I've applied a modified version of Greg Chapman's patch. I've included
the fixes without introducing the reorganization mentioned, for the sake
of stability. Also, the second fix mentioned in the patch don't fix the
mentioned problem anymore, because of the change introduced by patch
#720991 (by Greg as well). The new fix wasn't complicated though, and is
included as well.
As a note. It seems that there are other places that require the
"protection" of LASTMARK_SAVE()/LASTMARK_RESTORE(), and are just waiting
for someone to find how to break them. Particularly, I belive that every
recursion of SRE_MATCH() should be protected by these macros. I won't
do that right now since I'm not completely sure about this, and we don't
have much time for testing until the next release.
raising an exception. This is consistent with calling the
constructors for the other builtin types -- called without argument
they all return the false value of that type. (SF patch #724135)
Thanks to Alex Martelli.
New functions:
unsigned long PyInt_AsUnsignedLongMask(PyObject *);
unsigned PY_LONG_LONG) PyInt_AsUnsignedLongLongMask(PyObject *);
unsigned long PyLong_AsUnsignedLongMask(PyObject *);
unsigned PY_LONG_LONG) PyLong_AsUnsignedLongLongMask(PyObject *);
New and changed format codes:
b unsigned char 0..UCHAR_MAX
B unsigned char none **
h unsigned short 0..USHRT_MAX
H unsigned short none **
i int INT_MIN..INT_MAX
I * unsigned int 0..UINT_MAX
l long LONG_MIN..LONG_MAX
k * unsigned long none
L long long LLONG_MIN..LLONG_MAX
K * unsigned long long none
Notes:
* New format codes.
** Changed from previous "range-and-a-half" to "none"; the
range-and-a-half checking wasn't particularly useful.
New test test_getargs2.py, to verify all this.
getpwnam()/getpwuid() return consistent data.
Change test_grp to check that getgrall() and
getgrnam()/getgrgid() return consistent data.
Add error checks similar to test_pwd.py.
Port test___all__.py to PyUnit.
From SF patch #662807.
interpreted by slicing, so negative values count from the end of the
list. This was the only place where such an interpretation was not
placed on a list index.
A small fix for bug #545855 and Greg Chapman's
addition of op code SRE_OP_MIN_REPEAT_ONE for
eliminating recursion on simple uses of pattern '*?' on a
long string.
- range() now works even if the arguments are longs with magnitude
larger than sys.maxint, as long as the total length of the sequence
fits. E.g., range(2**100, 2**101, 2**100) is the following list:
[1267650600228229401496703205376L]. (SF patch #707427.)
- Expose NullTranslations and GNUTranslations to __all__
- Set the default charset to iso-8859-1. It used to be None, which
would cause problems with .ugettext() if the file had no charset
parameter. Arguably, the po/mo file would be broken, but I still think
iso-8859-1 is a reasonable default.
- Add a "coerce" default argument to GNUTranslations's constructor. The
reason for this is that in Zope, we want all msgids and msgstrs to be
Unicode. For the latter, we could use .ugettext() but there isn't
currently a mechanism for Unicode-ifying msgids.
The plan then is that the charset parameter specifies the encoding for
both the msgids and msgstrs, and both are decoded to Unicode when read.
For example, we might encode po files with utf-8. I think the GNU
gettext tools don't care.
Since this could potentially break code [*] that wants to use the
encoded interface .gettext(), the constructor flag is added, defaulting
to False. Most code I suspect will want to set this to True and use
.ugettext().
- A few other minor changes from the Zope project, including asserting
that a zero-length msgid must have a Project-ID-Version header for it to
be counted as the metadata record.
* Doc - add doc for when functions were added
* UserString
* string object methods
* string module functions
'chars' is used for the last parameter everywhere.
These changes will be backported, since part of the changes
have already been made, but they were inconsistent.
These never failed in 2.3, and the tests confirm it. They still blow up
in the 2.2 branch, despite that all the gc-vs-__del__ fixes from 2.3
have been backported (and this is expected -- 2.2 needs more work than
2.3 needed).
of PyObject_HasAttr(); the former promises never to execute
arbitrary Python code. Undid many of the changes recently made to
worm around the worst consequences of that PyObject_HasAttr() could
execute arbitrary Python code.
