Commit Graph

4934 Commits

Author SHA1 Message Date
Serhiy Storchaka 31a655411a Issue #17576: Deprecation warning emitted now when __int__() or __index__()
return not int instance.  Introduced _PyLong_FromNbInt() and refactored
PyLong_As*() functions.
2013-12-11 21:07:54 +02:00
Christian Heimes d3afe781b1 Silence expression result unused warnings with clang.
The PyObject_INIT() macros returns obj:

../cpython/Objects/methodobject.c:32:23: warning: expression result unused [-Wunused-value]
        PyObject_INIT(op, &PyCFunction_Type);
                      ^~
../cpython/Include/objimpl.h:139:69: note: expanded from macro 'PyObject_INIT'
    ( Py_TYPE(op) = (typeobj), _Py_NewReference((PyObject *)(op)), (op) )
                                                                    ^
1 warning generated.
2013-12-04 09:27:47 +01:00
Alexandre Vassalotti 2ccf8e969c Issue #6477: Merge with 3.3. 2013-11-30 17:58:53 -08:00
Alexandre Vassalotti 65846c6c51 Issue #6477: Keep PyNotImplemented_Type and PyNone_Type private. 2013-11-30 17:55:48 -08:00
Alexandre Vassalotti 3c23e7a5dc Issue #6477: Merge with 3.3. 2013-11-30 16:21:20 -08:00
Alexandre Vassalotti 19b6fa6ebb Issue #6477: Added support for pickling the types of built-in singletons. 2013-11-30 16:06:39 -08:00
Alexandre Vassalotti cc6e87b2cf Issue #19088: Merge with 3.3. 2013-11-30 01:05:51 -08:00
Alexandre Vassalotti 1a83070d9e Issue #19088: Fix incorrect caching of the copyreg module.
This fix does not cause any degradation in performance.
2013-11-30 00:53:09 -08:00
Benjamin Peterson 0ee22bf774 fix format spec recursive expansion (closes #19729) 2013-11-26 19:22:36 -06:00
Gregory P. Smith 65482570f1 Remove an errant extra \ within a docstring. 2013-11-25 00:31:31 -08:00
Gregory P. Smith 0e524047e2 Remove an errant extra \ within a docstring. 2013-11-25 00:30:56 -08:00
Raymond Hettinger 799f81b14c merge 2013-11-24 14:53:54 -08:00
Raymond Hettinger 97fc2ba6bf Document that @property can incorporate a docstring from the getter method. Improve readabilty with additional whitespace. 2013-11-24 14:53:29 -08:00
Alexandre Vassalotti c49477b184 Make Ellipsis and NotImplemented picklable through the reduce protocol. 2013-11-24 02:53:45 -08:00
Alexandre Vassalotti 4c05d3bc56 Make built-in methods picklable through the reduce protocol. 2013-11-24 02:41:05 -08:00
Zachary Ware a4b7a7548c Issue #3158: doctest can now find doctests in functions and methods
written in C.

As a part of this, a few doctests have been added to the builtins module
(on hex(), oct(), and bin()), a doctest has been fixed (hopefully on all
platforms) on float, and test_builtins now runs doctests in builtins.
2013-11-24 01:19:09 -06:00
Larry Hastings 44e2eaab54 Issue #19674: inspect.signature() now produces a correct signature
for some builtins.
2013-11-23 15:37:55 -08:00
Larry Hastings ebdcb50b8a Issue #19730: Argument Clinic now supports all the existing PyArg
"format units" as legacy converters, as well as two new features:
"self converters" and the "version" directive.
2013-11-23 14:54:00 -08:00
Christian Heimes 2489bd83f5 Issue #17810: Fixed NULL check in _PyObject_GetItemsIter()
CID 1131948:  Logically dead code  (DEADCODE)
2013-11-23 21:19:43 +01:00
Antoine Pitrou c9dc4a2a8a Issue #17810: Implement PEP 3154, pickle protocol 4.
Most of the work is by Alexandre.
2013-11-23 18:59:12 +01:00
Eric Snow b523f8433a Implement PEP 451 (ModuleSpec). 2013-11-22 09:05:39 -07:00
Nick Coghlan c72e4e6dcc Issue #19619: Blacklist non-text codecs in method API
str.encode, bytes.decode and bytearray.decode now use an
internal API to throw LookupError for known non-text encodings,
rather than attempting the encoding or decoding operation and
then throwing a TypeError for an unexpected output type.

