Tim Peters [Thu, 6 Dec 2001 23:16:09 +0000 (23:16 +0000)]
For ratecv, document that None should be the initial state argument (when
I was squashing spurious overflows in the implementation, I got hung up
on this point).
Fred Drake [Thu, 6 Dec 2001 22:59:54 +0000 (22:59 +0000)]
Visious hackery to solve a build-control problem related to our use of
distutils for the library modules built as shared objects. A better solution
appears possible, but with the threat that the distutils becomes more
magical ("complex").
This closes SF bug #458343.
Jack Jansen [Thu, 6 Dec 2001 22:58:56 +0000 (22:58 +0000)]
Don't fail on importing things with undefined references. Unfortunately we
still fail on importing modules that link with libraries that fail
their initialization code (such as windowing libraries when we don't have
access to the window server) and that is what I really wanted to fix.
Jack Jansen [Thu, 6 Dec 2001 21:47:20 +0000 (21:47 +0000)]
Test wether we are building on a case-insensitive filesystem (such
as OSX HFS+) and if so add an extension to the python executable, but
only in the build directory, not on the installed python.
Fix for SF bug #489671 (Neil Norwitz): memory leak in test_richcmp.
Had nothing to do with rich comparisons -- some stack cleanup code was
lost as a result of merging in Neil Schemenauer's generators patch.
Reinserted the stack cleanup code, skipping it when yielding.
Thomas Heller [Thu, 6 Dec 2001 20:44:19 +0000 (20:44 +0000)]
Use a version number of 0.0.0 instead of ???. The latter leads to
invalid filenames on Windows when building without specifying a
version number in the setup script.
See also
http://mail.python.org/pipermail/distutils-sig/2001-November/002656.html
Fix for #489669 (Neil Norwitz): memory leak in test_descr (unicode).
This is best reproduced by
while 1:
class U(unicode):
pass
U(u"xxxxxx")
The unicode_dealloc() code wasn't properly freeing the str and defenc
fields of the Unicode object when freeing a subtype instance. Fixed
this by a subtle refactoring that actually reduces the amount of code
slightly.
Fred Drake [Thu, 6 Dec 2001 16:51:41 +0000 (16:51 +0000)]
Be more careful about accessing attributes of the parent: if Tk has not been
initialized, this will be None, but the functions will still work (there will
simply be a bogus parent on the screen). Allowing the parent to be None
is useful when testing the functions from an interactive interpreter.
Add an optional keyword paramter "show" to the _QueryString class; when given
it is used to set the -show option to the entry widget. This allows passing
show="*" or the like to askstring(), making it useful for requesting
passwords/passphrases from the user.
This closes SF bug #438517.
Jeremy Hylton [Thu, 6 Dec 2001 15:48:16 +0000 (15:48 +0000)]
Fix memory leak in dict_to_map(), SF bug [ #485152 ] memory leak in test_scope.
PyCell_Set() incremenets the reference count, so the earlier XINCREF
causes a leak.
Also make a number of small performance improvements to the code on
the assumption that most of the time variables are not rebound across
a FastToLocals() / LocalsToFast() pair.
Replace uses of PyCell_Set() and PyCell_Get() with PyCell_SET() and
PyCell_GET(), since the frame is guaranteed to contain cells.
Jeremy Hylton [Thu, 6 Dec 2001 15:18:48 +0000 (15:18 +0000)]
Little stuff.
Add a missing DECREF in an obscure corner. If the str() or repr() of
an object passed to a string interpolation -- e.g. "%s" % obj --
returns a non-string, the returned object was leaked.
Repair an indentation glitch.
Replace a bunch of PyString_AsString() calls (and their ilk) with
macros.
Jeremy Hylton [Thu, 6 Dec 2001 14:34:58 +0000 (14:34 +0000)]
Fix [ #489673 ] memory leak in test_symtable: Free the st_future slot.
The st_future slot of the symtable is not freed by PySymtable_Free()
because it is shared by the symtable and compiling structs in
compiel.c. Since it is shared, it is explicitly deallocated when the
compiling struct is freed.
Fix a typo (probably caused by autocompletion <blush>) that caused a
leak when a class defined a __metaclass__. This fixes the problem
reported on python-dev by Ping; I dunno if it's the same as SF bug
#489669 (since that mentions Unicode).