Compatibility is hard to discuss, because the dangerous cases are
so perverse, and much of this appears to rely on implementation
accidents.
To start with, using hasattr() to check for __del__ wasn't only
dangerous, in some cases it was wrong: if an instance of an old-
style class didn't have "__del__" in its instance dict or in any
base class dict, but a getattr hook said __del__ existed, then
hasattr() said "yes, this object has a __del__". But
instance_dealloc() ignores the possibility of getattr hooks when
looking for a __del__, so while object.__del__ succeeds, no
__del__ method is called when the object is deleted. gc was
therefore incorrect in believing that the object had a finalizer.
The new method doesn't suffer that problem (like instance_dealloc(),
_PyObject_Lookup() doesn't believe __del__ exists in that case), but
does suffer a somewhat opposite-- and even more obscure --oddity:
if an instance of an old-style class doesn't have "__del__" in its
instance dict, and a base class does have "__del__" in its dict,
and the first base class with a "__del__" associates it with a
descriptor (an object with a __get__ method), *and* if that
descriptor raises an exception when __get__ is called, then
(a) the current method believes the instance does have a __del__,
but (b) hasattr() does not believe the instance has a __del__.
While these disagree, I believe the new method is "more correct":
because the descriptor *will* be called when the object is
destructed, it can execute arbitrary Python code at the time the
object is destructed, and that's really what gc means by "has a
finalizer": not specifically a __del__ method, but more generally
the possibility of executing arbitrary Python code at object
destruction time. Code in a descriptor's __get__() executed at
destruction time can be just as problematic as code in a
__del__() executed then.
So I believe the new method is better on all counts.
Bugfix candidate, but it's unclear to me how all this differs in
the 2.2 branch (e.g., new-style and old-style classes already
took different gc paths in 2.3 before this last round of patches,
but don't in the 2.2 branch).
externally unreachable objects with finalizers, and externally unreachable
objects without finalizers reachable from such objects. This allows us
to call has_finalizer() at most once per object, and so limit the pain of
nasty getattr hooks. This fixes the failing "boom 2" example Jeremy
posted (a non-printing variant of which is now part of test_gc), via never
triggering the nasty part of its __getattr__ method.
platforms which have dup(2). The makefile() method is built directly on top
of the socket without duplicating the file descriptor, allowing timeouts to
work properly. Includes a new test case (urllibnet) which requires the
network resource.
Closes bug 707074.
pack_float, pack_double, save_float: All the routines for creating
IEEE-format packed representations of floats and doubles simply ignored
that rounding can (in rare cases) propagate out of a long string of
1 bits. At worst, the end-off carry can (by mistake) interfere with
the exponent value, and then unpacking yields a result wrong by a factor
of 2. In less severe cases, it can end up losing more low-order bits
than intended, or fail to catch overflow *caused* by rounding.
Bugfix candidate, but I already backported this to 2.2.
In 2.3, this code remains in severe need of refactoring.
invalid, rather than returning a string of random garbage of the
estimated result length. Closes SF patch #703471 by Hye-Shik Chang.
Will backport to 2.2-maint (consider it done.)
ALERT! A month ago or so I made test_ossaudiodev.py require the
'audio' resource, but I didn't make the necessary changes to
regrtest.py. This means that *nobody* has been testing the oss module
all that time!
Contributed by Brett Cannon.
To prevent code duplication, I patched _strptime to use datetime's date
object to do Julian day, Gregorian, and day of the week calculations.
Patch also includes new regression tests to test results and the
calculation gets triggered.
Very minor comment changes and the contact email are also changed.
* Adds missing pop() methods to weakref.py
* Expands test suite to broaden coverage of objects with
a mapping interface.
Contributed by Sebastien Keim.
Quoting the path doesn't work on Win2K (cmd.exe) regardless, this is just
a hack to let the test pass again on Win2K (so long as Python isn't
installed in a path that does contain an embedded space). On Win2K it
looks like we'd also have to add a second pair of double quotes, around
the entire command line.
- the test was sloppy about filenames: "0-REGTYPE-TEXT" was used where
the archive held "/0-REGTYPE-TEXT".
- tarfile extracts all files in binary mode, but the test expected to be able to
read and compare text files in text mode. Use universal text mode.
