This is the mail archive of the
libstdc++@gcc.gnu.org
mailing list for the libstdc++ project.
Re: multiple definition of `std::__default_alloc_template<(bool)1,(int)0>::allocate(unsigned)'
- From: Loren James Rittle <rittle at latour dot rsch dot comm dot mot dot com>
- To: pfeifer at dbai dot tuwien dot ac dot at
- Cc: libstdc++ at gcc dot gnu dot org
- Date: Tue, 30 Apr 2002 17:56:49 -0500 (CDT)
- Subject: Re: multiple definition of `std::__default_alloc_template<(bool)1,(int)0>::allocate(unsigned)'
- Organization: Networks and Infrastructure Lab (IL02/2240), Motorola Labs
> To ensure mainline hasn't regressed too much, today I test-bootstrapped
> that instead of the 3.1-branch and got the following error when building
> an application of our on i386-unknown-freebsd4.5:
>
> plan.o: In function `std::__default_alloc_template<(bool)1,
>(int)0>::allocate(unsigned)':
> plan.o(.text+0x20b6): multiple definition of
>`std::__default_alloc_template<(bool)1, (int)0>::allocate(unsigned)'
> grounding.o(.text+0x215a): first defined here
>
> Any ideas?
Hi Gerald,
I have no guess at the moment. I bootstrapped gcc mainline last night
on i386-unknown-freebsd4.5 and i386-unknown-freebsd5.0. No such
messages appear in the testsuite log. The 4.5 bootstrap was against
binutils 2.12.X taken last Friday; the 5.0 bootstrap was againt
/usr/bin/{as,ld} (basically, FSF 2.12.0 with few local patches). I am
quickstrapping all my active trees (with a complete rebuild of
libstdc++-v3).
OTOH, the testsuite probably doesn't cover your case (two C++ files
linked together). Small test case?
BTW, do you bootstrap gcc against /usr/bin/{as,ld}? My current
suggestion on FreeBSD is to use that or binutils 2.12.X (taken from
that branch within the last few days).
Regards,
Loren