Tim Peters [Thu, 6 Dec 2001 06:23:26 +0000 (06:23 +0000)]
SF bug #488514: -Qnew needs work
Big Hammer to implement -Qnew as PEP 238 says it should work (a global
option affecting all instances of "/").
pydebug.h, main.c, pythonrun.c: define a private _Py_QnewFlag flag, true
iff -Qnew is passed on the command line. This should go away (as the
comments say) when true division becomes The Rule. This is
deliberately not exposed to runtime inspection or modification: it's
a one-way one-shot switch to pretend you're using Python 3.
ceval.c: when _Py_QnewFlag is set, treat BINARY_DIVIDE as
BINARY_TRUE_DIVIDE.
test_{descr, generators, zipfile}.py: fiddle so these pass under
-Qnew too. This was just a matter of s!/!//! in test_generators and
test_zipfile. test_descr was trickier, as testbinop() is passed
assumptions that "/" is the same as calling a "__div__" method; put
a temporary hack there to call "__truediv__" instead when the method
name is "__div__" and 1/2 evaluates to 0.5.
Three standard tests still fail under -Qnew (on Windows; somebody
please try the Linux tests with -Qnew too! Linux runs a whole bunch
of tests Windows doesn't):
test_augassign
test_class
test_coercion
I can't stay awake longer to stare at this (be my guest). Offhand
cures weren't obvious, nor was it even obvious that cures are possible
without major hackery.
Question: when -Qnew is in effect, should calls to __div__ magically
change into calls to __truediv__? See "major hackery" at tail end of
last paragraph <wink>.
The previous checkin to clear __slots__ variables did a little bit of
the work each time it found another base class. All the work is
contiguous, so we might as well do it all at once at the end.
Jack Jansen [Wed, 5 Dec 2001 22:46:23 +0000 (22:46 +0000)]
Link more modules with weak import, and add CarbonAccessors.o to all relevant PowerPC toolbox modules. This, in combination with recent IDE mods, makes the IDE work again under MacOS 8.1.
It was easier than I thought, assuming that no other things contribute
to the instance size besides slots -- a pretty good bet. With a test
suite, no less!
Tim Peters [Wed, 5 Dec 2001 22:30:21 +0000 (22:30 +0000)]
audioop_ratecv(): I left a potentially unsafe multiply unchecked
yesterday -- repair that. Also renamed the silly size_times_nchannels
to bytes_per_frame.
Fred Drake [Wed, 5 Dec 2001 22:10:44 +0000 (22:10 +0000)]
Fix memory leak in the parser module: There were two leaks in
parser_tuple2st() and a failure to propogate an error in
build_node_children() (masking yet another leak, of course!).
This closes SF bug #485133 (confirmed by Insure++).
At the PythonLabs meeting someone mentioned it would make Jim really
happy if one could delete the __dict__ attribute of an instance. I
love to make Jim happy, so here goes...
- New-style objects now support deleting their __dict__. This is for
all intents and purposes equivalent to assigning a brand new empty
dictionary, but saves space if the object is not used further.
Fred Drake [Wed, 5 Dec 2001 15:58:29 +0000 (15:58 +0000)]
Separate the script portion from the library portion; everything that
pertains to the script is now in the if __name__ == "__main__" block.
This is in response to a commenton python-dev from Neal Norwitz.
Tim Peters [Wed, 5 Dec 2001 06:05:07 +0000 (06:05 +0000)]
SF bug 482574: audioop.ratecv crashes.
Bugfix candidate.
A numerically naive computation of output buffer size caused crashes
and spurious MemoryErrors for reasonable arguments.
audioop_ratecv(): Avoid spurious overflow by careful reworking of the
buffer size computations, triggering MemoryError if and only if the
final buffer size can't be represented in a C int (although
PyString_FromStringAndSize may legitimately raise MemoryError even if
it does fit in a C int). All reasonable arguments should work as
intended now, and all unreasonable arguments should be cuaght.
Fred Drake [Wed, 5 Dec 2001 05:46:25 +0000 (05:46 +0000)]
Add a note to the description of the interaction between the softspace
attribute of file objects, the print statement, and other file operations.
This closes SF bug #484857.
Fred Drake [Wed, 5 Dec 2001 05:25:59 +0000 (05:25 +0000)]
Added documentation of the sendall() method, and a note to the send() method
that it does not guarantee that all data is sent.
This closes SF patch #474307.
Tim Peters [Tue, 4 Dec 2001 23:05:10 +0000 (23:05 +0000)]
SF bug #488480: integer multiply to return -max_int-1.
int_mul(): new and vastly simpler overflow checking. Whether it's
faster or slower will likely vary across platforms, favoring boxes
with fast floating point. OTOH, we no longer have to worry about
people shipping broken LONG_BIT definitions <0.9 wink>.
