lnotab-based tracing is very complicated and isn't documented very well. There
were at least 3 comment blocks purporting to document co_lnotab, and none did a
very good job. This patch unifies them into Objects/lnotab_notes.txt which
tries to completely capture the current state of affairs.
I also discovered that we've attached 2 layers of patches to the basic tracing
scheme. The first layer avoids jumping to instructions that don't start a line,
to avoid problems in if statements and while loops. The second layer
discovered that jumps backward do need to trace at instructions that don't
start a line, so it added extra lnotab entries for 'while' and 'for' loops, and
added a special case for backward jumps within the same line. I replaced these
patches by just treating forward and backward jumps differently.
made it try to set the line number from the trace callback for a 'call' event.
This patch makes the error message a little more helpful in that case, and
makes it a little less likely that a future editor will make the same mistake
in test_trace.
Most uses of PyCode_Addr2Line
(http://www.google.com/codesearch?q=PyCode_Addr2Line) are just trying to get
the line number of a specified frame, but there's no way to do that directly.
Forcing people to go through the code object makes them know more about the
guts of the interpreter than they should need.
The remaining uses of PyCode_Addr2Line seem to be getting the line from a
traceback (for example,
http://www.google.com/codesearch/p?hl=en#u_9_nDrchrw/pygame-1.7.1release/src/base.c&q=PyCode_Addr2Line),
which is replaced by the tb_lineno field. So we may be able to deprecate
PyCode_Addr2Line entirely for external use.
Most uses of PyCode_New found by http://www.google.com/codesearch?q=PyCode_New
are trying to build an empty code object, usually to put it in a dummy frame
object. This patch adds a PyCode_NewEmpty wrapper which lets the user specify
just the filename, function name, and first line number, instead of also
requiring lots of code internals.
the __doc__ into the subclass instance __dict__. The fix refactors
property_copy to call property_init in such a way that the __doc__
logic is re-executed correctly when getter_doc is 1, thus simplifying
property_copy.
PyUnicode_DecodeUTF8() once, remember the result and output it in a second
step. This avoids problems with counting UTF-8 bytes that ignores the effect
of using the replace error handler in PyUnicode_DecodeUTF8().
If anyone wants to clean up the documentation, feel free. It's my first documentation foray, and it's not that great.
Will port to py3k with a different strategy.
- simplify parsing and printing of complex numbers
- make complex(repr(z)) round-tripping work for complex
numbers involving nans, infs, or negative zeros
- don't accept some of the stranger complex strings
that were previously allowed---e.g., complex('1..1j')
int, long, and float __format__(), and it keeps their implementation
in sync with py3k.
Also added PyOS_double_to_string. This is the "fallback" version
that's also available in trunk, and should be kept in sync with that
code. I'll add an issue to document PyOS_double_to_string in the C
API.
There are many internal cleanups. Externally visible changes include:
- Implement PEP 378, Format Specifier for Thousands Separator, for
floats, ints, and longs.
- Issue #5515: 'n' formatting for ints, longs, and floats handles
leading zero formatting poorly.
- Issue #5772: For float.__format__, don't add a trailing ".0" if
we're using no type code and we have an exponent.
correctly rounded, using round-half-to-even. This ensures that the
value of float(n) doesn't depend on whether we're using 15-bit digits
or 30-bit digits for Python longs.
untrackable objects are not tracked by the garbage collector. This can
reduce the size of collections and therefore the garbage collection overhead
on long-running programs, depending on their particular use of datatypes.
(trivia: this makes the "binary_trees" benchmark from the Computer Language
Shootout 40% faster)
The basic algorithm remains the same; the most significant speedups
come from the following three changes:
(1) normalize by shifting instead of multiplying and dividing
(2) the old algorithm usually did an unnecessary extra iteration of
the outer loop; remove this. As a special case, this means that
long divisions with a single-digit result run twice as fast as
before.
(3) make inner loop much tighter.
Various benchmarks show speedups of between 50% and 150% for long
integer divisions and modulo operations.
sizeof(Py_UNICODE) == 2, PyUnicode_FromWideChar now converts
each character outside the BMP to the appropriate surrogate pair.
Thanks Victor Stinner for the patch.
(backport of r70452 from py3k to trunk)
For simple uses for str.format(), this makes the typing easier. Hopfully this
will help in the adoption of str.format().