The latter mechanism remains in place for third party non-text
encodings.
2013-11-22 22:39:36 +10:00
Victor Stinner 8455723cfb Close #19568: Fix bytearray_setslice_linear(), fix handling of
PyByteArray_Resize() failure: leave the bytearray object in an consistent state.

If growth < 0, handling the memory allocation failure is tricky here because
the bytearray object has already been modified.  If lo != 0, the operation is
completed, but a MemoryError is still raised and the memory block is not
shrinked. If lo == 0, the bytearray is restored in its previous state and a
MemoryError is raised.
2013-11-21 12:29:51 +01:00
Victor Stinner 35f2803a24 Close #19578: Fix list_ass_subscript(), handle list_resize() failure
Notify the caller of the failure (MemoryError exception).
2013-11-21 12:16:35 +01:00
Christian Heimes 985ecdcfc2 ssue #19183: Implement PEP 456 'secure and interchangeable hash algorithm'.
Python now uses SipHash24 on all major platforms.
2013-11-20 11:46:18 +01:00
Serhiy Storchaka dc2fd5101a Remove dead code committed in issue #12892. 2013-11-19 15:56:05 +02:00
Nick Coghlan f1de55fb33 Also chain codec exceptions that allow weakrefs
The zlib and hex codecs throw custom exception types with
weakref support if the input type is valid, but the data
fails validation. Make sure the exception chaining in the
codec infrastructure can wrap those as well.
2013-11-19 22:33:10 +10:00
Victor Stinner f91929b1d8 Issue #19646: repr(dict) now uses _PyUnicodeWriter API for better performances 2013-11-19 13:07:38 +01:00
Victor Stinner 88a9cd9b57 Issue #19513: repr(tuple) now uses _PyUnicodeWriter for better performances 2013-11-19 12:59:46 +01:00
Victor Stinner 4a58707a34 Add _PyUnicodeWriter_WriteASCIIString() function 2013-11-19 12:54:53 +01:00
Victor Stinner 4d3f109ad3 Issue #19513: Disable overallocation of the PyUnicodeWriter before the last write 2013-11-19 12:09:00 +01:00
Serhiy Storchaka 58cf607d13 Issue #12892: The utf-16* and utf-32* codecs now reject (lone) surrogates.
The utf-16* and utf-32* encoders no longer allow surrogate code points
(U+D800-U+DFFF) to be encoded.
The utf-32* decoders no longer decode byte sequences that correspond to
surrogate code points.
The surrogatepass error handler now works with the utf-16* and utf-32* codecs.

