Python 2.4.3 (and most likely previous versions) has a bug regarding signed integer overflows that is exposed when built with the current trunk of GCC (4.2). Detailed info can be found at [1]. The short version: > Michael Veklser wrote: > > It *is* a bug in python, here is the proof: > https://codespeak.net/viewvc/vendor/cpython/Python-r243/dist/src/Objects/intobject.c?revision=25647&view=markup > Function > > * i_divmod*(*register* *long* x, *register* *long* y, > > the following lines: > > / /* (-sys.maxint-1)/-1 is the only overflow case. *// > *if* (y == -1 && x < 0 && x == -x) > *return* DIVMOD_OVERFLOW; > > If overflow is barred then x==-x may happen only when x==0. > This conflicts with x<0, which means that the compiler may assume > that > x<0 && x==-x > always yields false. This may allow the compiler to eliminate the whole if > statement. Hence, clearly python is at fault. This bug causes the testsuite to fail in test_builtin: test_builtin make: *** [test] Floating point exception LD_LIBRARY_PATH=/var/tmp/portage/python-2.4.3-r4/work/Python-2.4.3: ./python -E -tt ./Lib/test/regrtest.py -l -w test_grammar test_opcodes test_operations test_builtin make: *** [test] Floating point exception It should be reproducible with: dirtyepic@tycho ~ $ python Python 2.4.3 (#1, Oct 11 2006, 19:20:54) [GCC 4.2.0-pre20061006 (experimental) (rev. 117527)] on linux2 Type "help", "copyright", "credits" or "license" for more information. >>> import sys >>> divmod(-sys.maxint-1, -1) Floating point exception The attached patch fixes the testsuite and example. dirtyepic@tycho ~ $ python Python 2.4.3 (#1, Oct 11 2006, 20:00:37) [GCC 4.2.0-pre20061006 (experimental) (rev. 117527)] on linux2 Type "help", "copyright", "credits" or "license" for more information. >>> import sys >>> divmod(-sys.maxint-1, -1) (2147483648L, 0L) >>> (sys.maxint+1, 0) (2147483648L, 0) >>> divmod(-sys.maxint-1, -1) == (sys.maxint+1, 0) True Like I said, this bug is also present in previous python versions, but considering it doesn't really show up unless GCC 4.2 is used, fixing those versions might not be worth it. I haven't checked 2.5-r1. [1] http://thread.gmane.org/gmane.comp.gcc.devel/80582/focus=81424
Created attachment 99611 [details, diff] python-2.4.3-gcc42.patch
*** Bug 148568 has been marked as a duplicate of this bug. ***
Looks like this is fixed in 2.4.4.
