Mailman 3 August 2007 - Python-Dev

Avoiding cascading test failures
by Alexandre Vassalotti 02 Sep '07

02 Sep '07

When I was fixing tests failing in the py3k branch, I found the number duplicate failures annoying. Often, a single bug, in an important method or function, caused a large number of testcase to fail. So, I thought of a simple mechanism for avoiding such cascading failures. My solution is to add a notion of dependency to testcases. A typical usage would look like this: @depends('test_getvalue') def test_writelines(self): ... memio.writelines([buf] * 100) self.assertEqual(memio.getvalue(), buf * 100) ... Here, running the test is pointless if test_getvalue fails. So by making test_writelines depends on the success of test_getvalue, we can ensure that the report won't be polluted with unnecessary failures. Also, I believe this feature will lead to more orthogonal tests, since it encourages the user to write smaller test with less dependencies. I wrote an example implementation (included below) as a proof of concept. If the idea get enough support, I will implement it and add it to the unittest module. -- Alexandre class CycleError(Exception): pass class TestCase: def __init__(self): self.graph = {} tests = [x for x in dir(self) if x.startswith('test')] for testname in tests: test = getattr(self, testname) if hasattr(test, 'deps'): self.graph[testname] = test.deps else: self.graph[testname] = set() def run(self): graph = self.graph toskip = set() msgs = [] while graph: # find tests without any pending dependencies source = [test for test, deps in graph.items() if not deps] if not source: raise CycleError for testname in source: if testname in toskip: msgs.append("%s... skipped" % testname) resolvedeps(graph, testname) del graph[testname] continue test = getattr(self, testname) try: test() except AssertionError: toskip.update(getrevdeps(graph, testname)) msgs.append("%s... failed" % testname) except: toskip.update(getrevdeps(graph, testname)) msgs.append("%s... error" % testname) else: msgs.append("%s... ok" % testname) finally: resolvedeps(graph, testname) del graph[testname] for msg in sorted(msgs): print(msg) def getrevdeps(graph, testname): """Return the reverse depencencies of a test""" rdeps = set() for x in graph: if testname in graph[x]: rdeps.add(x) if rdeps: # propagate depencencies recursively for x in rdeps.copy(): rdeps.update(getrevdeps(graph, x)) return rdeps def resolvedeps(graph, testname): for test in graph: if testname in graph[test]: graph[test].remove(testname) def depends(*args): def decorator(test): if hasattr(test, 'deps'): test.deps.update(args) else: test.deps = set(args) return test return decorator class MyTest(TestCase): @depends('test_foo') def test_nah(self): pass @depends('test_bar', 'test_baz') def test_foo(self): pass @depends('test_tin') def test_bar(self): self.fail() def test_baz(self): self.error() def test_tin(self): pass def error(self): raise ValueError def fail(self): raise AssertionError if __name__ == '__main__': t = MyTest() t.run()

3 4

Re: [Python-Dev] the FakeSocket class in httplib.py -- remove it?
by Brett Cannon 31 Aug '07

31 Aug '07

On 8/31/07, Bill Janssen <janssen(a)parc.com> wrote: > I'm wondering if it would be OK to remove the FakeSocket class in > httplib? Actually, it's currently a function that returns one of its > arguments. It's not documented as part of the module, and not used > elsewhere in the rest of the library. Since I assume that includes unit tests, I have no issue seeing it go away, especially if it is not being kept up-to-date. But I am willing to bet someone out there is using it so it might be best to add a DeprecationWarning for 2.6 and then remove in 3.0/2.7. -Brett

2 1

the FakeSocket class in httplib.py -- remove it?
by Bill Janssen 31 Aug '07

31 Aug '07

I'm wondering if it would be OK to remove the FakeSocket class in httplib? Actually, it's currently a function that returns one of its arguments. It's not documented as part of the module, and not used elsewhere in the rest of the library. Bill

1 0

Python 3.0a1 released!
by Guido van Rossum 31 Aug '07

31 Aug '07

The release is available from http://python.org/download/releases/3.0/ I'll send a longer announcement to python-list and python-announce-list. Please blog about this if you have a blog! Thanks to all who helped out! It's been a great ride. -- --Guido van Rossum (home page: http://www.python.org/~guido/)

1 0

Order of operations (was: PEP 238 - The // operator should truncate instead of floor)
by Facundo Batista 31 Aug '07