Based on patches by Victor Stinner and Kang-Hao (Kenny) Lu.
2013-11-19 11:32:41 +02:00
Victor Stinner b8fb197aa0 Issue #19513: Simplify list_repr() 2013-11-18 22:15:44 +01:00
Victor Stinner 5c733473f2 Issue #19513: repr(list) now uses the PyUnicodeWriter API, it is faster than
the PyAccu API
2013-11-18 21:11:57 +01:00
Victor Stinner 6989ba0174 Issue #19581: Change the overallocation factor of _PyUnicodeWriter on Windows
On Windows, a factor of 50% gives best performances.
2013-11-18 21:08:39 +01:00
Larry Hastings ed4a1c5703 Argument Clinic: rename "self" to "module" for module-level functions. 2013-11-18 09:32:13 -08:00
Ezio Melotti 745d54d2fa #17806: Added keyword-argument support for "tabsize" to str/bytes.expandtabs(). 2013-11-16 19:10:57 +02:00
Nick Coghlan 4b9b936429 Don't decref exc too soon 2013-11-16 00:34:13 +10:00
Victor Stinner 46ef31953e Issue #19429, #19437: fix error handling in the OSError constructor 2013-11-14 22:31:41 +01:00
Benjamin Peterson e109ee8205 fix refleaks 2013-11-13 23:49:49 -05:00
Benjamin Peterson 079c998872 adjust style 2013-11-13 23:25:01 -05:00
Christian Heimes 6a3db25c70 Issue #17828: _PyObject_GetDictPtr() may return NULL instead of a PyObject**
CID 1128792: Dereference null return value (NULL_RETURNS)
2013-11-14 01:47:14 +01:00
Christian Heimes 507eabdf11 Issue #17828: va_start() must be accompanied by va_end()
CID 1128793: Missing varargs init or cleanup (VARARGS)
2013-11-14 01:39:35 +01:00
Nick Coghlan 8b097b4ed7 Close #17828: better handling of codec errors
- output type errors now redirect users to the type-neutral
  convenience functions in the codecs module
- stateless errors that occur during encoding and decoding
  will now be automatically wrapped in exceptions that give
  the name of the codec involved
2013-11-13 23:49:21 +10:00
Victor Stinner 59799a8399 Don't use deprecated function PyUnicode_GET_SIZE()
Replace it with PyUnicode_GET_LENGTH() or PyUnicode_AsUnicodeAndSize()
2013-11-13 14:17:30 +01:00
Victor Stinner 3f36a5736b Issue #19515: Remove identifiers duplicated in the same file.
Patch written by Andrei Dorian Duma.
2013-11-12 21:39:02 +01:00
Victor Stinner 66b3270975 _Py_normalize_encoding(): explain how the value 6 was computed 2013-11-07 23:12:23 +01:00
Victor Stinner bd303c165b Issue #19512, #19515: remove shared identifiers, move identifiers where they
are used.

Move also _Py_IDENTIFIER() defintions to the top in modified files to remove
identifiers duplicated in the same file.
2013-11-07 23:07:29 +01:00
Victor Stinner 07e9e380f9 frameobject.c: Use an identifer instead of creating explicitly an interned
string for "__builtins__" literal string
2013-11-07 22:22:39 +01:00
Victor Stinner df23e30bea Fix _Py_normalize_encoding(): ensure that buffer is big enough to store "utf-8"
if the input string is NULL
2013-11-07 13:33:36 +01:00
Martin v. Löwis e75fc14813 Issue #19514: Deduplicate some _Py_IDENTIFIER declarations.
Patch by Andrei Dorian Duma.
2013-11-07 18:46:53 +01:00
Ezio Melotti a5b9599538 #17080: improve error message of float/complex when the wrong type is passed. 2013-11-07 19:18:34 +02:00
Victor Stinner 53e9ec48e5 Issue #19512: Use the new _PyId_builtins identifier 2013-11-07 00:43:05 +01:00
Victor Stinner ad14ccd047 Issue #19512: add _PyUnicode_CompareWithId() function
_PyUnicode_CompareWithId() is faster than PyUnicode_CompareWithASCIIString()
when both strings are equal and interned.