[ 555817 ] Flawed fcntl.ioctl implementation.
with my patch that allows for an array to be mutated when passed
as the buffer argument to ioctl() (details complicated by
backwards compatibility considerations -- read the docs!).
assertRaises. Fixed a repeated subtle bug in the inplace tests by
removing the possibilty that a self.fail() call could raise a
TypeError that the test catches by mistake.
Allow mixed-type __eq__ and __ne__ for Set objects. This is messier than
I'd like because Set *also* implements __cmp__. I know of one glitch now:
cmp(s, t) returns 0 now when s and t are both Sets and s == t, despite
that Set.__cmp__ unconditionally raises TypeError (and by intent). The
rub is that __eq__ gets tried first, and the x.__eq__(y) True result
convinces Python that cmp(x, y) is 0 without even calling Set.__cmp__.
rarely needed, but can sometimes be useful to release objects
referenced by the traceback held in sys.exc_info()[2]. (SF patch
#693195.) Thanks to Kevin Jacobs!
The problem is in sre_compile.py: the call to
_compile_charset near the end of _compile_info forgets to
pass in the flags, so that the info charset is not compiled
with re.U. (The info charset is used when searching to find
the first character at which a match could start; it is not
generated for patterns beginning with a repeat like '\w{1}'.)
test_nonrecursive_deep(): Reduced nesting depth to 60.
Not a bugfix candidate. 2.3 increased the number of stack frames
needed to pickle a list (in order to get implement the "list
batching" unpickling memory optimization new in 2.3).
between str, unicode, UserString and the string module
as possible. This increases code coverage in stringobject.c
from 83% to 86% and should help keep the string classes
in sync in the future. From SF patch #662807
time.sleep(1) sometimes delays for fractionally less than a second
resulting in too short of an interval for C's time.time() function
to create a distinct seed.
Fix off-by-1 error in normalize_line_endings():
when *p == '\0' the NUL was copied into q and q was auto-incremented,
the loop was broken out of,
then a newline was appended followed by a NUL.
So the function, in effect, was strcpy() but added two extra chars
which was caught by obmalloc in debug mode, since there was only
room for 1 additional newline.
Get test working under regrtest (added test_main).
the optional proto 2 slot state.
pickle.py, load_build(): CAUTION: Noted that cPickle's
load_build and pickle's load_build really don't do the same
things with the state, and didn't before this patch either.
cPickle never tries to do .update(), and has no backoff if
instance.__dict__ can't be retrieved. There are no tests
that can tell the difference, and part of what cPickle's
load_build() did looked accidental to me, so I don't know
what the true intent is here.
pickletester.py, test_pickle.py: Got rid of the hack for
exempting cPickle from running some of the proto 2 tests.
dictobject.c, PyDict_Next(): documented intended use.
test_linuxaudiodev.py) are no longer run by default. This is
because they don't always work, depending on your hardware and
software. To run these tests, you must use an invocation like
./python Lib/test/regrtest.py -u audio test_ossaudiodev
with an indented code block but no newline would raise SyntaxError.
This would have been a four-line change in parsetok.c... Except
codeop.py depends on this behavior, so a compilation flag had to be
invented that causes the tokenizer to revert to the old behavior;
this required extra changes to 2 .h files, 2 .c files, and 2 .py
files. (Fixes SF bug #501622.)
and loading them via the other, except for the special cases of this
Guido added to test_datetime.py for datetime module objects. The new
test_xpickle.py tries all of pickletester's AbstractPickleTests in
both x-module ways.
This changes the default __new__ to refuse arguments iff tp_init is the
default __init__ implementation -- thus making it a TypeError when you
try to pass arguments to a constructor if the class doesn't override at
least __init__ or __new__.
"Unsigned" (i.e., positive-looking, but really negative) hex/oct
constants with a leading minus sign are once again properly negated.
The micro-optimization for negated numeric constants did the wrong
thing for such hex/oct constants. The patch avoids the optimization
for all hex/oct constants.
This needs to be backported to Python 2.2!
descr_check(); it wasn't useful. Change the type argument of the
various _get() methods to PyObject * because the call signature of
tp_descr_get doesn't guarantee its type.
object is not a real str or unicode but an instance
of a subclass, construct the output via looping
over __getitem__. This guarantees that the result
is the same for function==None and function==lambda x:x
This doesn't happen for tuples, because filtertuple()
uses PyTuple_GetItem().
