On Thu, Mar 11, 2010 at 6:53 AM, Lisandro Dalcin <[email protected]> wrote: > On 11 March 2010 03:38, Ondrej Certik <[email protected]> wrote: >> On Tue, Mar 9, 2010 at 5:03 PM, Lisandro Dalcin <[email protected]> wrote: >>> Folks, I've been doing some housekeeping work in the testsuite to make >>> GCC warnings under control. In the process, I've added a few hacks to >>> remove warnings from unused stuff coming from NumPy headers. My fixes >>> work for NumPy 1.2.x and above, but I've not tested with older >>> versions... If anyone is running Cython testsuite with an older NumPy >>> and have any trouble, let me know and I'll fix things. >> >> This is the output of my tests: >> > > Many thanks... > >> >> warning: /home/ondrej/repos/cython-devel/Cython/Runtime/refnanny.pyx:90:5: >> Unraisable exception in function 'refnanny.GOTREF'. >> warning: /home/ondrej/repos/cython-devel/Cython/Runtime/refnanny.pyx:103:5: >> Unraisable exception in function 'refnanny.GIVEREF_and_report'. >> warning: /home/ondrej/repos/cython-devel/Cython/Runtime/refnanny.pyx:129:5: >> Unraisable exception in function 'refnanny.FinishContext'. > > This one I'm not sure about (since long time ago).. Looking at the > code, it seems that exception handling is done the right way, but > perhaps I'm missing something.. The idea is that these functions > absolutely cannot exit with an exception. > >> specmethdocstring.c:424: warning: >> ‘__pyx_doc_17specmethdocstring_1C___init__’ defined but not used >> specmethdocstring.c:446: warning: >> ‘__pyx_doc_17specmethdocstring_1C_3foo___get__’ defined but not used >> specmethdocstring.c:465: warning: >> ‘__pyx_doc_17specmethdocstring_1C_3foo___set__’ defined but not used >> specmethdocstring.cpp:424: warning: >> ‘__pyx_doc_17specmethdocstring_1C___init__’ defined but not used >> specmethdocstring.cpp:446: warning: >> ‘__pyx_doc_17specmethdocstring_1C_3foo___get__’ defined but not used >> specmethdocstring.cpp:465: warning: >> ‘__pyx_doc_17specmethdocstring_1C_3foo___set__’ defined but not used >> autotestdict.c:966: warning: >> ‘__pyx_doc_12autotestdict_11MyCdefClass___cinit__’ defined but not >> used >> autotestdict.c:988: warning: >> ‘__pyx_doc_12autotestdict_11MyCdefClass___dealloc__’ defined but not >> used >> autotestdict.c:1004: warning: >> ‘__pyx_doc_12autotestdict_11MyCdefClass___richcmp__’ defined but not >> used >> autotestdict.c:1024: warning: >> ‘__pyx_doc_12autotestdict_11MyCdefClass___nonzero__’ defined but not >> used >> autotestdict.cpp:966: warning: >> ‘__pyx_doc_12autotestdict_11MyCdefClass___cinit__’ defined but not >> used >> autotestdict.cpp:988: warning: >> ‘__pyx_doc_12autotestdict_11MyCdefClass___dealloc__’ defined but not >> used >> autotestdict.cpp:1004: warning: >> ‘__pyx_doc_12autotestdict_11MyCdefClass___richcmp__’ defined but not >> used >> autotestdict.cpp:1024: warning: >> ‘__pyx_doc_12autotestdict_11MyCdefClass___nonzero__’ defined but not >> used >> extpropertyref.c:416: warning: >> ‘__pyx_doc_14extpropertyref_4Spam_4eggs___get__’ defined but not used >> extpropertyref.cpp:416: warning: >> ‘__pyx_doc_14extpropertyref_4Spam_4eggs___get__’ defined but not used >> ---------------------------------------------------------------------- > > These are innocent warnings... These are docstrings from special > methods... Cython generates them, but as Python C-API does not have a > place to put them, then they are unused in the C code... > > Mmm... perhaps we should not emit these strings in the C code? After > all, we know they will not be used... What's the point of emitting > them?
They should not be emitted imho. All gcc warnings suck. Ondrej _______________________________________________ Cython-dev mailing list [email protected] http://codespeak.net/mailman/listinfo/cython-dev