Add also _PyId_builtins identifier for "builtins" common string.
2013-11-07 00:46:04 +01:00
Victor Stinner 3688aa9a04 Issue #19512: type_abstractmethods() and type_set_abstractmethods() now use an
identifier for the "__abstractmethods__" string
2013-11-06 18:59:18 +01:00
Victor Stinner 5fd2e5ae8a Issue #19512: Add a new _PyDict_DelItemId() function, similar to
PyDict_DelItemString() but using an identifier for the key
2013-11-06 18:58:22 +01:00
Victor Stinner 7a07e451a4 Issue #19512: Py_ReprEnter() and Py_ReprLeave() now use an identifier for the
"Py_Repr" dictionary key
2013-11-06 18:57:29 +01:00
Victor Stinner 21ea21ef6d Issue #19424: PyUnicode_CompareWithASCIIString() normalizes memcmp() result
to -1, 0, 1
2013-11-04 11:28:26 +01:00
Victor Stinner f0c7b2af05 Issue #16286: remove duplicated identity check from unicode_compare()
Move the test to PyUnicode_Compare()
2013-11-04 11:27:14 +01:00
Victor Stinner fd9e44db37 Issue #16286: optimize PyUnicode_RichCompare() for identical strings (same
pointer) for any operator, not only Py_EQ and Py_NE.

Code of bytes_richcompare() and PyUnicode_RichCompare() is now closer.
2013-11-04 11:23:05 +01:00
Victor Stinner c8bc5377ac Issue #16286: write a new subfunction bytes_compare_eq()
* cleanup bytes_richcompare()
* PyUnicode_RichCompare(): replace a test with a XOR
2013-11-04 11:08:10 +01:00
Victor Stinner e1b1592fd4 Issue #19424: Fix a compiler warning on comparing signed/unsigned size_t
Patch written by Zachary Ware.
2013-11-03 13:53:12 +01:00
Victor Stinner a6b9b071a3 Issue #19424: Fix a compiler warning
memcmp() just takes raw pointers
2013-10-30 18:27:13 +01:00
Victor Stinner 602f7cf0b9 Issue #19424: Optimize PyUnicode_CompareWithASCIIString()
Use fast memcmp() instead of a loop using the slow PyUnicode_READ() macro.
strlen() is still necessary to check Unicode string containing null bytes.
2013-10-29 23:31:50 +01:00
Antoine Pitrou 84745ab464 Issue #17936: Fix O(n**2) behaviour when adding or removing many subclasses of a given type. 2013-10-29 21:31:25 +01:00
Victor Stinner 68b674c9d4 Issue #19437: Fix _PyUnicode_New() (constructor of legacy string), set all
attributes before checking for error. The destructor expects all attributes to
be set. It is now safe to call Py_DECREF(unicode) in the constructor.
2013-10-29 19:31:43 +01:00
Victor Stinner 0b0c867178 Issue #19437: Fix PyObject_CallFunction(), handle Py_VaBuildValue() and
PyTuple_New() failure
2013-10-29 19:29:52 +01:00
Victor Stinner fa3ba4c3bc Issue #18609: Add a fast-path for "iso8859-1" encoding
On AIX, the locale encoding may be "iso8859-1", which was not a known syntax of
the legacy ISO 8859-1 encoding.

