I now get the following failures in test_gdb:

======================================================================
FAIL: test_pyup_command (test.test_gdb.StackNavigationTests)
Verify that the "py-up" command works
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/home/antoine/cpython/__svn__/Lib/test/test_gdb.py", line 599, in test_pyup_command
    $''')
  File "/home/antoine/cpython/__svn__/Lib/test/test_gdb.py", line 163, in assertMultilineMatches
    msg='%r did not match %r' % (actual, pattern))
AssertionError: 'Breakpoint 1 at 0x4508a0: file Objects/object.c, line 330.\n\nBreakpoint 1, PyObject_Print (op=42, fp=0x7ffff7535780, flags=1) at Objects/object.c:330\n330\t\treturn internal_print(op, fp, flags, 0);\n#7 (unable to read python frame information)\n' did not match '^.*\n#[0-9]+ Frame 0x[0-9a-f]+, for file .*gdb_sample.py, line 7, in bar \\(a=1, b=2, c=3\\)\n    baz\\(a, b, c\\)\n$'

======================================================================
FAIL: test_up_at_top (test.test_gdb.StackNavigationTests)
Verify handling of "py-up" at the top of the stack
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/home/antoine/cpython/__svn__/Lib/test/test_gdb.py", line 613, in test_up_at_top
    cmds_after_breakpoint=['py-up'] * 4)
  File "/home/antoine/cpython/__svn__/Lib/test/test_gdb.py", line 129, in get_stack_trace
    self.assertEquals(err, '')
AssertionError: 'Error occurred in Python command.\nError occurred in Python command.\n' != ''

======================================================================
FAIL: test_up_then_down (test.test_gdb.StackNavigationTests)
Verify "py-up" followed by "py-down"
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/home/antoine/cpython/__svn__/Lib/test/test_gdb.py", line 628, in test_up_then_down
    $''')
  File "/home/antoine/cpython/__svn__/Lib/test/test_gdb.py", line 163, in assertMultilineMatches
    msg='%r did not match %r' % (actual, pattern))
AssertionError: 'Breakpoint 1 at 0x4508a0: file Objects/object.c, line 330.\n\nBreakpoint 1, PyObject_Print (op=42, fp=0x7ffff7535780, flags=1) at Objects/object.c:330\n330\t\treturn internal_print(op, fp, flags, 0);\n#7 (unable to read python frame information)\n#3 Frame 0x805700, for file /home/antoine/cpython/__svn__/Lib/test/gdb_sample.py, line 10, in baz (args=(1, 2, 3))\n    print(42)\n' did not match '^.*\n#[0-9]+ Frame 0x[0-9a-f]+, for file .*gdb_sample.py, line 7, in bar \\(a=1, b=2, c=3\\)\n    baz\\(a, b, c\\)\n#[0-9]+ Frame 0x[0-9a-f]+, for file .*gdb_sample.py, line 10, in baz \\(args=\\(1, 2, 3\\)\\)\n    print\\(42\\)\n$'

======================================================================
FAIL: test_basic_command (test.test_gdb.PyBtTests)
Verify that the "py-bt" command works
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/home/antoine/cpython/__svn__/Lib/test/test_gdb.py", line 634, in test_basic_command
    cmds_after_breakpoint=['py-bt'])
  File "/home/antoine/cpython/__svn__/Lib/test/test_gdb.py", line 129, in get_stack_trace
    self.assertEquals(err, '')
AssertionError: 'Error occurred in Python command.\n' != ''

======================================================================
FAIL: test_print_after_up (test.test_gdb.PyPrintTests)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/home/antoine/cpython/__svn__/Lib/test/test_gdb.py", line 658, in test_print_after_up
    r".*\nlocal 'c' = 3\nlocal 'b' = 2\nlocal 'a' = 1\n.*")
  File "/home/antoine/cpython/__svn__/Lib/test/test_gdb.py", line 163, in assertMultilineMatches
    msg='%r did not match %r' % (actual, pattern))
AssertionError: 'Breakpoint 1 at 0x4508a0: file Objects/object.c, line 330.\n\nBreakpoint 1, PyObject_Print (op=42, fp=0x7ffff7535780, flags=1) at Objects/object.c:330\n330\t\treturn internal_print(op, fp, flags, 0);\n#7 (unable to read python frame information)\nUnable to read information on python frame\nUnable to read information on python frame\nUnable to read information on python frame\n' did not match ".*\\nlocal 'c' = 3\\nlocal 'b' = 2\\nlocal 'a' = 1\\n.*"

