diff options
author | Barry Warsaw <barry@python.org> | 2011-12-15 16:50:02 -0500 |
---|---|---|
committer | Barry Warsaw <barry@python.org> | 2011-12-15 16:50:02 -0500 |
commit | f2909c23abc4f8fa55d71673785f8e70a843f6ce (patch) | |
tree | 97adb74ab2ae9f262cd6ad11e049e50c86f1f178 | |
parent | 4c1c2eade1c5b383adad94a7a4fd6553873fecf0 (diff) | |
download | dbus-python-f2909c23abc4f8fa55d71673785f8e70a843f6ce.tar.gz |
- Added back the missing PY3PORT.rst file, with updates.
- Disallow appending unicode objects with 'y' (bytes) signatures. This now
requires either a bytes object or an integer. Update the tests to reflect
- this change.
- Fix broken __all__ in Python 3.
-rw-r--r-- | PY3PORT.rst | 227 | ||||
-rw-r--r-- | _dbus_bindings/message-append.c | 26 | ||||
-rw-r--r-- | dbus/types.py | 5 | ||||
-rw-r--r-- | test/cross-test-client.py | 9 | ||||
-rwxr-xr-x | test/run-test.sh | 1 | ||||
-rwxr-xr-x | test/test-standalone.py | 10 |
6 files changed, 249 insertions, 29 deletions
diff --git a/PY3PORT.rst b/PY3PORT.rst new file mode 100644 index 0000000..c873c2c --- /dev/null +++ b/PY3PORT.rst @@ -0,0 +1,227 @@ +=============================== +Porting python-dbus to Python 3 +=============================== + +This is an experimental port to Python 3.x where x >= 2. There are lots of +great sources for porting C extensions to Python 3, including: + + * http://python3porting.com/toc.html + * http://docs.python.org/howto/cporting.html + * http://docs.python.org/py3k/c-api/index.html + +I also consulted an early take on this port by John Palmieri and David Malcolm +in the context of Fedora: + + * https://bugs.freedesktop.org/show_bug.cgi?id=26420 + +although I have made some different choices. The patches in that tracker +issue also don't cover porting the Python bits (e.g. the test suite), nor the +pygtk -> pygi porting, both which I've also attempted to do in this branch. + +This document outlines my notes and strategies for doing this port. Please +feel free to contact me with any bugs, issues, disagreements, suggestions, +kudos, and curses. + +Barry Warsaw +barry@python.org +2011-11-11 + + +User visible changes +==================== + +You've got some dbus-python code that works great in Python 2. This branch +should generally allow your existing Python 2 code to continue to work +unchanged. There are a few changes you'll notice in Python 2 though:: + + - The minimum supported Python 2 version is 2.6. + - All object reprs are unicodes. This change was made because it greatly + simplifies the implementation and cross-compatibility with Python 3. + - Some values which were ints are now longs. Primarily, this affects the + type of the `variant_level` attributes. + - `dbus.Byte` can be constructed from a 1-character str or unicode object. + - Some exception strings have changed. + - `MethodCallMessage` and `SignalMessage` objects have better reprs now. + +What do you need to do to port that to Python 3? Here are the user visible +changes you should be aware of. Python 3.2 is the minimal required version:: + + - `ByteArray`s must be initialized with bytes objects, not unicodes. Use + `b''` literals in the constructor. This also works in Python 2, where + bytes objects are aliases for 8-bit strings. + - byte signatures (i.e. `y` type codes) must be passed either a length-1 + bytes object or an integer. unicodes are not allowed. + - `ByteArray` is now a subclass of `bytes`, where in Python 2 it is a + subclass of `str`. + - `dbus.Byte` can be constructed from a 1-character byte or str object, or an + integer. + - `dbus.UTF8String` is gone, use `dbus.String`. Also `utf8_string` arguments + are no longer allowed. + - All longs are now ints, since Python 3 has only a single int type. This + also means that the class hierarchy for the dbus numeric types has changed + (all derive from int in Python 3). + - Some exception strings have changed. + - `MethodCallMessage` and `SignalMessage` objects have better reprs now. + + +Bytes