Using a C codec instead of a Python codec is faster but also avoids tricky
issues during Python startup or complex code.
2013-10-29 11:34:05 +01:00
Victor Stinner bebba5059c fix indent 2013-10-29 10:56:34 +01:00
Victor Stinner cc64eb5b9f Issue #18408: Fix bytearrayiter.partition()/rpartition(), handle
PyByteArray_FromStringAndSize() failure (ex: on memory allocation failure)
2013-10-29 03:15:37 +01:00
Victor Stinner 986e224d5a Issue #18408: Fix error handling in PyBytes_FromObject()
_PyBytes_Resize(&new) sets new to NULL on error, don't call Py_DECREF() with NULL.
2013-10-29 03:14:22 +01:00
Victor Stinner a5afb58986 Issue #18408: Fix PyUnicode_AsUTF8AndSize(), raise MemoryError exception on
memory allocation failure
2013-10-29 01:28:23 +01:00
Victor Stinner 41bb43a71e Issue #18408: Add a new PyFrame_FastToLocalsWithError() function to handle
exceptions when merging fast locals into f_locals of a frame.
PyEval_GetLocals() now raises an exception and return NULL on failure.
2013-10-29 01:19:37 +01:00
Serhiy Storchaka f740d467bf Issue #19369: Optimized the usage of __length_hint__(). 2013-10-24 23:19:51 +03:00
Serhiy Storchaka c679227e31 Issue #1772673: The type of `char*` arguments now changed to `const char*`. 2013-10-19 21:03:34 +03:00
Serhiy Storchaka 55e092f545 Issue #19279: UTF-7 decoder no more produces illegal strings. 2013-10-19 20:39:28 +03:00
Serhiy Storchaka 35804e4c63 Issue #19279: UTF-7 decoder no more produces illegal strings. 2013-10-19 20:38:19 +03:00
Larry Hastings 3182680210 Issue #16612: Add "Argument Clinic", a compile-time preprocessor
for C files to generate argument parsing code.  (See PEP 436.)
2013-10-19 00:09:25 -07:00
Benjamin Peterson 94d08d908b upgrade unicode db to 6.3.0 (closes #19221) 2013-10-10 17:24:45 -04:00
Victor Stinner 6cf185dc06 Issue #18874: _PyObject_Malloc/Realloc/Free() now falls back on
_PyMem_RawMalloc/Realloc/Free, instead of _PyMem_Malloc/Realloc/Free.  So it
becomes possible to use the fast pymalloc allocator for the PYMEM_DOMAIN_MEM
domain (PyMem_Malloc/Realloc/Free functions).
2013-10-10 15:58:42 +02:00
Victor Stinner 7c74de4d00 Issue #18874: PyCode_New() now ensures that the filename is a ready Unicode
string. This change does nothing is most cases, but it is useful on Windows in
some cases.
2013-10-10 15:55:14 +02:00
Tim Peters 9259c21a63 Issue #19171: speed some cases of 3-argument long pow().
Reduce the base by the modulus when the base is larger than
the modulus.  This can unboundedly speed the "startup costs"
of doing modular exponentiation, particularly in cases where
the base is much larger than the modulus.  Original patch
by Armin Rigo, inspired by https://github.com/pyca/ed25519.