For example:
'The {} is {}'.format('sky', 'blue')
You can mix and matcth auto-numbering and named replacement fields:
'The {} is {color}'.format('sky', color='blue')
But you can't mix and match auto-numbering and specified numbering:
'The {0} is {}'.format('sky', 'blue')
ValueError: cannot switch from manual field specification to automatic field numbering
Will port to 3.1.
and cleanups in Objects/longobject.c. The most significant change is that
longs now use less memory: average savings are 2 bytes per long on 32-bit
systems and 6 bytes per long on 64-bit systems. (This memory saving already
exists in py3k.)
platforms. The previous code was fragile, depending on the twin
accidents that:
(1) in C, casting the double value 2.**63 to long returns the integer
value -2**63, and
(2) in Python, hash(-2**63) == hash(2**63).
There's already a test for this in test_hash.
- fix some places where counters into ob_digit were declared as
int instead of Py_ssize_t
- add (twodigit) casts where necessary
- fix code in _PyLong_AsByteArray that uses << on negative values
#4759: allow None as first argument of bytearray.translate(), for consistency with bytes.translate().
Also fix segfault for bytearray.translate(x, None) -- will backport this part to 3.0 and 2.6.
the __long__ slot is allowed to return either int or long, but the behaviour of
float objects should not change between 2.5 and 2.6.
Reviewed by Benjamin Peterson
exception afterwards (for a subsequent parameter), the user code will
not call PyBuffer_Release() and memory will leak.
Reviewed by Amaury Forgeot d'Arc.
match Python 2.5 speed despite the __instancecheck__ / __subclasscheck__
mechanism. In the process, fix a bug where isinstance() and issubclass(),
when given a tuple of classes as second argument, were looking up
__instancecheck__ / __subclasscheck__ on the tuple rather than on each
type object.
Reviewed by Benjamin Peterson and Raymond Hettinger.
* crashes on memory allocation failure found with failmalloc
* memory leaks found with valgrind
* compiler warnings in opt mode which would lead to invalid memory reads
* problem using wrong name in decimal module reported by pychecker
Update the valgrind suppressions file with new leaks that are small/one-time
leaks we don't care about (ie, they are too hard to fix).
TBR=barry
TESTED=./python -E -tt ./Lib/test/regrtest.py -uall (both debug and opt modes)
in opt mode:
valgrind -q --leak-check=yes --suppressions=Misc/valgrind-python.supp \
./python -E -tt ./Lib/test/regrtest.py -uall,-bsddb,-compiler \
-x test_logging test_ssl test_multiprocessing
valgrind -q --leak-check=yes --suppressions=Misc/valgrind-python.supp \
./python -E -tt ./Lib/test/regrtest.py test_multiprocessing
for i in `seq 1 4000` ; do
LD_PRELOAD=~/local/lib/libfailmalloc.so FAILMALLOC_INTERVAL=$i \
./python -c pass
done
At least some of these fixes should probably be backported to 2.5.
rewrite float.fromhex to only allow ASCII hex digits on all platforms.
(Tests for this are already present, but the test_float failures
on Solaris hadn't been noticed before.)
Reviewed by Antoine Pitrou.
Optimization of str.format() for cases with str, unicode, int, long,
and float arguments. This gives about 30% speed improvement for the
simplest (but most common) cases. This patch skips the __format__
dispatch, and also avoids creating an object to hold the format_spec.
Unfortunately there's a complication in 2.6 with int, long, and float
because they always expect str format_specs. So in the unicode
version of this optimization, just check for unicode objects. int,
float, long, and str can be added later, if needed.
long raise ValueError instead of returning 0. Also, change the error
message for conversion of an infinity to an integer, replacing 'long' by
'integer', so that it's appropriate for both long(float('inf')) and
int(float('inf')).
the freelist contained half-initialized objects with freed pointers.
The comment
/* XXX UNREF/NEWREF interface should be more symmetrical */
was copied from tupleobject.c, and appears in some other places.
I sign the petition.
was not always being done properly in some python types and extension
modules. PyMem_MALLOC, PyMem_REALLOC, PyMem_NEW and PyMem_RESIZE have
all been updated to perform better checks and places in the code that
would previously leak memory on the error path when such an allocation
failed have been fixed.
makes the float constructor behave in the same way as specified
by various other language standards, including C99, IEEE 754r,
and the IBM Decimal standard.
seen after a "import multiprocessing.reduction"
An instance of a weakref subclass can have attributes.
