If Py_NOGIL is defined and Py_SET_REFCNT() is called with a reference
count larger than UINT32_MAX, make the object immortal.
Set _Py_IMMORTAL_REFCNT constant type to Py_ssize_t to fix the
following compiler warning:
Include/internal/pycore_global_objects_fini_generated.h:14:24:
warning: comparison of integers of different signs: 'Py_ssize_t'
(aka 'long') and 'unsigned int' [-Wsign-compare]
if (Py_REFCNT(obj) < _Py_IMMORTAL_REFCNT) {
~~~~~~~~~~~~~~ ^ ~~~~~~~~~~~~~~~~~~~
PEP 683 (immortal objects) revealed some ways in which the Python documentation has been unnecessarily coupled to the implementation details of reference counts. In the end users should focus on reference ownership, including taking references and releasing them, rather than on how many reference counts an object has.
This change updates the documentation to reflect that perspective. It also updates the docs relative to immortal objects in a handful of places.
The Py_CLEAR(), Py_SETREF() and Py_XSETREF() macros now only evaluate
their arguments once. If an argument has side effects, these side
effects are no longer duplicated.
Use temporary variables to avoid duplicating side effects of macro
arguments. If available, use _Py_TYPEOF() to avoid type punning.
Otherwise, use memcpy() for the assignment to prevent a
miscompilation with strict aliasing caused by type punning.
Add _Py_TYPEOF() macro: __typeof__() on GCC and clang.
Add test_py_clear() and test_py_setref() unit tests to _testcapi.
The Py_CLEAR(), Py_SETREF() and Py_XSETREF() macros now only evaluate
their argument once. If an argument has side effects, these side
effects are no longer duplicated.
Add test_py_clear() and test_py_setref() unit tests to _testcapi.
svn+ssh://pythondev@svn.python.org/python/trunk
........
r59343 | georg.brandl | 2007-12-05 08:02:47 +0100 (Wed, 05 Dec 2007) | 2 lines
Fix typo.
........
r59347 | christian.heimes | 2007-12-05 13:31:44 +0100 (Wed, 05 Dec 2007) | 1 line
Fixed quoting and paths in the sqlite project file
........
r59348 | christian.heimes | 2007-12-05 13:45:11 +0100 (Wed, 05 Dec 2007) | 1 line
Fixed error in regrtest. I must have missed the spot.
........
r59350 | christian.heimes | 2007-12-05 13:49:14 +0100 (Wed, 05 Dec 2007) | 1 line
merge -r59315:59316 from py3k: Fix issue #1553: An errornous __length_hint__ can make list() raise a SystemError
........
r59352 | christian.heimes | 2007-12-05 13:52:34 +0100 (Wed, 05 Dec 2007) | 1 line
Added msg to Misc/NEWS
........
r59354 | andrew.kuchling | 2007-12-05 14:27:20 +0100 (Wed, 05 Dec 2007) | 1 line
Spelling fix
........
r59356 | georg.brandl | 2007-12-05 18:56:50 +0100 (Wed, 05 Dec 2007) | 3 lines
Add examples to csv, pprint and traceback docs.
Written by Ross for GHOP.
........
r59358 | raymond.hettinger | 2007-12-05 19:11:08 +0100 (Wed, 05 Dec 2007) | 1 line
Error checking was too aggressive (reported by Chris Tismer)
........
r59359 | georg.brandl | 2007-12-05 19:30:48 +0100 (Wed, 05 Dec 2007) | 2 lines
Add examples to re docs. Written for GHOP by Dan Finnie.
........
r59366 | georg.brandl | 2007-12-05 20:49:21 +0100 (Wed, 05 Dec 2007) | 2 lines
Fix markup.
........
r59367 | christian.heimes | 2007-12-05 20:57:54 +0100 (Wed, 05 Dec 2007) | 1 line
Updated documentation and build_tkinter.py script
........
r59368 | georg.brandl | 2007-12-05 21:03:57 +0100 (Wed, 05 Dec 2007) | 2 lines
Another markup fix.
........
r59369 | ronald.oussoren | 2007-12-05 21:07:36 +0100 (Wed, 05 Dec 2007) | 7 lines
This "fixes" compilation issues for the Carbon._OSA module on OSX Leopard
by purging bindings to OSA's debug API's. Those APIs we're completely
unsupported on OSX 10.4 and are no longer available on OSX 10.5.
Note that this patches a generated file. This is somewhat acceptable because
regenerating the file is non-trivial and wouldn't use system headers anyway.
........
r59370 | christian.heimes | 2007-12-05 21:10:38 +0100 (Wed, 05 Dec 2007) | 1 line
Fixed bug #1557 by using popen.communicate() before popen.wait()
........
