Clang buildworld failure due to multiple definitions of __isnanf

[Eric McCorkle]

On 6/28/11 1:42 AM, Benjamin Kaduk wrote:
> On Tue, 28 Jun 2011, Benjamin Kaduk wrote:
>
>> On Mon, 27 Jun 2011, Eric McCorkle wrote:
>>
>>> make.conf:
>>> CPUTYPE?=core2
>>> .if !defined(CC) || ${CC} == "cc"
>>> CC=clang
>>> CFLAGS=-Qunused-arguments
>>> .endif
>>> .if !defined(CXX) || ${CXX} == "c++"
>>> CXX=clang++
>>> .endif
>>> NO_WERROR=
>>> WERROR=
>>> NO_FSCHG=
>>> PERL_VERSION=5.12.3
>>>
>>> Just noticed, the CFLAGS would disable optimization, which would
>>> explain why no one else seems to see this. Still, I think the
>>> underlying issue warrants investigation.
>>
>> I seem to recall that our gcc-4.2 with -O0 is known to be buggy --
>> certainly I just tripped over a panic in a third-party kernel module
>> build using bsd.kmod.mk with -O0 that was fixed by compiling the
>> object in question with -O.
>> It is not terribly easy to search for, but my archives find at least
>> http://lists.freebsd.org/pipermail/freebsd-current/2007-August/076028.html
>>
>>
>> I would try with -O1 and confirm that your issues go away, and not
>> spend time on the core issue if it proves to be a gcc bug.
>>
>> Unfortunately, we are stuck with an old gcc in base ... though I have
>> yet to see if clang -O0 has issues with my module.
>
> Wow, I juxtaposed the two compilers due to working on too little sleep.
> Please disregard, and sorry for the noise.
>

Getting rid of the CFLAGS in my make.conf got rid of the multiple 
definitions (I still think the underlying issue merits investigation and 
resolution) for both compilers.