Merged from 3.3.
2013-10-05 16:55:38 -05:00
Tim Peters 81a93159d7 Issue #19171: speed some cases of 3-argument long pow().
Reduce the base by the modulus when the base is larger than
the modulus.  This can unboundedly speed the "startup costs"
of doing modular exponentiation, particularly in cases where
the base is much larger than the modulus.  Original patch
by Armin Rigo, inspired by https://github.com/pyca/ed25519.
2013-10-05 16:53:52 -05:00
Ezio Melotti e898153c0f #19069: merge with 3.3. 2013-10-06 00:46:58 +03:00
Ezio Melotti 7760b4eb4b #19069: use imperative mood in float object docstrings. Patch by Marco Buttu. 2013-10-06 00:45:11 +03:00
Ezio Melotti c1ae254717 #19068: merge with 3.3. 2013-10-06 00:39:44 +03:00
Ezio Melotti 488d244e1a #19068: use imperative mood in complex object docstrings. Patch by Marco Buttu. 2013-10-06 00:39:18 +03:00
Ezio Melotti b32a8111c9 #19067: merge with 3.3. 2013-10-06 00:38:19 +03:00
Ezio Melotti 5792ce151c #19067: use imperative mood in range object docstrings. Patch by Marco Buttu. 2013-10-06 00:36:45 +03:00
Antoine Pitrou 5b72075387 Hopefully fix Windows compilation error following 499a96611baa 2013-10-05 21:24:10 +02:00
Antoine Pitrou 5df8a8a1fd Issue #19087: Improve bytearray allocation in order to allow cheap popping of data at the front (slice deletion). 2013-10-05 21:12:18 +02:00
Antoine Pitrou 0e61ed8400 Issue #19014: memoryview.cast() is now allowed on zero-length views. 2013-10-03 19:56:54 +02:00
Antoine Pitrou 60b183407c Issue #19014: memoryview.cast() is now allowed on zero-length views. 2013-10-03 19:55:41 +02:00
Nick Coghlan a0f169cde8 Close #19078: memoryview now supports reversed
Patch by Claudiu Popa
2013-10-02 22:06:54 +10:00
Raymond Hettinger c13516b0a0 merge 2013-10-01 01:00:59 -07:00
Raymond Hettinger 2ff2190b62 Issue #18594: Fix the fast path for collections.Counter().
The path wasn't being taken due to an over-restrictive type check.
2013-10-01 00:55:43 -07:00
Nick Coghlan 6ba64f454d Close #18596: Support address sanity checking in clang/GCC
This patch appropriately marks known false alarms in the
small object allocator when address sanity checking is
enabled (patch contributed by Dhiru Kholia).
2013-09-29 00:28:55 +10:00
Georg Brandl 7cba5fd267 Fix minor typo. 2013-09-25 09:04:23 +02:00
Raymond Hettinger 710a67edfc Note that LINEAR_PROBES can be set to zero. 2013-09-21 20:17:31 -07:00
Raymond Hettinger 4ef0528b97 Minor beautification. Put updates and declarations in a more logical order. 2013-09-21 15:39:49 -07:00
Raymond Hettinger 0ce1953bf7 When LINEAR_PROBES=0, let the compiler remove the dead code on its own. 2013-09-21 14:07:18 -07:00
Raymond Hettinger c70a2b7bb9 Make the linear probe sequence clearer. 2013-09-21 14:02:55 -07:00
Tim Peters df099f5df6 Update internal comments to say _something_ about the "API ID".
Best I can tell, the possible values for this aren't documented anywhere.
2013-09-19 21:06:37 -05:00
Raymond Hettinger 8408dc581e Issue 18771: Make it possible to set the number linear probes at compile-time. 2013-09-15 14:57:15 -07:00
Raymond Hettinger 742d8716ff Put the defines in the logical section and fix indentation. 2013-09-08 00:25:57 -07:00
Raymond Hettinger 583cd03fd1 Minor code beautification. 2013-09-07 22:06:35 -07:00
Raymond Hettinger 4ea9080da9 Improve code clarity by removing two unattractive macros. 2013-09-07 21:01:29 -07:00
Raymond Hettinger 8f8839e10a Remove the freelist scheme for setobjects.
The setobject freelist was consuming memory but not providing much value.
Even when a freelisted setobject was available, most of the setobject
fields still needed to be initialized and the small table still required
a memset().  This meant that the custom freelisting scheme for sets was
providing almost no incremental benefit over the default Python freelist
scheme used by _PyObject_Malloc() in Objects/obmalloc.c.
2013-09-07 20:26:50 -07:00
Raymond Hettinger 04fd9dd52b Small rearrangement to bring together the three functions for probing the hash table. 2013-09-07 17:41:01 -07:00
Raymond Hettinger ae7b00e2d3 Move the overview comment to the top of the file. 2013-09-07 15:05:00 -07:00
Tim Peters b2372959ab Nerge 3.3 into default.
Issue #18942: sys._debugmallocstats() output was damaged on Windows.

_PyDebugAllocatorStats() called PyOS_snprintf() with a %zd format
code, but MS doesn't support that code.  Interpolated
PY_FORMAT_SIZE_T in place of the "z".
2013-09-05 23:04:26 -05:00
Tim Peters eaa3bcc370 Issue #18942: sys._debugmallocstats() output was damaged on Windows.
_PyDebugAllocatorStats() called PyOS_snprintf() with a %zd format
code, but MS doesn't support that code.  Interpolated
PY_FORMAT_SIZE_T in place of the "z".
2013-09-05 22:57:04 -05:00
Raymond Hettinger c56e0e3980 Minor touchups. 2013-09-02 16:32:27 -07:00
Raymond Hettinger 69492dab07 Factor-out the common code for setting a KeyError. 2013-09-02 15:59:26 -07:00
Raymond Hettinger a35adf5b09 Instead of XORed indicies, switch to a hybrid of linear probing and open addressing.
Modern processors tend to make consecutive memory accesses cheaper than
random probes into memory.

