Someone may have closed the file descriptor, with something like
f = open('test.test', 'w')
os.close(f.fileno())
f.close()
Protect against this by checking fd on windows before closing.
svn+ssh://pythondev@svn.python.org/python/trunk
........
r70558 | benjamin.peterson | 2009-03-23 17:29:45 -0500 (Mon, 23 Mar 2009) | 4 lines
comply with the evilJavaNamingScheme for attribute names
It seems my love of PEP 8 overrode the need for consistentcy
........
r70561 | benjamin.peterson | 2009-03-23 18:10:14 -0500 (Mon, 23 Mar 2009) | 1 line
refactor unittest docs
........
r70562 | benjamin.peterson | 2009-03-23 18:13:36 -0500 (Mon, 23 Mar 2009) | 1 line
forgot to document that setUp can be skipped (silly me...)
........
r70563 | benjamin.peterson | 2009-03-23 18:19:03 -0500 (Mon, 23 Mar 2009) | 1 line
update from CVS
........
r70568 | benjamin.peterson | 2009-03-23 19:35:20 -0500 (Mon, 23 Mar 2009) | 1 line
some cleanup and modernization
........
r70569 | benjamin.peterson | 2009-03-23 19:36:16 -0500 (Mon, 23 Mar 2009) | 1 line
remove special metadata
........
r70570 | benjamin.peterson | 2009-03-23 19:37:12 -0500 (Mon, 23 Mar 2009) | 1 line
update docstring
........
r70571 | benjamin.peterson | 2009-03-23 19:39:24 -0500 (Mon, 23 Mar 2009) | 1 line
add new skipping things to __all__
........
svn+ssh://pythondev@svn.python.org/python/trunk
........
r70546 | antoine.pitrou | 2009-03-23 19:41:45 +0100 (lun., 23 mars 2009) | 9 lines
Issue #4688: Add a heuristic so that tuples and dicts containing only
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)
........
svn+ssh://pythondev@svn.python.org/python/trunk
........
r70542 | mark.dickinson | 2009-03-23 18:25:13 +0000 (Mon, 23 Mar 2009) | 14 lines
Issue #5512: speed up the long division algorithm for Python longs.
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.
........
svn+ssh://pythondev@svn.python.org/python/trunk
........
r70523 | lars.gustaebel | 2009-03-22 21:09:33 +0100 (Sun, 22 Mar 2009) | 5 lines
Issue #5068: Fixed the tarfile._BZ2Proxy.read() method that would loop
forever on incomplete input. That caused tarfile.open() to hang when used
with mode 'r' or 'r:bz2' and a fileobj argument that contained no data or
partial bzip2 compressed data.
........
svn+ssh://pythondev@svn.python.org/python/trunk
........
r70342 | georg.brandl | 2009-03-13 14:03:58 -0500 (Fri, 13 Mar 2009) | 1 line
#5486: typos.
........
r70385 | benjamin.peterson | 2009-03-15 09:38:55 -0500 (Sun, 15 Mar 2009) | 1 line
fix tuple.index() error message #5495
........
r70386 | georg.brandl | 2009-03-15 16:32:06 -0500 (Sun, 15 Mar 2009) | 1 line
#5496: fix docstring of lookup().
........
r70387 | georg.brandl | 2009-03-15 16:37:16 -0500 (Sun, 15 Mar 2009) | 1 line
#5493: clarify __nonzero__ docs.
........
r70389 | georg.brandl | 2009-03-15 16:43:38 -0500 (Sun, 15 Mar 2009) | 1 line
Fix a small nit in the error message if bool() falls back on __len__ and it returns the wrong type: it would tell the user that __nonzero__ should return bool or int.
........
r70390 | georg.brandl | 2009-03-15 16:44:43 -0500 (Sun, 15 Mar 2009) | 1 line
#5491: clarify nested() semantics.
........
r70392 | georg.brandl | 2009-03-15 16:46:00 -0500 (Sun, 15 Mar 2009) | 1 line
#5488: add missing struct member.
........
r70393 | georg.brandl | 2009-03-15 16:47:42 -0500 (Sun, 15 Mar 2009) | 1 line
#5478: fix copy-paste oversight in function signature.
........
r70395 | georg.brandl | 2009-03-15 16:51:48 -0500 (Sun, 15 Mar 2009) | 1 line
#5276: document IDLESTARTUP and .Idle.py.
........
r70400 | georg.brandl | 2009-03-15 16:59:37 -0500 (Sun, 15 Mar 2009) | 3 lines
Fix markup in re docs and give a mail address in regex howto, so that
the recommendation to send suggestions to the author can be followed.
........
r70405 | georg.brandl | 2009-03-15 17:11:07 -0500 (Sun, 15 Mar 2009) | 7 lines
Move the previously local import of threading to module level.
This is cleaner and avoids lockups in obscure cases where a Queue
is instantiated while the import lock is already held by another thread.
OKed by Tim Peters.
........
r70406 | hirokazu.yamamoto | 2009-03-15 17:43:14 -0500 (Sun, 15 Mar 2009) | 1 line
Added skip for old MSVC.
........
r70418 | georg.brandl | 2009-03-16 14:42:03 -0500 (Mon, 16 Mar 2009) | 1 line
Add token markup.
