python-numpy: fix build for libn32
PATCH 1/1 python-numpy: fix build for libn32
Fix do_compile failure for libn32. To reproduce, use the following config.
MACHINE = "qemumips64"
require conf/multilib.conf MULTILIB_GLOBAL_VARIANTS_append = " libn32" MULTILIBS ?= "multilib:lib32 multilib:libn32" DEFAULTTUNE_virtclass-multilib-lib32 ?= "mips" DEFAULTTUNE_virtclass-multilib-libn32 ?= "mips64-n32"
The error message is as following.
numpy/core/include/numpy/npy_common.h:206:10: error: #error Unsupported size for type off_t
Signed-off-by Chen Qi <[email protected]>
--- .../python-numpy/files/mipsarchn32eb/_numpyconfig.h | 1 + 1 file changed, 1 insertion(+)
meta/recipes-devtools/python-numpy/files/mipsarchn32eb/_numpyconfig.h +1 -0
@@ -12,6 +12,7 @@#define NPY_SIZEOF_PY_INTPTR_T 8#define NPY_SIZEOF_PY_LONG_LONG 8#define NPY_SIZEOF_LONGLONG 8 +#define NPY_SIZEOF_OFF_T 8#define NPY_NO_SMP 0#define NPY_HAVE_DECL_ISNAN #define NPY_HAVE_DECL_ISINF
Signature · 1 line
2.17.1