Small sets can fit into L1 cache, so they get less benefit.  But they do
come out ahead because the consecutive probes don't probe the same key
more than once and because the randomization step occurs less frequently
(or not at all).

For the open addressing step, putting the perturb shift before the index
calculation gets the upper bits into play sooner.
2013-09-02 03:23:21 -07:00
Raymond Hettinger 6c3c1ccd1b Update copyright. 2013-08-31 21:34:24 -07:00
Raymond Hettinger 95c0d67581 Further reduce the cost of hash collisions by inspecting an additional nearby entry. 2013-08-31 21:27:08 -07:00
Ethan Furman fb13721b1b Close #18780: %-formatting now prints value for int subclasses with %d, %i, and %u codes. 2013-08-31 10:18:55 -07:00
Raymond Hettinger afe890923f Tighten-up the lookkey() logic and beautify the code a bit.
Use less code by moving many of the steps from the initial
lookup into the main search loop.

Beautify the code but keep the overall logic unchanged.
2013-08-28 20:59:31 -07:00
Serhiy Storchaka 46e1ce214b Issue #18783: Removed existing mentions of Python long type in docstrings,
error messages and comments.
2013-08-27 20:17:03 +03:00
Serhiy Storchaka 9594942716 Issue #18783: Removed existing mentions of Python long type in docstrings,
error messages and comments.
2013-08-27 19:40:23 +03:00
Victor Stinner 33824f6fd7 Restore changeset 5bd9db528aed (issue #18408)
"Issue #18408: PyObject_Str(), PyObject_Repr() and type_call() now fail with an
assertion error if they are called with an exception set (PyErr_Occurred()).

As PyEval_EvalFrameEx(), they may clear the current exception and so the caller
looses its exception."
2013-08-26 14:05:19 +02:00
Victor Stinner e51321020c Issue #18408: _PyObject_Dump() now saves/restores the current exception
So it can be called even if an exception was raised
2013-08-26 13:49:06 +02:00
Tim Peters 1544fc5312 Various clarifications based on feedback & questions over the years. 2013-08-24 15:31:07 -05:00
Tim Peters ec8147ba55 Various clarifications based on feedback & questions over the years.
(grafted from 23181bf411a16287a0a54e910fc0f9ecd2764bf0)
2013-08-24 15:15:19 -05:00
Antoine Pitrou 9d95254bb7 Issue #18772: fix the gdb plugin after the set implementation changes 2013-08-24 21:07:07 +02:00
Antoine Pitrou 91541931f5 Back out 5bd9db528aed (issue #18408). It caused unsolved buildbot failures. 2013-08-23 23:18:20 +02:00
Raymond Hettinger bfc1e1a9cd Add the same dummy type that is used in dictionaries. 2013-08-23 03:22:15 -05:00
Tim Peters a1db94554b Add line explaining the "%sort" test. 2013-08-22 18:42:02 -05:00
Tim Peters 01e75a699d Add line explaining the "%sort" test.
(grafted from 1ea833ecaf5a9d43a886e9e73b4e2551d0d5b548)
2013-08-22 18:32:53 -05:00
Raymond Hettinger fcf3b500ba Issue 18797: Remove unneeded refcount adjustments for dummy objects.
It suffices to keep just one reference when the object is created.
2013-08-22 08:20:31 -07:00
Raymond Hettinger 5bb1b1dd6f Hoist the global dummy lookup out of the inner loop for set_merge(). 2013-08-21 01:34:18 -07:00
Raymond Hettinger 929cbac307 Remove a redundant hash table probe (this was artifact from an earlier draft of the patch). 2013-08-20 23:03:28 -07:00
Raymond Hettinger ae9e616a00 Issue 18772: Restore set dummy object back to unicode and restore the identity checks in lookkey().
The Gdb prettyprint plugin depended on the dummy object being displayable.
Other solutions besides a unicode object are possible.  For now, get it
back up and running.