........
r70438 | benjamin.peterson | 2009-03-17 15:29:51 -0500 (Tue, 17 Mar 2009) | 1 line
I thought this was begging for an example
........
r70464 | benjamin.peterson | 2009-03-18 15:58:09 -0500 (Wed, 18 Mar 2009) | 1 line
a much better example
........
r70468 | benjamin.peterson | 2009-03-18 22:04:31 -0500 (Wed, 18 Mar 2009) | 1 line
close files after comparing them
........
........
r70286 | raymond.hettinger | 2009-03-09 19:06:05 -0500 (Mon, 09 Mar 2009) | 1 line
Fix markup.
........
r70290 | raymond.hettinger | 2009-03-09 20:07:30 -0500 (Mon, 09 Mar 2009) | 1 line
Update url for the spec.
........
r70292 | raymond.hettinger | 2009-03-09 23:40:24 -0500 (Mon, 09 Mar 2009) | 1 line
Clarify the meaning of normal and subnormal.
........
r70293 | raymond.hettinger | 2009-03-09 23:49:21 -0500 (Mon, 09 Mar 2009) | 1 line
Add a version tag to the decimal module.
........
r70295 | raymond.hettinger | 2009-03-10 03:16:05 -0500 (Tue, 10 Mar 2009) | 1 line
Update the decimal FAQ for the from_float() classmethod and improve the recipe for remove_exponent() to make it cut and pasteable.
........
r70296 | raymond.hettinger | 2009-03-10 04:31:48 -0500 (Tue, 10 Mar 2009) | 1 line
Small optimization for corner case where maxlen==0.
........
r70298 | raymond.hettinger | 2009-03-10 07:50:59 -0500 (Tue, 10 Mar 2009) | 1 line
For collections.deque() objects, expose the maxlen parameter as a read-only attribute.
........
r70300 | raymond.hettinger | 2009-03-10 08:04:30 -0500 (Tue, 10 Mar 2009) | 1 line
Fix typo.
........
r70305 | brett.cannon | 2009-03-10 23:51:06 -0500 (Tue, 10 Mar 2009) | 5 lines
Require implementations for warnings.showwarning() support the 'line' argument.
Was a DeprecationWarning for not supporting it since Python 2.6.
Closes issue #3652.
........
r70315 | raymond.hettinger | 2009-03-11 19:25:03 -0500 (Wed, 11 Mar 2009) | 1 line
Add reference to solution for a commonly asked question.
........
r70319 | raymond.hettinger | 2009-03-11 19:31:58 -0500 (Wed, 11 Mar 2009) | 1 line
Issue 5477: Fix buglet in the itertools documentation.
........
r70466 | raymond.hettinger | 2009-03-18 17:13:20 -0500 (Wed, 18 Mar 2009) | 1 line
Use mixin methods where possible. (2.7 only -- these don't all exist in 3.0)
........
r70470 | raymond.hettinger | 2009-03-19 10:21:10 -0500 (Thu, 19 Mar 2009) | 6 lines
Improve implementation with better underlying data structure
for O(1) deletions. Big-Oh performance now the same as regular
dictionaries. Uses a doubly-linked list instead of a list/seq
to track insertion order.
........
r70472 | raymond.hettinger | 2009-03-19 14:24:43 -0500 (Thu, 19 Mar 2009) | 1 line
Silence a compiler warning.
........
r70473 | raymond.hettinger | 2009-03-19 14:59:58 -0500 (Thu, 19 Mar 2009) | 6 lines
* Add clearer comment to initialization code.
* Add optional argument to popitem() -- modeled
after Anthon van der Neut's C version.
* Fix method markup in docs.
........
r70475 | raymond.hettinger | 2009-03-19 18:12:41 -0500 (Thu, 19 Mar 2009) | 6 lines
* Add implementation notes.
* Re-order methods so that those touching the underlying data
structure come first and the derived methods come last.
........
r70477 | raymond.hettinger | 2009-03-19 18:22:25 -0500 (Thu, 19 Mar 2009) | 1 line
Fix typo
........
r70479 | mark.dickinson | 2009-03-20 10:51:55 -0500 (Fri, 20 Mar 2009) | 3 lines
Issue #4258: Use 30-bit digits for Python longs, on 64-bit platforms.
Backport of r70459.
........
r70485 | raymond.hettinger | 2009-03-20 13:25:49 -0500 (Fri, 20 Mar 2009) | 1 line
Add MutableSet example.
........
svn+ssh://pythondev@svn.python.org/python/trunk
........
r70499 | hirokazu.yamamoto | 2009-03-21 19:32:52 +0900 | 1 line
There is no macro named SIZEOF_SSIZE_T. Should use SIZEOF_SIZE_T instead.
........
when packing an integer: for example, struct.pack('=L', -1) now
raises struct.error instead of returning b'\xff\xff\xff\xff'.
Thanks Andreas Schawo for the patch.
svn+ssh://pythondev@svn.python.org/python/trunk
........
r70489 | mark.dickinson | 2009-03-20 23:16:14 +0000 (Fri, 20 Mar 2009) | 4 lines
Rewrite Py_ARITHMETIC_RIGHT_SHIFT so that it's valid for all signed
integer types T, not just those for which "unsigned T" is legal.
........
- new configure option --enable-big-digits
- new structseq sys.int_info giving information about the internal format
By default, 30-bit digits are enabled on 64-bit machines but
disabled on 32-bit machines.
