In issues 7238 [1], 16482 [2], 17697 [3] and 17277 [4], the line number may be incorrect when the global trace function has been removed but not the frame f_trace function. A simple test (see below) in issue 17288 [5] crashes the interpreter when setting f_lineno in a generator from a return trace function. All those issues are few months old. There is a patch at issue 17277 [4] that fixes the first 4 issues. There is also a patch for issue 17288 [5]. ###### Setting f_lineno in a generator ###### $ cat jump.py def gen(): for i in range(1): yield i lineno = 4 for i in gen(): pass $ python -m pdb jump.py
/tmp/jump.py(1)<module>() -> def gen(): (Pdb) import sys; print(sys.version) 3.4.0a1+ (default:975d1e180689, Sep 6 2013, 09:26:12) [GCC 4.3.2] (Pdb) break 3 Breakpoint 1 at /tmp/jump.py:3 (Pdb) continue /tmp/jump.py(3)gen() -> yield i (Pdb) step --Return-- /tmp/jump.py(3)gen()->0 -> yield i (Pdb) jump 2 /tmp/jump.py(2)gen()->0 -> for i in range(1): (Pdb) continue Segmentation fault
[1] http://bugs.python.org/issue7238 [2] http://bugs.python.org/issue16482 [3] http://bugs.python.org/issue17697 [4] http://bugs.python.org/issue17277 [5] http://bugs.python.org/issue17288 Xavier