vs. Strings +================= + +All strings in dbus are defined as UTF-8: + +http://dbus.freedesktop.org/doc/dbus-specification.html#message-protocol-signatures + +However, the dbus C API accepts `char*` which must be UTF-8 strings NUL +terminated and no other NUL bytes. + +This page describes the mapping between Python types and dbus types: + + http://dbus.freedesktop.org/doc/dbus-python/doc/tutorial.html#basic-types + +Notice that it maps dbus `string` (`'s'`) to `dbus.String` (unicode) or +`dbus.UTF8String` (str). Also notice that there is no direct dbus equivalent +of Python's bytes type (although dbus does have byte arrays), so I am mapping +dbus strings to unicodes in all cases, and getting rid of `dbus.UTF8String` in +Python 3. I've also added a `dbus._BytesBase` type which is unused in Python +2, but which forms the base class for `dbus.ByteArray` in Python 3. This is +an implementation detail and not part of the public API. + +In Python 3, object paths (`'o'` or `dbus.ObjectPath`), signatures (`'g'` or +`dbus.Signature`), bus names, interfaces, and methods are all strings. A +previous aborted effort was made to use bytes for these, which at first blush +may makes some sense, but on deeper consideration does not. This approach +also tended to impose too many changes on user code, and caused lots of +difficult to track down problems. + +In Python 3, all such objects are subclasses of `str` (i.e. `unicode`). + +(As an example, dbus-python's callback dispatching pretty much assumes all +these things are strings. When they are bytes, the fact that `'foo' != b'foo'` +causes dispatch matching to fail in difficult to debug ways. Even bus names +are not immune, since they do things like `bus_name[:1] == ':'` which fails in +multiple ways when `bus_name` is a bytes. For sanity purposes, these are all +unicode strings now, and we just eat the complexity at the C level.) + +I am using `#include <bytesobject.h>`, which exposes the PyBytes API to Python +2.6 and 2.7, and I have converted all internal PyString calls to PyBytes +calls. Where this is inappropriate, we'll use PyUnicode calls explicitly. +E.g. all repr() implementations now return unicodes. Most of these changes +shouldn't be noticed, even in existing Python 2 code. + +Generally, I've left the descriptions and docstrings saying "str" instead of +"unicode" since there's no distinction in Python 3. + +APIs which previously returned PyStrings will usually return PyUnicodes, not +PyBytes. + + +Ints vs. Longs +============== + +Python 3 only has PyLong types; PyInts are gone. For that reason, I've +switched all PyInt calls to use PyLong in both Python 2 and Python 3. Python +3.0 had a nice `<intobject.h>` header that aliased PyInt to PyLong, but that's +gone as of Python 3.1, and the minimal required Python 3 version is 3.2. + +In the above page mapping basic types, you'll notice that the Python int type +is mapped to 32-bit signed integers ('i') and the Python long type is mapped +to 64-bit signed integers ('x'). Python 3 doesn't have this distinction, so +ints map to 'i' even though ints can be larger in Python 3. Use the +dbus-specific integer types if you must have more exact mappings. + +APIs which accepted ints in Python 2 will still do so, but they'll also now +accept longs. These APIs obviously only accept longs in Python 3. + +Long literals in Python code are an interesting thing to have to port. Don't +use them if you want your code to work in both Python versions. + +`dbus._IntBase` is removed in Python 3, you only have `dbus._LongBase`, which +inherits from a Python 3 int (i.e. a PyLong). Again, this is an +implementation detail that users should never care about. + + +Macros +====== + +In types-internal.h, I define `PY3K` when `PY_MAJOR_VERSION` >= 3, so you'll +see ifdefs on the former symbol within the C code. + +Python 3 really could use a PY_REFCNT() wrapper for ob_refcnt access. + + +PyCapsule vs. PyCObject +======================= + +`_dbus_bindings._C_API` is