The identity checks in lookkey() need to be there to prevent the dummy
object from leaking through Py_RichCompareBool() into user code in the
rare circumstance where the dummy's hash value exactly matches the hash
value of the actual key being looked up.
2013-08-20 22:28:24 -07:00
Raymond Hettinger 3c0a4f5def Issue18771: Reduce the cost of hash collisions for set objects. 2013-08-19 07:36:04 -07:00
Raymond Hettinger 07351a0449 Remove the else-clause because the conditions are no longer mutually exclusive. 2013-08-17 02:39:46 -07:00
Raymond Hettinger 237b34b074 Use a known unique object for the dummy entry.
This lets us run PyObject_RichCompareBool() without
first needing to check whether the entry is a dummy.
2013-08-17 02:31:53 -07:00
Serhiy Storchaka 8fa8ee3970 Issue #18701: Remove support of old CPython versions (<3.0) from C code. 2013-08-17 00:48:02 +03:00
Raymond Hettinger 8ad3919577 Hoist the global "dummy" lookup outside of the reinsertion loop. 2013-08-15 02:18:55 -07:00
Raymond Hettinger d06eeb4a24 merge 2013-08-13 18:20:55 -07:00
Raymond Hettinger b1b915c796 Issue 18719: Remove a false optimization
Remove an unused early-out test from the critical path for
dict and set lookups.

When the strings already have matching lengths, kinds, and hashes,
there is no additional information gained by checking the first
characters (the probability of a mismatch is already known to
be less than 1 in 2**64).
2013-08-13 18:16:34 -07:00
Antoine Pitrou 9ed5f27266 Issue #18722: Remove uses of the "register" keyword in C code. 2013-08-13 20:18:52 +02:00
Raymond Hettinger c629d4c9a2 Replace outdated optimization with clearer code that compiles better.
Letting the compiler decide how to optimize the multiply by five
gives it the freedom to make better choices for the best technique
for a given target machine.

For example, GCC on x86_64 produces a little bit better code:

Old-way (3 steps with a data dependency between each step):

    shrq    $5, %r13
    leaq    1(%rbx,%r13), %rax
    leaq    (%rax,%rbx,4), %rbx

New-way (3 steps with no dependency between the first two steps
         which can be run in parallel):

    leaq    (%rbx,%rbx,4), %rax     # i*5
    shrq    $5, %r13                # perturb >>= PERTURB_SHIFT
    leaq    1(%r13,%rax), %rbx      # 1 + perturb + i*5
2013-08-05 22:24:50 -07:00
Antoine Pitrou 58720d6145 Issue #17934: Add a clear() method to frame objects, to help clean up expensive details (local variables) and break reference cycles. 2013-08-05 23:26:40 +02:00
Raymond Hettinger c86d7e989c Silence compiler warning for an unused declaration 2013-08-04 12:00:36 -07:00
Raymond Hettinger e56666d17f Silence compiler warning about an uninitialized variable 2013-08-04 11:51:03 -07:00
Raymond Hettinger 5ed1b38a7d merge 2013-08-04 11:51:35 -07:00
Serhiy Storchaka 579ddc2fd4 Issue #16741: Fix an error reporting in int(). 2013-08-03 21:14:05 +03:00
Serhiy Storchaka f6d0aeeadc Issue #16741: Fix an error reporting in int(). 2013-08-03 20:55:06 +03:00
Mark Dickinson 02515f7a3a Minor consistency fixes for some longobject.c exception messages:
- replace 'long int' / 'long' by 'int'
 - fix capitalization of "Python" in PyLong_AsUnsignedLong
 - "is too large" -> "too large", for consistency with other messages.
2013-08-03 12:08:22 +01:00