(This was discussed on SF bug #665835).
* Removed the ifilter flag wart by splitting it into two simpler functions.
* Fixed comment tabbing in C code.
* Factored module start-up code into a loop.
Documentation:
* Re-wrote introduction.
* Addede examples for quantifiers.
* Simplified python equivalent for islice().
* Documented split of ifilter().
Sets.py:
* Replace old ifilter() usage with new.
tickle the 2.2.2 __cmp__ bug test_datetime used to tickle, so the
workarounds for that bug no longer make sense in the test suite (which I'm
still trying to keep as closely in synch as possible with Zope3's
version).
__ne__ no longer complain if they don't know how to compare to the other
thing. If no meaningful way to compare is known, saying "not equal" is
sensible. This allows things like
if adatetime in some_sequence:
and
somedict[adatetime] = whatever
to work as expected even if some_sequence contains non-datetime objects,
or somedict non-datetime keys, because they only call __eq__.
It still complains (raises TypeError) for mixed-type comparisons in
contexts that require a total ordering, such as list.sort(), use as a
key in a BTree-based data structure, and cmp().
the tests will remain in sync:
"""
Tres discovered a weird bug when a datetime is pickled, caused by the
shadowing of __year, __month, __day and the use of proxies.
Here's a quick fix and a quick unit test. I don't quite understand
why this wasn't caught by the pickling unit tests.
"""
atomically, but deepcopy() didn't support this at all.
I don't see any reason for this, so I'm adding ClassType
to the set of types that are deep-copied atomically.
* Fixed typo in exception message for times()
* Filled in missing times_traverse()
* Document reasons that imap() did not adopt a None fill-in feature
* Document that count(sys.maxint) will wrap-around on overflow
* Add overflow test to islice()
* Check that starmap()'s argument returns a tuple
* Verify that imap()'s tuple re-use is safe
* Make a similar tuple re-use (with safety check) for izip()
Right now the test cases create a files and a directory in the temp.
directory. Raymond suggested checking files in to the test/ directory,
simplifying the setup/teardown methods; is that worth doing?
Apparently MAC OS 9 doesn't have POSIX-conforming timestamps. A test
fails as a result, but at least for this specific test it's easy enough
to get the POSIX epoch out of it.
guarantee to keep valid pointers in its slots.
tests: Moved ExtensionSaver from test_copy_reg into pickletester, and
use it both places. Once extension codes get assigned, it won't be
safe to overwrite them willy nilly in test suites, and ExtensionSaver
does a thorough job of undoing any possible damage.
Beefed up the EXT[124] tests a bit, to check the smallest and largest
codes in each opcode's range too.
blindly assumed that tp_as_sequence->sq_item always returns
a str or unicode object. This might fail with str or unicode
subclasses.
This patch checks whether the object returned from __getitem__
is a str/unicode object and raises a TypeError if not (and
the filter function returned true).
Furthermore the result for __getitem__ can be more than one
character long, so checks for enough memory have to be done.
module. This increases code coverage of Python/sysmodule.c
from 68% to 77% (on Linux).
The script doesn't exercise the error branch that handles an evil
or lost sys.excepthook in Python/pythonrun.c::PyErr_PrintEx().
Also this script might not work on Jython in its current form.
From SF patch #662807.
extension implemented flush() was fixed. Scott also rewrite the
zlib test suite using the unittest module. (SF bug #640230 and
patch #678531.)
Backport candidate I think.
are actually getting generated. Add helpered method
ensure_opcode_in_pickle to do a correct job checking for that. Changed
test_long1(), test_long4(), and test_short_tuples() to use it.
huge. On older Linux systems, the C library's strtod() apparently
gives up before seeing the end of the string when it sees so many
digits that it thinks the result must be Infinity. (It is wrong, BTW
-- there could be an "e-10000" hiding behind 10,000 digits.) The
shorter shuge still tests what it's testing, without relying on
strtod() doing a super job.
exception, ResourceDenied. This is used to distinguish between tests that
are skipped for other reasons (platform support, missing data, etc.) from
those that are skipped because a "resource" has not been enabled. This
prevents those tests from being reported as unexpected skips for the
platform; those should only be considered unexpected skips if the resource
were enabled.
loops. Renamed DATA and BINDATA to DATA0 and DATA1. Included
disassemblies, but noted why we can't test them. Added XXX comment to
cPickle about a mysterious comment, where pickle and cPickle diverge
in how they number PUT indices.