an attribute exposed to Python in the module. In +Python 2, this is a PyCObject, but these do not exist in Python >= 3.2, so it +is replaced with a PyCapsules for Python 3. However, since PyCapsules were +only introduced in Python 2.7, and I want to support Python 2.6, PyCObjects +are still used when this module is compiled for Python 2. + + +Python level compatibility +========================== + +`from dbus import _is_py3` gives you a flag to check if you must do something +different in Python 3. In general I use this flag to support both versions in +one set of sources, which seems better than trying to use 2to3. It's not part +of the dbus-python public API, so you may not need it. + + +Miscellaneous +============= + +The PyDoc_STRVAR() documentation is probably out of date. Once the API +choices have been green-lighted upstream, I'll make a pass through the code to +update them. It might be tricky based on any differences between Python 2 and +Python 3. + +There were a few places where I noticed what might be considered bugs, +unchecked exception conditions, or possible reference count leaks. In these +cases, I've just fixed what I can and hopefully haven't made the situation +worse. + +`dbus_py_variant_level_get()` did not check possible error conditions, nor did +their callers. When `dbus_py_variant_level_get()` encounters an error, it now +returns -1, and callers check this. + +As much as possible, I've refrained from general code cleanups (e.g. 80 +columns), unless it just bugged me too much or I touched the code for reasons +related to the port. I've also tried to stick to existing C code style, +e.g. through the use of pervasive `Py_CLEAR()` calls, comparison against NULL +usually with `!foo`, and such. As Bart Simpson might write on his classroom +blackboard:: + + This is not a rewrite + This is not a rewrite + This is not a rewrite + This is not a rewrite + ... + +and so on. Well, mostly ;). + +I think I fixed a reference leak in `DBusPyServer_set_auth_mechanisms()`. +`PySequence_Fast()` returns a new reference, which wasn't getting decref'd in +any return path. + + - Instantiation of metaclasses uses different, incompatible syntax in Python + 2 and 3. You have to use direct calling of the metaclass to work across + versions, i.e. `Interface = InterfaceType('Interface', (object,), {})` + - `iteritems()` and friends are gone. I dropped the "iter" prefixes. + - `xrange() is gone. I changed them to use `range()`. + - `isSequenceType()` is gone in Python 3, so I use a different idiom there. + - `__next__()` vs. `next()` + - `PyUnicode_FromFormat()` `%V` flag is a clever hack! + - `sys.version_info` is a tuple in Python 2.6, not a namedtuple. i.e. there + is no `sys.version_info.major` + - `PyArg_Parse()`: No 'y' code in Python 2; in Python 3, no equivalent of 'z' + for bytes objects. + + +Open issues +=========== + +Here are a few things that still need to be done, or for which there may be +open questions:: + + - Update all C extension docstrings for accuracy. diff --git a/_dbus_bindings/message-append.c b/_dbus_bindings/message-append.c index c08f498..bbf1286 100644 --- a/_dbus_bindings/message-append.c +++ b/_dbus_bindings/message-append.c @@ -587,35 +587,21 @@ _message_iter_append_byte(DBusMessageIter *appender, PyObject *obj) if (PyBytes_Check(obj)) { if (PyBytes_GET_SIZE(obj) != 1) { - PyErr_Format(PyExc_ValueError, "Expected a string of " - "length 1 byte, but found %d bytes", - (int) PyBytes_GET_SIZE(obj)); + PyErr_Format(PyExc_ValueError, + "Expected a length-1 bytes but found %d bytes", + (int)PyBytes_GET_SIZE(obj)); return -1; } y = *(unsigned char *)PyBytes_AS_STRING(obj); } - else if (PyUnicode_Check(obj)) { - PyObject *obj_as_bytes = PyUnicode_AsUTF8String(obj); - - if (!obj_as_bytes) - return -1; - if (PyBytes_GET_SIZE(obj_as_bytes) != 1) { - PyErr_Format(PyExc_ValueError, "Expected a string of " - "length 1 byte, but found %d