If such a weakref holds the only strong reference to the object,
deleting the weakref will delete the object. In this case,
the callback must not be called, because the ref object is being deleted!
Added checks for integer overflows, contributed by Google. Some are
only available if asserts are left in the code, in cases where they
can't be triggered from Python code.
occurances of PyBytes_ in the code to their original PyString_ names. The
bytesobject.c file will be renamed back to stringobject.c in a future checkin.
renamed Include/bytesobject.h to Include/bytearrayobject.h
renamed Include/stringobject.h to Include/bytesobject.h
added Include/stringobject.h with aliases
Adds 'n' as a format specifier for integers, to mirror the same
specifier which is already available for floats. 'n' is the same as
'd', but inserts the current locale-specific thousands grouping.
I added this as a stringlib function, but it's only used by str type,
not unicode. This is because of an implementation detail in
unicode.format(), which does its own str->unicode conversion. But the
unicode version will be needed in 3.0, and it may be needed by other
code eventually in 2.6 (maybe decimal?), so I left it as a stringlib
implementation. As long as the unicode version isn't instantiated,
there's no overhead for this.
Renamed copy_reg to copyreg in the standard library, to avoid
spurious warnings and ease later merging to py3k branch. Public
documentation remains intact.
The patch also adds acosh, asinh, atanh, log1p and copysign to all platforms. Finally it fixes differences between platforms like different results or exceptions for edge cases. Have fun :)
close() will now raise an IOError if any operations on the file object
are currently in progress in other threads.
Most code was written by Antoine Pitrou (pitrou). Additional testing,
documentation and test suite cleanup done by me (gregory.p.smith).
Fixes issue 815646 and 595601 (as well as many other bugs and
references to this problem dating back to the dawn of Python).
Rather than sprinkle casts throughout the code, change Py_CHARMASK to
always cast it's result to an unsigned char. This should ensure we
do the right thing when accessing an array with the result.
svn+ssh://pythondev@svn.python.org/python/branches/trunk-bytearray
........
r61750 | christian.heimes | 2008-03-22 20:47:44 +0100 (Sat, 22 Mar 2008) | 1 line
Copied files from py3k w/o modifications
........
r61752 | christian.heimes | 2008-03-22 20:53:20 +0100 (Sat, 22 Mar 2008) | 7 lines
Take One
* Added initialization code, warnings, flags etc. to the appropriate places
* Added new buffer interface to string type
* Modified tests
* Modified Makefile.pre.in to compile the new files
* Added bytesobject.c to Python.h
........
r61754 | christian.heimes | 2008-03-22 21:22:19 +0100 (Sat, 22 Mar 2008) | 2 lines
Disabled bytearray.extend for now since it causes an infinite recursion
Fixed serveral unit tests
........
r61756 | christian.heimes | 2008-03-22 21:43:38 +0100 (Sat, 22 Mar 2008) | 5 lines
Added PyBytes support to several places:
str + bytearray
ord(bytearray)
bytearray(str, encoding)
........
r61760 | christian.heimes | 2008-03-22 21:56:32 +0100 (Sat, 22 Mar 2008) | 1 line
Fixed more unit tests related to type('') is not unicode
........
r61763 | christian.heimes | 2008-03-22 22:20:28 +0100 (Sat, 22 Mar 2008) | 2 lines
Fixed more unit tests
Fixed bytearray.extend
........
r61768 | christian.heimes | 2008-03-22 22:40:50 +0100 (Sat, 22 Mar 2008) | 1 line
Implemented old buffer interface for bytearray
........
r61772 | christian.heimes | 2008-03-22 23:24:52 +0100 (Sat, 22 Mar 2008) | 1 line
Added backport of the io module
........
r61775 | christian.heimes | 2008-03-23 03:50:49 +0100 (Sun, 23 Mar 2008) | 1 line
Fix str assignement to bytearray. Assignment of a str of size 1 is interpreted as a single byte
........
r61805 | christian.heimes | 2008-03-23 19:33:48 +0100 (Sun, 23 Mar 2008) | 3 lines
Fixed more tests
Fixed bytearray() comparsion with unicode()
Fixed iterator assignment of bytearray
........
r61809 | christian.heimes | 2008-03-23 21:02:21 +0100 (Sun, 23 Mar 2008) | 2 lines
str(bytesarray()) now returns the bytes and not the representation of the bytearray object
Enabled and fixed more unit tests
........