31 Aug '07

2007/8/29, "Martin v. Löwis" <martin(a)v.loewis.de>: > It would have been good if you had indicated what result you had > expected instead. I assume 0; to get 0, you have to write -(3/2)+3/2, > or 0-3/2+3/2. Wow, that caught me: >>> -3/2+3/2 -1 >>> 0-3/2+3/2 0 >>> I'm not talking about division here, just the fact that in the first case, it was (-3) / 2 and not -(3/2). It's a surprise to me, taking into account that >>> -3**2 -9 >>> 0-3**2 -9 Here, the behavior is always the same: -(3**2). Do you know why? Thanks! -- . Facundo Blog: http://www.taniquetil.com.ar/plog/ PyAr: http://www.python.org/ar/

8 9

Notification of assignment in new tracker
by "Martin v. Löwis" 30 Aug '07

30 Aug '07

Erik just fixed the tracker so that the Assigned-to person gets email for any message added to an issue. However, *no* email is sent if one merely changes the Assigned-to field. Sending a message for every status change would cause significant amounts of spam, so mere status changes are performed "silently". So if you assign an issue to somebody but yourself, you should leave a message also so that this person actually learns about the assignment. Doing so is probably a good idea, anyway, as you should explain why you did this assignment. Regards, Martin

1 0

HAVE_DECL_TZNAME in pyconfig.h.in, but no test for it ?
by Alexander Neundorf 30 Aug '07

30 Aug '07

Hi, in current python svn there is a #undef HAVE_DECL_TZNAME but I can't find a configure test for it. Did I miss something ? Bye Alex

2 1

Re: [Python-Dev] Order of operations
by Scott Dial 30 Aug '07

30 Aug '07

Martin v. Löwis wrote: > Alexandre Vassalotti schrieb: >> C doesn't have an exponentiation operator. You use the pow() function, instead: > > Right. Then Scott's claim ("Because it is consistent with C's precedence > rules.") can't be true - from C precedence, it does not follow that ** > must have higher precedence than unary -. > I believe the confusion with my claim is that I thought the oddity in question was that -3/2 != 0-3/2, but I believe you took the other half about exponentiation as the oddity in question that -3**2 == -(3**2). My original comment was to say that unary - was higher precedence than division because of C's ordering. With respect to the original question, then my answer was poor as I ignored the exponentiation part. Since we tacked this on, my answer was to be an implicit zen answer of "because it makes the most sense." Sorry for the added noise. -Scott -- Scott Dial scott(a)scottdial.com scodial(a)cs.indiana.edu

1 0

Re: [Python-Dev] Other SSL issues in the tracker have been marked
by Guido van Rossum 30 Aug '07

30 Aug '07

On 8/28/07, Bill Janssen <janssen(a)parc.com> wrote: > > The problematic test is this one: > > > > File "C:\buildbot_py25\trunk.mcintyre-windows\build\lib\test\test_ssl.py", line 77, in testSSLconnect > > s.connect(("pop.gmail.com", 995)) > > File "C:\buildbot_py25\trunk.mcintyre-windows\build\lib\ssl.py", line 170, in connect > > if self._sslobj or (self.getsockname()[1] != 0): > > File "<string>", line 1, in getsockname > > error: (10022, 'Invalid argument') > > Actually, I think this is a bug in the Windows implementation of the > socket interface. The question is, what should s.getsockname() return > (or throw) if called on an unbound socket instance? Unix decided one > way, and Windows another (and not a particularly good way). I'll post > an issue to the tracker. Well, that's just what you get with two independent implementations of a spec. We don't try to hide the differences between the sockets stack in Unix and Windows -- you'll just have to work around it. -- --Guido van Rossum (home page: http://www.python.org/~guido/)

3 6

newb question
by Ty Newton 29 Aug '07

29 Aug '07

Hi, I'm new to the Python source and I have a (hopefully) quick question that someone more familiar with it should be able to answer so I can continue on my way to understanding how Python is put together. I have started looking at the parser and have gotten a little confused about how the grammar is instantiated (perhaps the wrong term). Here's how I understand things at the moment. graminit.c contains the definition of the grammar in terms of static structures. Everything seems simple enough until I get to the DFA (in graminit.c): static dfa dfas[84] = { {256, "single_input", 0, 3, states_0, "\004\050\014\000\000\000\000\025\074\005\023\310\011\020\004\000\300\020\222\006\201"}, ... I assume that the last bit of my snippet (\004\050\014\ ...) is a bitset structure. When I look at bitset it is defined as a char array. Can someone explain how this works please? I've never come across escape sequences like this; I've only ever seen \0 (nul) before; not \2, \3 etc. or are they not escape sequences, but literal forward slashes. Thanks, Ty

2 1