Jeff Whitaker wrote:
>
>>Ronald Florence wrote:
>>> formulabase.h: In function `void
>>> __static_initialization_and_destruction_0(int,
>>> int)':
>>> formulabase.h:32: no matching function for call to
>>> `std::type_info::type_info()
>>> '
>>> /usr/include/gcc/darwin/3.1/g++-v3/typeinfo:78: candidates are:
>>> std::type_info::type_info(const char*)
>>> /usr/include/gcc/darwin/3.1/g++-v3/typeinfo:72:
>>> std::type_info::type_info(const std::type_info&)
>
>>Angus Leeming wrote:
>>
>>Is it getting confused with type_info in ref_inset.h?
>> struct type_info {
>> ///
>> string latex_name;
>> ///
>> string gui_name;
>> ///
>> string short_gui_name;
>> };
>>
>>What happens if you comment out #include "ref_inset.h" in
>>formulabase.C. (Yeah, yeah, you'll get other errors, but does this
>>one go away?)
>
> Angus: I'm having this problem too - and no, it doesn't help to
> comment out include "ref_inset.h" in formulabase.C.
>
> -Jeff
Hello, Jeff. I'm afraid that this appears to be a bug in the gcc
shipped by Apple. Apparently, it's the same bug as SuSE had in their
version of the compiler and they have squashed it. Also apparently,
Apple know about the problem, it has a bug number, but they haven't
addressed it yet.
I'm totally ignorant, but have gleaned this information from the lyx
users' list. I would suggest though that Apple will probably fix the
problem in the not too distant future.
If you can't wait that long, I understand that you should download an
'official' gcc from the fsf and compile that. (Do they have
precompiled binaries there too?)
HTH,
--
Angus