r61812 | christian.heimes | 2008-03-23 21:53:08 +0100 (Sun, 23 Mar 2008) | 3 lines
Clear error PyNumber_AsSsize_t() fails
Use CHARMASK for ob_svall access
disabled a test with memoryview again
........
r61819 | christian.heimes | 2008-03-23 23:05:57 +0100 (Sun, 23 Mar 2008) | 1 line
Untested updates to the PCBuild directory
........
r61917 | christian.heimes | 2008-03-26 00:57:06 +0100 (Wed, 26 Mar 2008) | 1 line
The type system of Python 2.6 has subtle differences to 3.0's. I've removed the Py_TPFLAGS_BASETYPE flags from bytearray for now. bytearray can't be subclasses until the issues with bytearray subclasses are fixed.
........
r61920 | christian.heimes | 2008-03-26 01:44:08 +0100 (Wed, 26 Mar 2008) | 2 lines
Disabled last failing test
I don't understand what the test is testing and how it suppose to work. Ka-Ping, please check it out.
........
r61930 | christian.heimes | 2008-03-26 12:46:18 +0100 (Wed, 26 Mar 2008) | 1 line
Re-enabled bytes warning code
........
r61933 | christian.heimes | 2008-03-26 13:20:46 +0100 (Wed, 26 Mar 2008) | 1 line
Fixed a bug in the new buffer protocol. The buffer slots weren't copied into a subclass.
........
r61934 | christian.heimes | 2008-03-26 13:25:09 +0100 (Wed, 26 Mar 2008) | 1 line
Re-enabled bytearray subclassing - all tests are passing.
........
Added 0b and 0o literals to tokenizer.
Modified PyOS_strtoul to support 0b and 0o inputs.
Modified PyLong_FromString to support guessing 0b and 0o inputs.
Renamed test_hexoct.py to test_int_literal.py and added binary tests.
Added upper and lower case 0b, 0O, and 0X tests to test_int_literal.py
Added "Z" format_char to PyOS_ascii_formatd to support empty float presentation type.
Renamed buf_size in PyOS_ascii_formatd to more accurately reflect it's meaning.
Modified format.__float__ to use the new "Z" format as the default.
Added test cases.
across platforms: it should now raise OverflowError on all
platforms. (Previously it raised OverflowError only on
non IEEE 754 platforms.)
Also fix the (already existing) test for this behaviour
so that it actually raises TestFailed instead of just
referencing it.
which forbids constructing types that have it set. The effect is to speed
./python.exe -m timeit -s 'import abc' -s 'class Foo(object): __metaclass__ = abc.ABCMeta' 'Foo()'
up from 2.5us to 0.201us. This fixes issue 1762.
Highlights:
- Adding PyObject_Format.
- Adding string.Format class.
- Adding __format__ for str, unicode, int, long, float, datetime.
- Adding builtin format.
- Adding ''.format and u''.format.
- str/unicode fixups for formatters.
The files in Objects/stringlib that implement PEP 3101 (stringdefs.h,
unicodedefs.h, formatter.h, string_format.h) are identical in trunk
and py3k. Any changes from here on should be made to trunk, and
changes will propogate to py3k).
_PyLong_Format. In longobject.c, changed long_format to
_PyLong_Format. In intobject.c, changed uses of PyOS_snprintf to
_PyInt_Format instead.
_PyLong_Format is similar to py3k's routine of the same name, except
it has 2 additional parameters: addL and newstyle. addL was existing
in long_format, and controls adding the trailing "L". This is
unneeded in py3k. newstyle is used to control whether octal prepends
"0" (the pre-2.6 style), or "0o" (the 3.0 sytle).
PyNumber_ToBase is needed for PEP 3127 (Integer Literal Support and
Syntax) and PEP 3101 (Advanced String Formatting).
This changeset does not need merging into py3k.
Jeffrey Yasskin
Eric Smith
Mark Dickinson
Hirokazu Yamamoto
Benjamin Peterson
Georg Brandl
R. David Murray
Antoine Pitrou
Walter Dörwald
Martin v. Löwis
Kristján Valur Jónsson
Raymond Hettinger
Amaury Forgeot d'Arc
Nick Coghlan
Alexandre Vassalotti
Andrew M. Kuchling
Christian Heimes
Andrew MacIntyre
Neal Norwitz
Brett Cannon
Gregory P. Smith
Robert Schuppenies
Skip Montanaro
Hye-Shik Chang
Travis E. Oliphant
Steven Bethard
Guido van Rossum
Facundo Batista