> "import site failed" > OverflowError: signed integer is greater than the maximum. > > > This is happening in the convertsimple() routine when it tries to > return a signed int: > > ival = PyInt_AsLong(arg) > > the ival is larger than what is defined in INT_MAX. > > Why is this happening in a standard HP 64 bit build?
Can you provide a complete gdb/dbx backtrace? Some function tries to convert a Python int into a C int, using the "i" conversion character. Python int uses C long for internal representation, and that particular C long happens to be larger than INT_MAX. This is quite reasonable to happen in principle, but shouldn't happen on interpreter startup. So the questions are: - what are the callers of convertsimple here? (in particular, where does the call to PyArg_ParseTuple come from?) - what is the value of ival? - where does that number come from? The first two questions are best answered with a C debugger. Depending on the answer, the third question may nor may not need an answer. Good luck, Martin P.S. If you are asking in the more abstract sense "why is that happening to me?", the answer is "because you are using an uncommon platform on which Python sees little or no testing". To work around, try a 32-bit build, or switch to Solaris, OS X, Debian Linux, or (heaven forbid) MS Windows :-) -- http://mail.python.org/mailman/listinfo/python-list
