Hi,
I recently updated to gcc3.3 from the gcc December2002 update I downloaded from ADC.
Now I have an internal compiler error when compiling director_generic.cpp (which I think is the first file to include boost stuff).
Does anyone have the same kind of errors under MacOS X ? What dev CD version is it ?
Here is my ouput :
g++ -DHAVE_CONFIG_H -I. -I. -I../../.. -I../../../src -I../../../src/boost -DWITH_MACOSX_BUNDLE=1 -framework OpenGL -I/System/Library/Frameworks/Carbon.framework/Libraries/CIncludes -framework GLUT -I/sw/include -I/sw/include -I/sw/include -DJPEG_SUPPORT -DHAVE_AL=1 -I/Library/Frameworks/OpenAL.framework/Headers -DNETCOMM_NOSOUND -DNETCOMM -I/System/Library/Frameworks/QuickTime.framework/Headers -I/usr/include/python2.2 -DHAVE_PYTHON=1 -DUSE_BOOST_128=1 -gdwarf-2 -g3 -Wall -pipe -c -o director_generic.o `test -f 'director_generic.cpp' || echo './'`director_generic.cpp
cc1plus: warning: `dwarf-2': unknown or unsupported -g option
In file included from ../../../src/boost/boost/type_traits/conversion_traits.hpp:26,
from ../../../src/boost/boost/type_traits.hpp:15,
from ../../../src/boost/boost/function/function_base.hpp:25,
from ../../../src/boost/boost/function.hpp:22,
from ../../../src/boost/boost/python/errors.hpp:14,
from ../../../src/boost/boost/python/reference.hpp:19,
from ../../../src/boost/boost/python/objects.hpp:14,
from ../../../src/python/python_class.h:19,
from director_generic.cpp:11:
../../../src/boost/boost/type_traits/arithmetic_traits.hpp:83: warning: use of
`long double' type; its size may change in a future release
../../../src/boost/boost/type_traits/arithmetic_traits.hpp:83: warning: (Long
double usage is reported only once for each file.
../../../src/boost/boost/type_traits/arithmetic_traits.hpp:83: warning: To
disable this warning, use -Wno-long-double.)
../../../src/boost/boost/function/function_template.hpp: In member function `
void boost::function0<R, Policy, Mixin, Allocator>::clear() [with R = void,
Policy = boost::empty_function_policy, Mixin = boost::empty_function_mixin,
Allocator = std::allocator<boost::function_base>]':
../../../src/boost/boost/function/function_template.hpp:290: instantiated from `boost::function0<R, Policy, Mixin, Allocator>::~function0() [with R = void, Policy = boost::empty_function_policy, Mixin = boost::empty_function_mixin, Allocator = std::allocator<boost::function_base>]'
../../../src/boost/boost/python/errors.hpp:28: instantiated from `bool boost::python::handle_exception(T) [with T = void (*)()]'
../../../src/boost/boost/python/errors.hpp:35: instantiated from here
../../../src/boost/boost/function/function_template.hpp:364: internal compiler error: in
c_expand_expr, at c-common.c:4752
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://developer.apple.com/bugreporter> for instructions.
{standard input}FATAL:.abort detected. Assembly stopping.
make[3]: *** [director_generic.o] Error 1
make[2]: *** [all-recursive] Error 1
make[1]: *** [all-recursive] Error 1
make: *** [all-recursive] Error 1
Problem with MacOS X
-
- Developer
- Posts: 153
- Joined: Wed Jan 22, 2003 4:28 am
- Location: Reunion Island
- Contact:
-
- Lead Network Developer
- Posts: 2560
- Joined: Sun Jan 12, 2003 9:13 am
- Location: Palo Alto CA
- Contact:
Oh that bug...
I think gcc 3.3 needs boost 1.2.9
don't use enable-boost1.2.8 when configuring; use:
./configure --enable-debug ... ... ... --enable-boost129
instead.
boost 1.2.8 relies on some quirky compiler bugs in order to work.
boost 1.2.9 is better about that, but doesn't work on gcc 2.9.x
Looks like you are using boost 1.2.8.gcc wrote:... -I/usr/include/python2.2 -DHAVE_PYTHON=1 -DUSE_BOOST_128=1 -gdwarf-2 -g3 ...
I think gcc 3.3 needs boost 1.2.9
don't use enable-boost1.2.8 when configuring; use:
./configure --enable-debug ... ... ... --enable-boost129
instead.
boost 1.2.8 relies on some quirky compiler bugs in order to work.
boost 1.2.9 is better about that, but doesn't work on gcc 2.9.x
-
- Developer
- Posts: 153
- Joined: Wed Jan 22, 2003 4:28 am
- Location: Reunion Island
- Contact:
Mmm... I'm afraid I already tried with boost129 with the same kind of error.
I've always been using boost128 with gcc3.1 under MacOS X and it always worked fine.
If I remember well Daniel told me explicitly to use boost128 under macosx.
Now about gcc3.3 I don't know maybe it needs boost 129 but this isn't the problem now since even with that one gcc crashes.
But anyway... this is a bad problem I hadn't before with gcc3.1... I will just switch back.
Now I have another (new) internal error in networking stuff I added recently. I'll investigate that one.
I've always been using boost128 with gcc3.1 under MacOS X and it always worked fine.
If I remember well Daniel told me explicitly to use boost128 under macosx.
Now about gcc3.3 I don't know maybe it needs boost 129 but this isn't the problem now since even with that one gcc crashes.
But anyway... this is a bad problem I hadn't before with gcc3.1... I will just switch back.
Now I have another (new) internal error in networking stuff I added recently. I'll investigate that one.
-
- Lead Network Developer
- Posts: 2560
- Joined: Sun Jan 12, 2003 9:13 am
- Location: Palo Alto CA
- Contact:
Hellcatv said that I could use his mac while he was gone, so I decided to see what command line it uses.
Well it gave a billion errors, so I re-checked it out and compiled... Now I am just waiting to see wat happens, as it doesn't seem to give errors.
'll post again when it either works or doesn't work.
EDIT: it is using version 3.1, so i guess that is the only one that dorks
Well it gave a billion errors, so I re-checked it out and compiled... Now I am just waiting to see wat happens, as it doesn't seem to give errors.
'll post again when it either works or doesn't work.
EDIT: it is using version 3.1, so i guess that is the only one that dorks
-
- Developer
- Posts: 153
- Joined: Wed Jan 22, 2003 4:28 am
- Location: Reunion Island
- Contact:
Ok, I'll let gcc 3.3 for macosx aside for now.
I found a way to avoid the compiler error I seem to have caused with gcc3.1.
Everything is back to normal then.
Under linux I'm gcc 3.3 with no problem at all. But as for others I have to use boost 1.2.8 because 1.2.9 indeed produces compilation errors.
I found a way to avoid the compiler error I seem to have caused with gcc3.1.
Everything is back to normal then.
Under linux I'm gcc 3.3 with no problem at all. But as for others I have to use boost 1.2.8 because 1.2.9 indeed produces compilation errors.