Tim Peters [Tue, 4 Dec 2001 20:06:11 +0000 (20:06 +0000)]
Stop defining NDEBUG in Python.h, because it can interfere with
extensions that #include Python.h. See (rejected) patch 487634 for
more detail. I'll open a new bug report for the rest needed here.
Fred Drake [Tue, 4 Dec 2001 19:47:46 +0000 (19:47 +0000)]
Include a warning that scripts should not have the same name as standard
modules, or the module cannot be properly imported. (Based on a suggestion
sent to python-docs.)
Update the displayed dir() of the sys and __builtin__ module with Python 2.2.
Fred Drake [Tue, 4 Dec 2001 19:20:43 +0000 (19:20 +0000)]
Talk about str() in the discussion of string representations of values, and
give examples for which str() and repr() yield different results.
This closes SF bug #485446.
Ka-Ping Yee [Tue, 4 Dec 2001 18:45:17 +0000 (18:45 +0000)]
Add "file" argument to Hook constructor.
By default, save sys.stdout in self.file when a Hook instance is created
(e.g. when cgitb.enable() is called).
Fix SF bug #486144: Uninitialized __slot__ vrbl is None.
There's now a new structmember code, T_OBJECT_EX, which is used for
all __slot__ variables (except __weakref__, which has special behavior
anyway). This new code raises AttributeError when the variable is
NULL rather than converting NULL to None.
Fred Drake [Tue, 4 Dec 2001 17:03:54 +0000 (17:03 +0000)]
Fix SF bug #479967: Appearantly I broke something that made the index
insertion work. This fix makes things at least somewhat more explicit, and
adds a little sanity checking (and verbosity!) to
add_bbl_and_idx_dummy_commands().
Fix SF bug #486144: Uninitialized __slot__ vrbl is None.
There's now a new structmember code, T_OBJECT_EX, which is used for
all __slot__ variables (except __weakref__, which has special behavior
anyway). This new code raises AttributeError when the variable is
NULL rather than converting NULL to None.
Jack Jansen [Tue, 4 Dec 2001 13:30:29 +0000 (13:30 +0000)]
Patch by Jason Harper to allow IDE to work again under MacOS 8.1. Plus appearance support for Wlist frames and focussing. Plus commented-out appearance support for the same for Wtext, which still needs some work.
Jeremy Hylton [Tue, 4 Dec 2001 02:41:46 +0000 (02:41 +0000)]
SF bug #488687 reported by Neal Norwitz
The error for assignment to __debug__ used ste->ste_opt_lineno instead
of n->n_lineno. The latter was at best incorrect; often the slot was
uninitialized. Two fixes here: Use the correct lineno for the error.
Initialize ste_opt_lineno in PySymtable_New(); while there are no
current cases where it is referenced unless it has already been
assigned to, there is no harm in initializing it.
Fred Drake [Mon, 3 Dec 2001 22:15:56 +0000 (22:15 +0000)]
Slightly improved indexing for the string-% operator, thanks to comments
from Skip Montanaro. There is one weirdness in the final index for HTML, but
that is low priority.
Fix the final two issues in Armin Rigo's SF bug #488477: apply_slice()
and assign_slice() weren't properly DECREF'ing the temporary slice
object they created. (Shame on me. :-)
unpack_iterable(): Add a missing DECREF in an error case. Reported by
Armin Rigo (SF bug #488477). Added a testcase to test_unpack_iter()
in test_iter.py.
Barry Warsaw [Mon, 3 Dec 2001 19:26:40 +0000 (19:26 +0000)]
decode(), encode(): Accepting the minor optimizations from SF patch
#486375, but not the rest of it, since that changes the documented
semantics of encode().
Fred Drake [Mon, 3 Dec 2001 18:27:22 +0000 (18:27 +0000)]
Re-word the intro slightly to avoid reader misunderstanding: strings are not
mutable! We do not want to shock anyone.
This closes SF bug #483805.
Re-factor so that the description of the "access" keyword parameter is not
repeated in both the descriptions of mmap(). Also, only make sure the first
description of mmap() appears in the index. The the index link is followed,
the first is now used to locate the page on the screen; chances are really good
both will be visible. This avoids the problem that the index entry for the
second is selected and the first version is not visible, making the reader
consider that mmap() is not available on Windows.