Assorted code cleanups; e.g., sizeof(char) is 1 by definition, so there's
no need to do things like multiply by sizeof(char) in hairy malloc
arguments. Fixed an undetected-overflow bug in readline_file().
longobject.c: Fixed a really stupid bug in the new _PyLong_NumBits.
pickle.py: Fixed stupid bug in save_long(): When proto is 2, it
wrote LONG1 or LONG4, but forgot to return then -- it went on to
append the proto 1 LONG opcode too.
Fixed equally stupid cancelling bugs in load_long1() and
load_long4(): they *returned* the unpickled long instead of pushing
it on the stack. The return values were ignored. Tests passed
before only because save_long() pickled the long twice.
Fixed bugs in encode_long().
Noted that decode_long() is quadratic-time despite our hopes,
because long(string, 16) is still quadratic-time in len(string).
It's hex() that's linear-time. I don't know a way to make decode_long()
linear-time in Python, short of maybe transforming the 256's-complement
bytes into marshal's funky internal format, and letting marshal decode
that. It would be more valuable to make long(string, 16) linear time.
pickletester.py: Added a global "protocols" vector so tests can try
all the protocols in a sane way. Changed test_ints() and test_unicode()
to do so. Added a new test_long(), but the tail end of it is disabled
because it "takes forever" under pickle.py (but runs very quickly under
cPickle: cPickle proto 2 for longs is linear-time).
anymore either, so don't. This also allows to get rid of obscure code
making __getnewargs__ identical to __getstate__ (hmm ... hope there
wasn't more to this than I realize!).
longer needs to be public, and shoudn't be public because all datetime
objects are immutable. The Python implementation has changed
accordingly, but still need to change the C implementation.
__module__ is the string name of the module the function was defined
in, just like __module__ of classes. In some cases, particularly for
C functions, the __module__ may be None.
Change PyCFunction_New() from a function to a macro, but keep an
unused copy of the function around so that we don't change the binary
API.
Change pickle's save_global() to use whichmodule() if __module__ is
None, but add the __module__ logic to whichmodule() since it might be
used outside of pickle.
error handers in the Unicode codecs: Negative
positions are treated as being relative to the end of
the input and out of bounds positions result in an
IndexError.
Also update the PEP and include an explanation of
this in the documentation for codecs.register_error.
Fixes a small bug in iconv_codecs: if the position
from the callback is negative *add* it to the size
instead of substracting it.
From SF patch #677429.
classes have a __reduce__ that returns (self.__class__,
self.__getstate__()). tzinfo.__reduce__() is a bit smarter, calling
__getinitargs__ and __getstate__ if they exist, and falling back to
__dict__ if it exists and isn't empty.
for this iconv() implementation in the init function.
For encoding: use a byteswapped version of the input if
neccessary.
For decoding: byteswap every piece returned by iconv()
if neccessary (but not those pieces returned from the
callback)
Comment out test_sane() in the test script, because
whether this works depends on whether byte swapping
is neccessary or not (an on Py_UNICODE_SIZE)
types. The special handling for these can now be removed from save_newobj().
Add some testing for this.
Also add support for setting the 'fast' flag on the Python Pickler class,
which suppresses use of the memo.
only get run by test_pickle.py now (& not by test_cpickle.py). This
should be undone when protocol 2 is implemented in cPickle too.
test_cpickle should pass again.
Gustavo Niemeyer
Martin v. Löwis
Guido van Rossum
Mark Hammond
Brett Cannon
Neal Norwitz
Fred Drake
Thomas Heller
Jack Jansen
Walter Dörwald
Tim Peters
Barry Warsaw
Skip Montanaro
Raymond Hettinger
Jeremy Hylton
Thomas Wouters
Michael W. Hudson
Andrew M. Kuchling
Just van Rossum
Jason Tishler
Marc-André Lemburg