bytes", - (int)PyBytes_GET_SIZE(obj_as_bytes)); - Py_CLEAR(obj_as_bytes); - return -1; - } - y = *(unsigned char *)PyBytes_AS_STRING(obj_as_bytes); - Py_CLEAR(obj_as_bytes); - } else { long i = PyLong_AsLong(obj); if (i == -1 && PyErr_Occurred()) return -1; if (i < 0 || i > 0xff) { - PyErr_Format(PyExc_ValueError, "%d outside range for a " - "byte value", (int)i); + PyErr_Format(PyExc_ValueError, + "%d outside range for a byte value", + (int)i); return -1; } y = i; diff --git a/dbus/types.py b/dbus/types.py index a134495..c33acff 100644 --- a/dbus/types.py +++ b/dbus/types.py @@ -1,7 +1,7 @@ -__all__ = ('ObjectPath', 'ByteArray', 'Signature', 'Byte', 'Boolean', +__all__ = ['ObjectPath', 'ByteArray', 'Signature', 'Byte', 'Boolean', 'Int16', 'UInt16', 'Int32', 'UInt32', 'Int64', 'UInt64', 'Double', 'String', 'Array', 'Struct', 'Dictionary', - 'UTF8String', 'UnixFd') + 'UnixFd'] from _dbus_bindings import ( Array, Boolean, Byte, ByteArray, Dictionary, Double, Int16, Int32, Int64, @@ -11,3 +11,4 @@ from _dbus_bindings import ( from dbus._compat import is_py2 if is_py2: from _dbus_bindings import UTF8String + __all__.append('UTF8String') diff --git a/test/cross-test-client.py b/test/cross-test-client.py index ecc6f8a..74d957c 100644 --- a/test/cross-test-client.py +++ b/test/cross-test-client.py @@ -234,7 +234,8 @@ class Client(SignalTestsImpl): # "Single tests" if have_signatures: self.assert_method_eq(INTERFACE_SINGLE_TESTS, 6, 'Sum', [1, 2, 3]) - self.assert_method_eq(INTERFACE_SINGLE_TESTS, 6, 'Sum', ['\x01', '\x02', '\x03']) + self.assert_method_eq(INTERFACE_SINGLE_TESTS, 6, 'Sum', + [b'\x01', b'\x02', b'\x03']) self.assert_method_eq(INTERFACE_SINGLE_TESTS, 6, 'Sum', [Byte(1), Byte(2), Byte(3)]) self.assert_method_eq(INTERFACE_SINGLE_TESTS, 6, 'Sum', ByteArray(b'\x01\x02\x03')) @@ -282,7 +283,7 @@ class Client(SignalTestsImpl): self.assert_method_eq(INTERFACE_TESTS, '\xa9', 'IdentityString', i) if have_signatures: - self.assert_method_eq(INTERFACE_TESTS, Byte(0x42), 'IdentityByte', '\x42') + #self.assert_method_eq(INTERFACE_TESTS, Byte(0x42), 'IdentityByte', '\x42') self.assert_method_eq(INTERFACE_TESTS, True, 'IdentityBool', 42) self.assert_method_eq(INTERFACE_TESTS, 42, 'IdentityInt16', 42) self.assert_method_eq(INTERFACE_TESTS, 42, 'IdentityUInt16', 42) @@ -342,7 +343,9 @@ class Client(SignalTestsImpl): 'IdentityByteArray', ByteArray(b'\x01\x02\x03')) if have_signatures: - self.assert_method_eq(INTERFACE_TESTS, [1,2,3], 'IdentityByteArray', ['\x01', '\x02', '\x03']) + self.assert_method_eq(INTERFACE_TESTS, [1,2,3], + 'IdentityByteArray', + [b'\x01', b'\x02', b'\x03']) self.assert_method_eq(INTERFACE_TESTS, [False,True], 'IdentityBoolArray', [False,True]) if have_signatures: self.assert_method_eq(INTERFACE_TESTS, [False,True,True], 'IdentityBoolArray', [0,1,2]) diff --git a/test/run-test.sh b/test/run-test.sh index 516e876..2530b3f 100755 --- a/test/run-test.sh +++ b/test/run-test.sh @@ -63,6 +63,7 @@ fi dbus-monitor > "$DBUS_TOP_BUILDDIR"/test/monitor.log & echo "PYTHONPATH=$PYTHONPATH" +echo "PYTHON=$PYTHON" echo "running test-standalone.py" $PYTHON "$DBUS_TOP_SRCDIR"/test/test-standalone.py || die "test-standalone.py failed" diff --git a/test/test-standalone.py b/test/test-standalone.py index a76ad25..48cbe09 100755 --- a/test/test-standalone.py +++ b/test/test-standalone.py @@ -265,10 +265,12 @@ class TestMessageMarshalling(unittest.TestCase): def test_get_args_options(self): aeq = self.assertEqual s = _dbus_bindings.SignalMessage('/', 'foo.bar', 'baz') - s.append('b', 'bytes', -1, 1, 'str', 'var', signature='yayiusv') - aeq(s.get_args_list(), [ord('b'), - [ord('b'),ord('y'),ord('t'),ord('e'), ord('s')], - -1, 1, 'str', 'var']) + s.append(b'b', b'bytes', -1, 1, 'str', 'var', signature='yayiusv') + aeq(s.get_args_list(), [ + ord('b'), + [ord('b'),ord('y'),ord('t'),ord('e'), ord('s')], + -1, 1, 'str', 'var' + ]) byte, bytes, int32, uint32, string, variant = s.get_args_list() aeq(byte.__class__, types.Byte) aeq(bytes.__class__, types.Array) |