: GCC 4.3.1 fails with
>>> cannot compute suffix of object files
>>>
>>
>>
>>> Is this the right mailing list to ask questions about BUILDROOT ?
>>>
>>> Should I ask some official maintainers directly instead ?
>>>
>>> If there is a communication policy described somewhere, please,
>>> refer me to it.
>>>
>>> Thanks,
>>> -- Maxim
>>>
>>>
>>> Maxim Grigoriev wrote:
>>>
>>>> There is a buildroot issue 0004384:
>>>>
>>>>
>>> http://lists.uclibc.org/pipermail/buildroot/2008-September/010792.html
>>>
>>>> I understand it's a generic architecture-independent GCC 4.3.1+
>>>> compiler
>>>> problem
>>>> mentioned several times in different places ( e.g. bugzilla PRs
>>>> 35577, 35693 ).
>>>>
>>>> On the buildroot mailing, list I found a reference to the patch
>>>>
>>>> gcc-4.3.1-compile.patch
>>>> introduced by "antezedens" somewhere around 09-10-08.
>>>> But, I was not able to find this patch. Looks like it was withdrawn.
>>>>
>>>> * * * * *
>>>>
>>>> My question is: is this build failure supposed to be fixed on
>>>> the BUILDROOT makefiles/scripts level ? Or, setting
>>>> LD_LIBRARY_PATH=$PWD/toolchain_build_arm/mpfr/lib:$LD_LIBRARY_PATH
>>>>
>>>> from the top of buildroot tree is assumed to be a standard
>>>> setup for the buildroot build ? Is it documented anywhere ?
>>>>
>>>> Thanks,
>>>> -- Maxim
>>>>
>>>>
>>>>
>>>>
>>>> _______________________________________________
>>>> buildroot mailing list
>>>> buildroot at busybox.net
>>>> http://lists.busybox.net/mailman/listinfo/buildroot
>>>>
>>> _______________________________________________
>>> buildroot mailing list
>>> buildroot at busybox.net
>>> http://lists.busybox.net/mailman/listinfo/buildroot
>>>
>>
>>
>
More information about the buildroot
mailing list