On Sun, 04 Oct 2015 12:04:18 -0400 Scott Kitterman <[email protected]> wrote: > Package: python-cbor > Version: 0.1.21-1 > Severity: important > > It is a best practice to build extension modules for all supported python > versions. Now that python3.5 is supported, it should be a simple matter > of changing the python3-all/python3-dev build-deps to python3-all-dev, except > that the compile fails with python3.5. > > creating build/temp.linux-x86_64-3.5 > creating build/temp.linux-x86_64-3.5/c > x86_64-linux-gnu-gcc -pthread -DNDEBUG -g -fwrapv -O2 -Wall -Wstrict- prototypes -g -O2 -fstack-protector-strong -Wformat -Werror=format-security - D_FORTIFY_SOURCE=2 -fPIC -Ic/ -I/usr/include/python3.5m -c c/cbormodule.c -o build/temp.linux-x86_64-3.5/c/cbormodule.o > c/cbormodule.c: In function 'decodeFloat64': > c/cbormodule.c:148:5: warning: dereferencing type-punned pointer will break strict-aliasing rules [-Wstrict-aliasing] > return PyFloat_FromDouble(*((double*)(&aux))); > ^ > c/cbormodule.c: In function 'inner_dumps': > c/cbormodule.c:1200:2: warning: dereferencing type-punned pointer will break strict-aliasing rules [-Wstrict-aliasing] > tag_u64_out(CBOR_7, *((uint64_t*)(&val)), out, &pos); > ^ > c/cbormodule.c: In function 'PyInit__cbor': > c/cbormodule.c:1412:10: error: 'PyModuleDef {aka struct PyModuleDef}' has no member named 'm_reload' > modef.m_reload = NULL; > ^ > > BUILD ERROR: > command 'x86_64-linux-gnu-gcc' failed with exit status 1 > > Note that this will be severity serious when python3.5 is made the default > python3, but it would be nice to get it fixed before that and support all > versions. Multiple version support greatly eases changing the default.
Actually, now that python3.5 is default, it fails to compile the C extension for either python3.4 or python3.5 so a usable (but much slower) package results. I'll leave this as an important bug as I think it's a significant regression, but not RC after all. Scott K
