My first guess is that it's some issue with SWIG/gcc. What version of
gcc and swig are you using? Swig should be generating code that gcc
doesn't complain about, however that is not always the case. The other
option is that the error is resulting from some optimization that
exposes a bug in our code. Do you get the same error if you compile a
debug build?
What is the source code to the swig::Py_Sequence_Ref<T>::operator T()
function in build/ALPHA_SE/params/params_wrap.cc? My version doesn't
have a variable v in it.
You could compile the source without -Werror and the build would
complete since it's only a warning and not an error.
Ali
On Dec 6, 2007, at 1:48 PM, Nicolas Zea wrote:
After going through and changing python parameters references to use
the "const <object>Params *p" format in the constructor, and
removing the preprocessor python things at the bottom of the file,
I'm getting the following compile error:
g++ -o build/ALPHA_SE/params/params_wrap.o -c -pipe -fno-strict-
aliasing -Wall -Wno-sign-compare -Werror -Wundef -g -O3 -
DTHE_ISA=ALPHA_ISA -DTRACING_ON=1 -Iext/dnet -I/opt/local/Library/
Frameworks/Python.framework/Versions/2.4/include/python2.4 -Ibuild/
libelf -Ibuild/ALPHA_SE build/ALPHA_SE/params/params_wrap.cc
cc1plus: warnings being treated as errors
build/ALPHA_SE/params/params_wrap.cc: In member function
‘swig::PySequence_Ref<T>::operator T() const [with T = int]’:
build/ALPHA_SE/params/params_wrap.cc:3217: warning: ‘v’ may be used
uninitialized in this function
scons: *** [build/ALPHA_SE/params/params_wrap.o] Error 1
scons: building terminated because of errors.
Can anyone shed some light on what may be going wrong here? The
simulator compiled fine before I tried to transfer over this custom
object, but I'm not familiar with swig so am unsure how to go about
finding the source of this problem.
-Nick_______________________________________________
m5-users mailing list
[email protected]
http://m5sim.org/cgi-bin/mailman/listinfo/m5-users
_______________________________________________
m5-users mailing list
[email protected]
http://m5sim.org/cgi-bin/mailman/listinfo/m5-users
