Issue1181301
Created on 2005-04-12 07:37 by mwh, last changed 2005-05-27 15:24 by mwh.
|
msg48200 - (view) |
Author: Michael Hudson (mwh) |
Date: 2005-04-12 07:37 |
|
As discussed on python-dev, this makes the
_PyFloat_{Pack,Unpack}{4,8} functions copy bytes around when
peering at 1.5 has the right result.
|
|
msg48201 - (view) |
Author: Michael Hudson (mwh) |
Date: 2005-04-12 17:18 |
|
Logged In: YES
user_id=6656
New patch, which attacks comments in floatobject.h and implements Tim's
idea of refusing to unpack an IEEE special on a non-IEEE platform.
|
|
msg48202 - (view) |
Author: Michael Hudson (mwh) |
Date: 2005-04-14 07:51 |
|
Logged In: YES
user_id=6656
This version attacks comments some more, uses more interesting probe
values and, um, actually works on a little-endian platform (previous
versions had a very stupid bug).
|
|
msg48203 - (view) |
Author: Michael Hudson (mwh) |
Date: 2005-04-19 16:06 |
|
Logged In: YES
user_id=6656
Stop me if this is getting silly.
This adds class methods __getformat__ and __setformat__ to the float
type and uses them in a new test file, test_float (they have docstrings that
warn people not to use them).
I'd like to check this in soon. Any objections?
|
|
msg48204 - (view) |
Author: Michael Hudson (mwh) |
Date: 2005-05-27 15:24 |
|
Logged In: YES
user_id=6656
Checked in as:
Lib/test/test_float.py revision 1.1
Include/floatobject.h revision 2.23
Include/pythonrun.h revision 2.66
Objects/floatobject.c revision 2.136
Python/pythonrun.c revision 2.214
|
|
| Date |
User |
Action |
Args |
| 2005-04-12 07:37:40 | mwh | create | |
|