mirror of
https://github.com/python/cpython.git
synced 2025-10-06 23:21:06 +00:00
Merged revisions 64365,64370,64406,64408-64409,64412,64416-64417,64420-64421,64425-64428 via svnmerge from
svn+ssh://pythondev@svn.python.org/python/trunk ........ r64365 | raymond.hettinger | 2008-06-17 19:56:57 -0500 (Tue, 17 Jun 2008) | 1 line Fix double decref. ........ r64370 | mark.dickinson | 2008-06-18 04:20:17 -0500 (Wed, 18 Jun 2008) | 2 lines Typo fix ........ r64406 | andrew.kuchling | 2008-06-19 09:02:30 -0500 (Thu, 19 Jun 2008) | 1 line Only include update_lines_cols() function when it's actually going to be used ........ r64408 | amaury.forgeotdarc | 2008-06-19 14:57:39 -0500 (Thu, 19 Jun 2008) | 2 lines test_macos can be skipped on non-mac platforms. ........ r64409 | andrew.kuchling | 2008-06-19 15:33:31 -0500 (Thu, 19 Jun 2008) | 1 line Put threading in front of thread ........ r64412 | amaury.forgeotdarc | 2008-06-19 16:17:12 -0500 (Thu, 19 Jun 2008) | 3 lines In test_site, correctly escape backslashes in path names. This allows the test to pass when the username begins with a lowercase 't'... ........ r64416 | vinay.sajip | 2008-06-19 17:40:17 -0500 (Thu, 19 Jun 2008) | 2 lines Bug #3136: fileConfig()'s disabling of old loggers is now conditional via an optional disable_existing_loggers parameter, but the default value is such that the old behaviour is preserved. Thanks to Leandro Lucarella for the patch. ........ r64417 | vinay.sajip | 2008-06-19 17:41:08 -0500 (Thu, 19 Jun 2008) | 1 line Updated with fix for #3136. ........ r64420 | andrew.kuchling | 2008-06-19 21:05:57 -0500 (Thu, 19 Jun 2008) | 1 line Various items ........ r64421 | andrew.kuchling | 2008-06-19 21:11:42 -0500 (Thu, 19 Jun 2008) | 1 line Fix comment typos ........ r64425 | andrew.kuchling | 2008-06-20 06:39:54 -0500 (Fri, 20 Jun 2008) | 1 line Various items ........ r64426 | mark.dickinson | 2008-06-20 09:53:43 -0500 (Fri, 20 Jun 2008) | 4 lines Issue #3004: Minor fix to slice.indices(). slice(-10).indices(9) now returns (0, 0, 1) instead of (0, -1, 1), and slice(None, 10, -1).indices(10) returns (9, 9, -1) instead of (9, 10, -1). ........ r64427 | mark.dickinson | 2008-06-20 10:17:41 -0500 (Fri, 20 Jun 2008) | 2 lines Fix outdated count of the number of new math module functions. ........ r64428 | mark.dickinson | 2008-06-20 10:26:19 -0500 (Fri, 20 Jun 2008) | 2 lines Fix another typo in math_sum comment ........
This commit is contained in:
parent
464109a893
commit
fea6a94b3b
7 changed files with 30 additions and 14 deletions
|
@ -382,11 +382,11 @@ FUNC1(tanh, tanh, 0,
|
|||
sum([1e+308, 1e+308, 1e-308]) raises an OverflowError due to the
|
||||
overflow of the first partial sum.
|
||||
|
||||
Note 3: The itermediate values lo, yr, and hi are declared volatile so
|
||||
aggressive compilers won't algebraicly reduce lo to always be exactly 0.0.
|
||||
Note 3: The intermediate values lo, yr, and hi are declared volatile so
|
||||
aggressive compilers won't algebraically reduce lo to always be exactly 0.0.
|
||||
Also, the volatile declaration forces the values to be stored in memory as
|
||||
regular doubles instead of extended long precision (80-bit) values. This
|
||||
prevents double rounding because any addition or substraction of two doubles
|
||||
prevents double rounding because any addition or subtraction of two doubles
|
||||
can be resolved exactly into double-sized hi and lo values. As long as the
|
||||
hi value gets forced into a double before yr and lo are computed, the extra
|
||||
bits in downstream extended precision operations (x87 for example) will be
|
||||
|
@ -614,7 +614,6 @@ math_factorial(PyObject *self, PyObject *arg)
|
|||
|
||||
error:
|
||||
Py_DECREF(result);
|
||||
Py_XDECREF(iobj);
|
||||
return NULL;
|
||||
}
|
||||
|
||||
|
|
Loading…
Add table
Add a link
Reference in a new issue