======================================================================
FAIL: test_locals_after_up (test.test_gdb.PyLocalsTests)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/home/antoine/cpython/__svn__/Lib/test/test_gdb.py", line 684, in test_locals_after_up
    r".*\na = 1\nb = 2\nc = 3\n.*")
  File "/home/antoine/cpython/__svn__/Lib/test/test_gdb.py", line 163, in assertMultilineMatches
    msg='%r did not match %r' % (actual, pattern))
AssertionError: 'Breakpoint 1 at 0x4508a0: file Objects/object.c, line 330.\n\nBreakpoint 1, PyObject_Print (op=42, fp=0x7ffff7535780, flags=1) at Objects/object.c:330\n330\t\treturn internal_print(op, fp, flags, 0);\n#7 (unable to read python frame information)\nUnable to read information on python frame\n' did not match '.*\\na = 1\\nb = 2\\nc = 3\\n.*'

----------------------------------------------------------------------


Basically, it seems the Python API for gdb is a bit unstable or changing. Here is what manually trying the new commands produces:

(gdb) py-bt 
#4 (unable to read python frame information)
Traceback (most recent call last):
  File "/home/antoine/cpython/__svn__/python-gdb.py", line 1341, in invoke
    frame.print_summary()
  File "/home/antoine/cpython/__svn__/python-gdb.py", line 1195, in print_summary
    pyop = self.get_pyop()
  File "/home/antoine/cpython/__svn__/python-gdb.py", line 1168, in get_pyop
    return PyFrameObjectPtr.from_pyobject_ptr(f)
  File "/home/antoine/cpython/__svn__/python-gdb.py", line 349, in from_pyobject_ptr
    return cls(gdbval)
TypeError: __init__() takes exactly 3 arguments (2 given)
Error occurred in Python command.

(gdb) py-up 
Traceback (most recent call last):
  File "/home/antoine/cpython/__svn__/python-gdb.py", line 1309, in invoke
    move_in_stack(move_up=True)
  File "/home/antoine/cpython/__svn__/python-gdb.py", line 1289, in move_in_stack
    iter_frame.print_summary()
  File "/home/antoine/cpython/__svn__/python-gdb.py", line 1195, in print_summary
    pyop = self.get_pyop()
  File "/home/antoine/cpython/__svn__/python-gdb.py", line 1168, in get_pyop
    return PyFrameObjectPtr.from_pyobject_ptr(f)
  File "/home/antoine/cpython/__svn__/python-gdb.py", line 349, in from_pyobject_ptr
    return cls(gdbval)
TypeError: __init__() takes exactly 3 arguments (2 given)
Error occurred in Python command.


$ gdb --version
GNU gdb (GDB) 7.1-1mdv2010.1 (Mandriva Linux release 2010.1)
Copyright (C) 2010 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law.  Type "show copying"
and "show warranty" for details.
This GDB was configured as "x86_64-mandriva-linux-gnu".
For bug reporting instructions, please see:
<http://www.gnu.org/software/gdb/bugs/>.

See #8482 for the "Unable to read information on python frame" issue.

My intuition about an API problem seems to be mistaken. If I compile Python in debug mode without any optimizations ("CFLAGS='' ./configure --with-pydebug"), all tests pass. The default builds of Python use -O2 nowadays, even in debug mode (issue8625), which makes some of the tests fail.

Perhaps test_gdb, or some of its tests, should be skipped when compiler optimizations were enabled. Now we have to find out how to get that information.

Here is a patch which skips test_gdb when Python was built with something else than -O0.

Thanks.  The patch looks good to me, and appears to also fix issue 8482 and issue 9163: compiler optimization across all different compilers and configurations can somewhat arbitrarily break the ability for the debugger to work, and skipping the test in the face of optimization seems the sanest course of action.

Minor nit: given that final_opt is set to "" at the start of the loop, the first conditional in:
    if final_opt and final_opt != '-O0':
appears to be redundant.

Setting "Stage" to "commit review" - I'd commit it (modulo the above nit), but I don't have rights to do so; looking for a committer to review this.

> Minor nit: given that final_opt is set to "" at the start of the loop, > the first conditional in:
>     if final_opt and final_opt != '-O0':
> appears to be redundant.

I don't think it is (the empty string is false, but "" != "-O0" is true).

Oops, my bad.  Patch looks good as is.

Patch committed in 82647 (3.2) and 82648 (2.7). Thanks!